Q2 2026 eMemory Technology Inc Earnings Call

August 14, 2026

Speaker #1: Good afternoon, and welcome to eMemory's Q2 2026 webcast investor conference. Joining us today are our Chairman, Dr. Charles Hsu; Head of IR, Ms. Li Jingchen; Director of the Finance Department, Mr. Joseph Hsia; and Head of Digital Marketing, Dr. Felix Hsu.

Li Ching-Xiang: Good afternoon, and welcome to eMemory's Q2 2026 webcast investor conference. Joining us today is our Chairman, Dr. Charles Hsu, Head of IR, Ms. Li Ching-Xiang, Director of the Finance Department, Mr. Joseph Hsia, and Head of Digital Marketing, Dr. Felix Hsu. The format of today's event will be as follows. First, eMemory's Chairman, Dr. Charles Hsu, will give an opening remark. Afterwards, our financial officer, Mr. Joseph Hsia, will present a review of our financial results. Following that, Dr. Charles Hsu will share our business outlook. Next, Dr. Felix Hsu will give a talk titled, "Securing the Next Generation of AI Infrastructure, the Hardware Anchor for Caliptra Root of Trust to Secure Chiplets and Compute Express Link or CXL." Then, we will conclude today's conference with the Q&A section, where our management team will answer your questions.

Operator: Good afternoon, and welcome to eMemory's Q2 2026 webcast investor conference. Joining us today is our Chairman, Dr. Charles Hsu, Head of IR, Ms. Li-Jeng Chen, Director of the Finance Department, Mr. Joseph Hsia, and Head of Digital Marketing, Dr. Felix Hsu. The format of today's event will be as follows. First, eMemory's Chairman, Dr. Charles Hsu, will give an opening remark. Afterwards, our financial officer, Mr. Joseph Hsia, will present a review of our financial results. Following that, Dr. Charles Hsu will share our business outlook. Next, Dr. Felix Hsu will give a talk titled, "Securing the Next Generation of AI Infrastructure, the Hardware Anchor for Caliptra Root of Trust to Secure Chiplets and Compute Express Link or CXL." Then, we will conclude today's conference with the Q&A section, where our management team will answer your questions.

Speaker #1: The format of today's event will be as follows: First, eMemory's chairman, Dr. Charles Hsu, will give an opening remark. Afterwards, our financial officer, Mr. Joseph Hsia, will present the review of our financial results. Following that, Dr. Charles Hsu will share our business outlook. Next, Dr. Felix Hsu will give a talk titled, "Securing the Next Generation of AI Infrastructure: The Hardware Anchor for Calyptra Root of Trust to Secure Chiplets and Compute Express Link, or CXL." Then, we will conclude today's conference with the Q&A section, where our management team will answer your questions.

Speaker #1: Please feel free to submit your questions in the input box on the webcast window throughout the conference. As a reminder, this conference is being recorded and a webcast replay will be available after the conference is finished.

Li Ching-Xiang: Please feel free to submit your questions in the input box on the webcast window throughout the conference. As a reminder, this conference is being recorded and a webcast replay will be available after the conference is finished. For more information, please visit the company's website under the investor relations section. As usual, before we begin, we would like to remind everyone that today's presentation may contain forward-looking statements subject to risk factors associated with the semiconductor and IP business. Please refer to the cautionary statement on page 3 of today's presentation. Now, I would like to give the floor over to eMemory's Chairman, Dr. Charles Hsu.

Operator: Please feel free to submit your questions in the input box on the webcast window throughout the conference. As a reminder, this conference is being recorded and a webcast replay will be available after the conference is finished. For more information, please visit the company's website under the investor relations section. As usual, before we begin, we would like to remind everyone that today's presentation may contain forward-looking statements subject to risk factors associated with the semiconductor and IP business. Please refer to the cautionary statement on page 3 of today's presentation. Now, I would like to give the floor over to eMemory's Chairman, Dr. Charles Hsu.

Speaker #1: For more information, please visit the company's website under the Investor Relations section. As usual, before we begin, we would like to remind everyone that today's presentation may contain forward-looking statements subject to risk factors associated with the semiconductor and IP business.

Speaker #1: Please refer to the cautionary statement on page 3 of today's presentation. Now, I would like to give the floor to eMemory's chairman, Dr. Charles Hsu.

Speaker #2: Okay. Good afternoon, shareholders and investors. Welcome to our investor conference. First, I would like to report on our latest progress in security IP.

Charles Hsu: Okay. Good afternoon, shareholders and investors. Welcome to our investor conference. First, I would like to report on our latest progress in security IP. Over the past few years, our security IP has evolved from the foundational OTP and the PUF into a PUF root of trust. Today, as security requirements for AI servers continue to rise, our technology has taken a major step forward, advancing from a standalone security IP to the design of integration of comprehensive security subsystems. This transition is crucial for us. We are no longer providing clients with just OTP, PUF, or root of trust. We are now delivering fully integrated system-level security IP that combines encryption technology, software, firmware, and anti-tampering protections. In other words, our role within our clients' chips has become far more critical, allowing us to deliver significantly higher value.

Speaker #2: Over the past few years, our security IP has evolved from a foundational OTP and path into a top rule-of-trust. Today, as security requirements for AI servers continue to rise, our technology has taken a major step forward, advancing from a standalone security IP to a design integrating comprehensive security systems.

Speaker #2: This transition is crucial for us. We are no longer providing clients with just OTP path or rule-of-trust; we are now delivering fully integrated, system-level security IP that combines encryption technology, software firmware, and anti-tampering protections.

Speaker #2: In other words, our role within our clients' chips has become far more critical, allowing us to deliver significantly higher value. Consequently, the potential license fee and also royalty are expected to increase significantly as well.

Charles Hsu: Consequently, the potential license fee and also royalty are expected to increase significantly as well. We have a license to several major global memory manufacturers to integrate our security IP into SSD, solid state disk, systems for AI data centers, providing the security subsystems required to comply with Caliptra standards. Beyond SSDs, we have also expanded into another vital area of AI servers, which is memory expansion, IE. As AI model grows increasingly large, the memory capacity required by AI servers is surging rapidly. As a result, CXL is becoming a dominant interface technology for AI server memory expansion. CXL switch and their related controller chips similarly require a complete hardware security architectures. We are currently working with several CXL chip makers to supply both security IP and the security subsystems. Connecting the dots over the past few years reveal a clear trajectory for our security IP applications.

Speaker #2: We have licensed to several major global memory manufacturers to integrate our security IP into SSD (solid-state disk) systems. For AI data centers, we are providing the security subsystems required to comply with Calyptra standards.

Speaker #2: Beyond SSDs, we have also expanded into another vital area of AI servers, which is memory expansion. I see. As AI models grow increasingly large, the memory capacity required by AI servers is surging rapidly.

Speaker #2: As a result, CXL is becoming a dominant interface technology for AI servers and memory expansion. CXL switches and their related controller chips similarly require complete hardware security architectures.

Speaker #2: We are currently working with several CXL chip makers to supply both security IP and security subsystems. So, connecting the dots over the past few years reveals a clear trajectory for our security IP applications.

Speaker #2: Last year, we entered AI AGI CPUs and also BMC, and this year we are expanding into AI data center solid-state disk security subsystems. Now, we are penetrating CXL switch for AI server memory expansion.

Charles Hsu: Last year, we entered AI, AGI, CPUs, and also BMC. This year, we expand into AI data center, solid state disk security subsystems. Now we are pre-penetrating CXL switch for the AI server memory expansions. This means our security IP is expanding from a single chip inside an AI server to multiple critical chips. Furthermore, many of these advanced chips are utilizing cutting-edge process nodes, such as 3 nanometer, and generating much higher licensing and royalties than corporate average. Another highly favorable trend for us is collateral. Driven jointly by global tech giants such as Microsoft, Google, Intel, AMD, and NVIDIA, we are seeing an increasing number of AI and data center chips and the system adopting the Caliptra security architecture. This indicates that the number of chip requiring root of trust, security subsystem and the related security IP will grow significantly moving forward.

Speaker #2: This means our security IP is expanding from a single chip inside an AI server to multiple chips. Furthermore, many of these advanced chips are utilizing cutting-edge process nodes such as 3-nanometer and are generating much higher licensing and royalties than the corporate average.

Speaker #2: Another highly favorable trend for us is Calyptra. Driven jointly by global giants such as Microsoft, Google, Intel, AMD, and NVIDIA, we are seeing an increase in the number of AI and data center chips and systems adopting the Calyptra security architecture.

Speaker #2: This indicates that the number of chips required in rule-of-trust security subsystems and related security IP will grow significantly moving forward. Therefore, we remain highly confident in the long-term growth of our security IP.

Speaker #2: We think the AI server and data center markets, and next I would like to also share some updates on another key growth driver we are heavily focused on, which we call logic flash technologies.

Charles Hsu: Therefore, we remain highly confident in the long-term growth of our security IP within the AI server and the data center markets. Next, I would like to also share some updates on another key growth driver we are heavily focused on, which is the Logic-Flash technologies. Currently, our collaboration with foundries on Logic-Flash technology is accelerating applications extend beyond embedded Flash to include standalone Flash products as well. We believe the significance of Logic-Flash goes far beyond adding another IP technology to our portfolios. More importantly, it presents an opportunity to leverage a more competitive process and the cost structure to gradually replace portion of existing legacy Flash technologies. Flash memory represents a massive market with exceptionally broad applications. Once Logic-Flash enters mass production and it begins progressively replacing conventional Flash technologies, the market opportunity and the long-term impact on our future revenue and profitability will be far-reaching.

Speaker #2: Currently, our collaboration with foundries on logic flash technology is accelerating. Applications extend beyond embedded flash to include standalone flash products as well. We believe the significance of logic flash goes far beyond adding another IP technology to our portfolios.

Speaker #2: More importantly, it presents an opportunity to leverage a more competitive process and cost structure to gradually replace portions of existing legacy flash technologies. Flash memory represents a massive market with exceptionally broad applications.

Speaker #2: Once logic flash enters mass production and begins progressively replacing conventional flash technologies, the market opportunity and the long-term impact on our future revenue and profitability will be far-reaching.

Speaker #2: For these reasons, we have strong confidence in our multi-year growth outlook ahead. Next, I would like to invite our financial officer, Joseph. Joseph, please present our second quarter performance.

Charles Hsu: For these reasons, we have a strong confidence in our multi-year growth outlook ahead. Next, I would like to invite our financial officer, Joseph, to present our Q2 performance. Thank you.

Speaker #2: Thank you.

Speaker #3: Good afternoon, everyone. Now let's begin with our 2026 second quarter financial results. The second quarter revenue was $1,097 million, up 0.2% sequentially and up 17.1% year over year.

Speaker #3: Operating expenses were $431 million, down 0.3% sequentially and up 10.2% year over year. As a result, our operating income was $666 million, with an increase of 0.6% sequentially and an increase of 22% year over year.

Joseph Hsia: Good afternoon, everyone. Now let's begin with our 2026 Q2 financial results. The Q2 revenue was TWD 1,097 million, up 0.2% sequentially and up 17.1% year over year. Operating expenses were TWD 431 million, down 0.3% sequentially and up 10.2% year over year. In results, our operating income was TWD 666 million, with an increase of 0.6% sequentially and an increase of 22% year over year. Operating margin also increased by 2.2 percentage points sequentially and increased by 2.4 percentage points year over year to 60.7%. Earning income amounting to TWD 580 million, experienced a decrease of 2.8% sequentially, but an increase of 44.9% year over year. The EPS for this quarter was NT$7.77. Next, let's move on to revenue contributions by licensing and royalty. First of all, licensing in the Q2 accounted for 39.2% of the total revenue, increasing 12.8% sequentially and up 35.1% year over year.

Speaker #3: Operating margin also increased by 0.2 percentage points sequentially and increased by 2.4 percentage points year over year to 60.7%. Our net income, amounting to $580 million, experienced a decrease of 2.8% sequentially but an increase of 44.9% year over year.

Speaker #3: DPS for this quarter was 7.77. Next, let's move on to revenue contributions by licensing and royalty. First of all, licensing in the second quarter accounted for 39.2% of total revenue, increasing 12.8% sequentially and up 35.1% year over year.

Speaker #3: On a US dollar basis, licensing grew by 12.8% quarter over quarter and 32.6% year over year. Royalty in the second quarter contributed 60.8% of our total revenue, decreasing 6.4% sequentially but increasing 7.8% year over year.

Speaker #3: On a US dollar basis, there was a 6.7% decrease quarter over quarter, but an increase of 8.6% year over year. Overall, the revenue increased by 0.2% quarter over quarter and was up 17.1% year over year.

Joseph Hsia: On the USD basis, licensing grew by 12.8% quarter over quarter and 32.6% year over year. Royalty in Q2 contributed 60.8% of our total revenue, decreasing 6.4% sequentially, but increasing 7.8% year over year. On USD basis, there was a 6.7% decrease quarter over quarter, but an increase of 8.6% year over year. Overall, the revenue increased by 0.2% quarter over quarter and up 17.1% year over year. On USD basis, the growth was 0.1% quarter over quarter and 16.9% year over year. For H1 2026, the licensing and royalty revenues are as follows. First of all, licensing in H1 accounted for 37% of our total revenue, increasing 45.2% year over year. On USD basis, licensing grew by 45.9% year over year.

Speaker #3: On a US dollar basis, the growth was 0.1% quarter over quarter and 16.9% year over year. And for the first half of 2026, the licensing and royalty revenues are as follows.

Speaker #3: First of all, licensing in the first half accounted for 37% of our total revenue, increasing 45.2% year over year. And on a US dollar basis, licensing grew by 45.9% year over year.

Speaker #3: And royalty in the first half contributed 63% of the total revenue, increasing 7% year over year, and on a US dollar basis, there was a 9.8% increase year over year.

Speaker #3: And to get the total revenue for the first half increased by 18.5% compared to the previous quarter. On a US dollar basis, the increase was stronger at 20.9% year-over-year.

Speaker #3: And with that, I will comment further on our revenue contribution by specific IPs. First of all, Corporate accounted for 21.9% of total revenue in the second quarter.

Joseph Hsia: Royalty in H1 contributed 63% of the total revenue, increasing 7% year over year. On the USD basis, there was 9.8% increase year over year. Together, the total revenue for H1 increased by 18.5% compared to the previous quarter, and on USD basis, the growth was even stronger at 20.9% year over year. With that, I will comment further on our revenue contribution by specific IPs. First of all, NeoBit accounted for 21.9% of total revenue in Q2. The licensing revenue increased 11.2% sequentially, but decreasing 22.4% year over year, while royalty increased by 9% sequentially and increasing 9.2% year over year. For NeoFuse technology, it accounted for 56.6% of the total revenue in Q2. The licensing revenue was up by 24.8% sequentially and up by 55.6% year over year.

Speaker #3: The licensing revenue increased 11.2% sequentially but decreased 22.4% year over year, while royalty increased by 9% sequentially and increased 9.2% year over year. For NeoFuse technology, it accounted for 56.6% of the total revenue in the second quarter.

Speaker #3: The licensing revenue was up by 24.8% sequentially and up by 55.6% year over year. In terms of royalty, neofield royalty decreased by 13.3% sequentially but increased by 4.4% year over year.

Speaker #3: And for path-based security IPs, it contributed 10.9% of total revenue. The licensing revenue decreased 13.8% sequentially but increased 133.5% year over year. In terms of royalty, path-based royalty increased by 213.9% sequentially and increased by over 1,600% year over year.

Joseph Hsia: In terms of royalty, NeoFuse royalty decreased by 13.3% sequentially, but increased by 4.4% year over year. For PUF-based secured IPs, it contributed 10.9% of total revenue. The licensing revenue decreased 13.8% sequentially, but increased 133.5% year over year. In terms of royalty, PUF-based royalty increased by 213.9% sequentially and increased by over 1,600% year over year. Lastly, for NeoMTP technology accounted for 10.6% of the total revenue in Q2. The licensing revenue increased by 40.2% sequentially and increased by 9.2% year over year. The royalty from NeoMTP was up 7.2% sequentially and increased by 19.2% year over year. Together, for H1 2026, the revenue by technology are as follows.

Speaker #3: And lastly, for MTP technology, it accounted for 10.6% of total revenue in the second quarter. The licensing revenue increased by 40.2% sequentially and increased by 9.3% year over year.

Speaker #3: The royalty from MTP was up 7.2% sequentially and increased by 19.2% year over year. And to get the first half of 2026, the revenue by technology is as follows:

Speaker #3: First of all, NeoFuse licensing revenue decreased by 18.5% year-over-year, but the royalty increased by 2.4%. Together, they accounted for 12.9% of the total revenue for the first half of 2026.

Speaker #3: For NeoFuse, the licensing revenue increased by 37.9%, and royalty revenue increased by 6% year over year, rising to 58.2% of our total revenue in the first half.

Speaker #3: And for path-based security IPs, the licensing revenue increased by 234.6% year over year. The royalty increased by over 1,300% year over year, and together accounted for 11.6% of our total revenue in the first half.

Joseph Hsia: First of all, NeoBit licensing revenue decreased by 18.5% year over year, but the royalty increased by 2.4% and together accounting for 20.9% of the total revenue for H1 2026. For NeoFuse, the licensing revenue increased by 37.9%, and the royalty also increased by 6% year over year, contributing to 58.2% of our total revenue in H1. For PUF-based secured IPs, the licensing revenue increased by 264.6% year over year, and the royalty increased by over 1,300% year over year, and together accounting for 11.6% of our total revenue in H1. Lastly, for NeoMTP technology, the licensing revenue increased by 19.6% and the royalty increased by 30.4% year over year, and together accounting for 9.3% of the total revenue. Now let's take a look at our royalties for 8-inch and 12-inch wafers.

Speaker #3: And lastly, for MTP technology, the licensing revenue increased by 19.6% and royalty increased by 30.4% year over year, together accounting for 9.3% of the total revenue.

Speaker #3: And now let's take a look at our royalties for 8-inch and 12-inch quarters. First of all, 8-inch and 12-inch accounted for 37.5% of the royalties, up 4.2% sequentially, and up 1.2% year over year.

Speaker #3: And on a US dollar basis, this represents a sequential increase of 4% and a year-over-year increase of 3.1%. For 12 and 12, it contributed 62.5% of the total royalties, down 11.9% sequentially but up 12.3% year over year.

Speaker #3: On a US dollar basis, this represents a sequential decrease of 12.1%, but a year-over-year increase of 12.2%. In total, 156 product tape-outs were completed in the second quarter, and we will provide more information in our management report, which will be released shortly after this earnings call.

Joseph Hsia: First of all, 8-inch wafers accounted for 37.5% of the royalties, up 4.2% sequentially and up 1.2% year over year. On US dollar basis, this represents a sequential increase of 4% and a year over year increase of 3.1%. For 12-inch wafers, it contributed 62.5% of the total royalties, down 11.9% sequentially, but increased 12.3% year over year. On US dollar basis, this represents a sequential decrease of 12.1%, but a year over year increase of 12.2%. In total, 156 product tape-outs were completed in Q2, and we will provide more information in our management report, which will be released shortly after this earnings call. Next, I would like to invite our Chairman, Charles, to share a little bit more about our future outlook. Thank you.

Speaker #3: And next, I would like to invite our Chairman, Charles, to share a little bit more about our future outlook. Thank you.

Speaker #2: Okay. In the following section, I will address our future outlook. As far as the licensing revenue is concerned, licensing will continue its strong momentum due to robust demand for our technologies from leading-edge to legacy process node security and the next generation fresh technologies.

Speaker #2: And for the royalty revenue, the royalty revenue growth is expected to accelerate, driven by the higher ASP from the new advanced norm applications and the new application ramps, and also expanding the PUF royalty contribution and growing mix of higher royalty rate of MTP-related applications.

Charles Hsu: Okay, in the following section, I will address our future outlook. As far as the licensing revenue concerned, licensing will continue its strong momentum due to robust demand for our technologies from leading edge to the legacy process node security and the next generation Flash technologies. For the royalty revenue, growth is expected to accelerate, driven by the higher ASP from the new advanced node applications and the new application ramps, also expanding the PUF royalty contribution and a growing mix of higher royalty rate of MTP-related applications.

Speaker #2: And as far as new technologies, for the advanced norm OTP and the path-based hardware security continue and qualify next generation OTP and the path-based hardware security solutions for two nanometer get all around technology and also sub two nanometer nodes.

Speaker #2: And meeting the growing customer demand in device identity and key protections, and to secure both, and also hardware root of trust. And for the next, another new technology will be next-generation 1T flash—1T NeoFlash technology is advancing across both embedded and standalone applications, and its part logic process-compatible architecture offers greater scalability, lower process complexity, and better cost efficiency.

Charles Hsu: As far as new technologies, for the advanced node, OTP and the PUF-based hardware security continue to develop and qualify next generation OTP and the PUF-based hardware security solutions for 2 nanometer Gate-all-around technology and also sub 2 nanometer nodes and meeting growing customer demand in the device identity and key protections and secure boots and also hardware root of trust. Another new technology will be next generation 1T Flash. Our 1T NeoFlash technology is advancing across embedded also standalone applications, and its logic process compatible architecture offers greater scalability, lower process complexity, and better cost efficiency. In the future, outlook for the business development platforms, which we have five items, the first is chiplet security platform.

Speaker #2: And in the future, the outlook for the business development platforms, we have five items. The first is the capability security platform. We continue to work with ecosystem partners on an end-to-end security framework for chiplet-based systems, covering supply chain traceability, identity, and authentication.

Speaker #2: And also secure provisioning, as well as day-to-day communication, and also key management and hardware root of trust—addressing the increasingly complex security challenges of AI.

Speaker #2: And also the advancing packaging and heterogeneous integration among today's geopolitical environments. Okay. And the second item is the data center security and collateral platform.

Charles Hsu: We continue to work with ecosystem partners on an end-to-end security framework for chiplet-based systems covering supply chain traceability, identity and authentication, also secure provisioning, also the die to die communication, also the key management and the hardware root of trust, addressing the increasingly complex security challenges of AI and also the advancing packaging and heterogeneous integration among today's geopolitical environments.

Speaker #2: Targeting data center and AI servers, we continue to upsell, expanding the path rule of trust from hardware rule of trust solutions into a collateral-compatible security subsystem.

Speaker #2: An integration service that reduces integration complexity and accelerates customer deployments. Another business platform is the AI compute and also the rule of trust platform.

Felix Hsu: Okay. The second item is the data center security and the Caliptra platform. Targeting the data center and the AI servers, we continue to upsell expanding PUF, our Root of Trust, from a hardware Root of Trust solution into a Caliptra-compatible security subsystem and the integration service that reduce the integration complexity and accelerate customer deployments. Another business platform is on the AI compute and also Root of Trust platform. We are extending collaboration across CPU, AI accelerator, and AI ASIC ecosystem to integrate chip-level Root of Trust, secure boot device identity, and also secure key protection into the next generation AI computing platform, strengthening trust and also security from silicon to a system. Another business platform we are developing is PUFhsm server and security as a service platform.

Charles Hsu: Okay. The second item is the data center security and the Caliptra platform. Targeting the data center and the AI servers, we continue to upsell expanding PUF, our Root of Trust, from a hardware Root of Trust solution into a Caliptra-compatible security subsystem and the integration service that reduce the integration complexity and accelerate customer deployments. Another business platform is on the AI compute and also Root of Trust platform. We are extending collaboration across CPU, AI accelerator, and AI ASIC ecosystem to integrate chip-level Root of Trust, secure boot device identity, and also secure key protection into the next generation AI computing platform, strengthening trust and also security from silicon to a system. Another business platform we are developing is PUFhsm server and security as a service platform.

Speaker #2: We are extending collaboration across CPU, AI accelerator, and AI ASIC ecosystem to integrate chip level rule of trust secure boot device identity and also secure key protection into the next generation AI computing platform strengthening trust and also from system in from system to system.

Speaker #2: And another business platform we are developing is HSMH server and security service platform. The path-based HSMH server combines device identity, key, and certification management and secure OTA updates signature verification and privacy protection and post-quantum cryptography migration early opportunities OTA and PKI and also HSM integration with potential to expand into the industrial control and edge AI and smart device medical and also the data centers.

Felix Hsu: The PUF-based PUFhsm server combines device identity, key and certification management, and the secure OTA updates, signature verification, and the privacy protection and the post-quantum cryptography migration. Early opportunities are progressing in the automotive, OTA, and PKI, and also HSM integration, with potential to expand it into the industrial control and the edge AI, smart device medical, and also the data centers. The last platform is, we call it the post-quantum security platform. We continue to strengthen our PUF-PQC portfolio with attack resistant hardware security, including side-channel protection to support the transition to post-quantum security standards. Okay. Next, I would like to pass to Felix, our head of digital marketing, to share with our feature topic today. Felix? Please. Thank you.

Charles Hsu: The PUF-based PUFhsm server combines device identity, key and certification management, and the secure OTA updates, signature verification, and the privacy protection and the post-quantum cryptography migration. Early opportunities are progressing in the automotive, OTA, and PKI, and also HSM integration, with potential to expand it into the industrial control and the edge AI, smart device medical, and also the data centers. The last platform is, we call it the post-quantum security platform. We continue to strengthen our PUF-PQC portfolio with attack resistant hardware security, including side-channel protection to support the transition to post-quantum security standards. Okay. Next, I would like to pass to Felix, our head of digital marketing, to share with our feature topic today. Felix? Please. Thank you.

Speaker #2: And the last platform is what we call the post-quantum security platform. So we continue to strengthen our path PQC portfolio with attack system-resistant hardware security, including side-channel protection, to support the transition to post-quantum security standards.

Speaker #2: Okay, so next I would like to pass to Felix, our Head of Digital Marketing, to share with us our future topic today. Felix, please. Thank you.

Speaker #3: Modern data centers depend on thousands of interconnected servers, each processing sensitive workloads, credentials, and proprietary data. But system-level security ultimately begins much deeper—inside the silicon.

Speaker #3: A server board contains chiplets of CPUs, accelerators, BMCs, networking storage controllers, and other programmable devices. Each component loads firmware and communicates across shared interfaces.

Speaker #3: Creating multiple points where trust must be established. Without a consistent hardware root of trust, compromised firmware, counterfeit components, leaked keys, or unauthenticated communication can allow an attacker to move across the board and persist below the operating system.

[Company Representative] (eMemory): Modern data centers depend on thousands of interconnected servers, each processing sensitive workloads, credentials, and proprietary data. System-level security ultimately begins much deeper inside the silicon. A server board contains chiplets of CPUs, accelerators, BMCs, networking storage controllers, and other programmable devices. Each component loads firmware and communicates across shared interfaces, creating multiple points where trust must be established. Without a consistent hardware Root of Trust, compromised firmware, counterfeit components, leaked keys, or unauthenticated communication can allow an attacker to move across the board and persist below the operating system. Caliptra addresses this challenge with an open source Root of Trust architecture designed for data center class silicon. It establishes a standardized security foundation that can be integrated into each major System-on-a-Chip. During boot, the Caliptra subsystem begins from immutable code, authenticates its firmware, measures the device state, and derives protected identities and cryptographic keys.

[Video Narrator]: Modern data centers depend on thousands of interconnected servers, each processing sensitive workloads, credentials, and proprietary data. System-level security ultimately begins much deeper inside the silicon. A server board contains chiplets of CPUs, accelerators, BMCs, networking storage controllers, and other programmable devices. Each component loads firmware and communicates across shared interfaces, creating multiple points where trust must be established. Without a consistent hardware Root of Trust, compromised firmware, counterfeit components, leaked keys, or unauthenticated communication can allow an attacker to move across the board and persist below the operating system. Caliptra addresses this challenge with an open source Root of Trust architecture designed for data center class silicon. It establishes a standardized security foundation that can be integrated into each major System-on-a-Chip. During boot, the Caliptra subsystem begins from immutable code, authenticates its firmware, measures the device state, and derives protected identities and cryptographic keys.

Speaker #3: Calipra addresses this challenge with an open-source root of trust architecture designed for data center-class silicon. It establishes a standardized security foundation that can be integrated into each major system-on-chip.

Speaker #3: During boot, the Calipra subsystem begins from immutable code, authenticates its firmware, measures device state, and derives protected identities and cryptographic keys. It can then support secure boots, signed firmware updates, and device attestations. However, the security architecture still requires a trustworthy physical source for device secrets, lifecycle data, and entropy.

Speaker #3: PUF RT provides this silicon-level foundation directly beneath the Calipra core. PUF generates a device-unique secret without permanently storing the key. OTP securely retains configuration and lifecycle information, while the hardware noise source supplies high-quality entropy for cryptographic operations.

Speaker #3: With Calipra deployed across critical devices, components can authenticate one another, verify firmware, establish protected channels, exchange keys, and produce evidence of their security state.

[Company Representative] (eMemory): It can then support secure boot, signed firmware updates, and device attestation. However, the security architecture still requires a trustworthy physical source for device secrets, life cycle data, and entropy. PUFrt provides this silicon level foundation directly beneath the Caliptra core. PUF generates a device-unique secret without permanently storing the key. OTP securely retains configuration and life cycle information, while the hardware noise source supplies high-quality entropy for cryptographic operations. With Caliptra deployed across critical devices, components can authenticate one another, verify firmware, establish protected channels, exchange keys, and produce evidence of their security state. The result is board-level security built from independently verifiable chips, extending a consistent chain of trust from each component across the server platform and throughout the data center.

[Video Narrator]: It can then support secure boot, signed firmware updates, and device attestation. However, the security architecture still requires a trustworthy physical source for device secrets, life cycle data, and entropy. PUFrt provides this silicon level foundation directly beneath the Caliptra core. PUF generates a device-unique secret without permanently storing the key. OTP securely retains configuration and life cycle information, while the hardware noise source supplies high-quality entropy for cryptographic operations. With Caliptra deployed across critical devices, components can authenticate one another, verify firmware, establish protected channels, exchange keys, and produce evidence of their security state. The result is board-level security built from independently verifiable chips, extending a consistent chain of trust from each component across the server platform and throughout the data center.

Speaker #3: The result is board-level security built from independently verifiable chips, extending a consistent chain of trust from each component across the server platform and throughout the data center.

Speaker #1: eMemory and PUF Security: your trusted partner in providing non-volatile memory and PUF-based security intellectual properties.

Speaker #2: Okay. Hi everyone. What you have just viewed in this video is one part of a much bigger movement happening around Calipra and hardware root security.

Speaker #2: Before I begin this talk, I'd like to briefly mention some related activities we've just been involved in. Earlier this week in Taipei, we participated in a Calipra workshop focused on implementation and on bringing the specification closer to production silicon.

Speaker #2: We also presented at two sessions at OCPAPEC, where we discussed Calipra and chiplet security from both the architecture and systems perspective. Those discussions reinforced something we've been seeing more broadly across the industry.

[Company Representative] (eMemory): eMemory and PUFsecurity, your trusted partner in providing non-volatile memory and PUF-based security intellectual properties.

[Video Narrator]: eMemory and PUFsecurity, your trusted partner in providing non-volatile memory and PUF-based security intellectual properties.

Speaker #2: These security technologies are becoming increasingly important as AI servers and data center infrastructure become more distributed and more interconnected. So, rather than covering everything around Calipra, I'd like to use the next few minutes to just focus on the underlying reason why this movement is gaining momentum.

Felix Hsu: Okay. Hi, everyone. What you have just viewed in this video is one part of a much bigger movement happening around Caliptra and the hardware rooted security. Before I begin this talk, I like to briefly mention some related activities we have just been involved in. Earlier this week in Taipei, we participated in a Caliptra workshop focused on implementation and on bringing the specification closer to production silicon. We also presented at 2 sessions at OCP APAC, where we discussed Caliptra and chiplet security from both the architecture and systems perspective. Those discussions reinforce something we have been seeing more broadly across the industry. These security technologies are becoming increasingly important as AI servers and data center infrastructure become more distributed and more interconnected.

Speaker #2: And instead of starting from security, I would actually like to start from how AI compute itself has been changing. AI systems are increasingly being built across multiple dies, accelerators, memory devices, and high-speed fabrics.

Speaker #2: And as more components are connected together, more of these components and their interfaces also need to be individually trusted and protected. And that's the story I'd like to present today.

Speaker #2: So, AI compute can no longer scale in a single die. When we talk about scaling AI compute, simply making chips larger is no longer enough.

Speaker #2: In the past, if we wanted more performance, we could build a larger die or use more advanced process nodes. But there are practical limits that we will approach, including die size, yield, cost, power, memory bandwidth, and I/O—all of which start to become constraints.

Felix Hsu: Rather than covering everything around Caliptra, I would like to use the next few minutes to just focus on the underlying reason why this movement is gaining momentum. Instead of starting from security, I would actually like to start from how AI compute itself has been changing. AI systems are increasingly being built across multiple dies, accelerators, memory devices, and high-speed fabrics. As more components are connected together, more of these components and their interfaces also need to be individually trusted and protected. That is the story I like to present today. AI compute can no longer scale in a single die. When we talk about scaling AI compute, simply making chips larger is no longer enough. In the past, we wanted more performance, we could build a larger die or use more advanced process nodes.

Speaker #2: So, increasingly, the industry is scaling by composition. There are two important architectural trends shown on this slide, and I want to separate them because they're related, but they're not the same thing.

Speaker #2: The first one is chiplets, which operate inside the package. Instead of building one enormous, monolithic SoC containing every function, we can divide the system into multiple sized, specialized dies.

Speaker #2: For example, we have a CPU compute die, GPU or MPU accelerator dies, management dies, a security die, or dies with other specialized functions.

Speaker #2: Importantly, these dies do not necessarily need to use the same process technology. The compute portion may benefit the most from advanced nodes, but other dies—such as those for management or security functions—may be more economical on less advanced nodes.

Felix Hsu: These are practical limits that we will approach, including die size, yield, cost, power, memory bandwidth, and IO all start to become constraints. Increasingly, the industry is scaling by composition. There are 2 important architectural trends shown on this slide, and I want to separate them because they are related, but they are not the same thing. The first one is chiplets, which operate inside the package. Instead of building one enormous monolithic SoC containing every function, we can divide the system into multiple specialized dies. For example, we have a CPU compute die, a GPU or NPU accelerator dies. We have management dies, a security die, or dies with other specialized functions. Importantly, these dies do not necessarily need to use the same process technology.

Speaker #2: So chiplets give us modularity, reuse, better economics, and another way to continue scaling the processor package. But even if we scale inside the package, we eventually hit another boundary.

Speaker #2: The processor socket and the server itself. And that's where Compute Express Links, or CXLs, come in. CXL operates at the system level, outside the processor package.

Speaker #2: Here you can see the servers on the left connected through a CXL fabric switch to a CXL Type 3 memory shelf on the right.

Speaker #2: Instead of memory belonging permanently to one CPU socket, CXL allows memory and other devices to participate in a coherent fabric. That means memory can increasingly be expanded, pooled, shared, and dynamically assigned across systems.

Felix Hsu: The compute portion may benefit the most from advanced nodes, but other dies, such as management, security functions, it may be more economical on less advanced nodes. Chiplets give us modularity, reuse, better economics, and another way to continue scaling the processor package. But even if we scale inside the package, we eventually hit another boundary, the processor socket and the server itself. That is where Compute Express Link, CXLs, come in. CXL operates at the system level outside the processor package. Here you can see the servers on the left connected through a CXL fabric switch to a CXL type 3 memory shelf on the right. Instead of memory belonging permanently to one CPU socket, CXL allows memory and other devices to participate in a coherent fabric. That means memory can increasingly be expanded, pooled, shared, and dynamically assigned across systems.

Speaker #2: So you could think of it as chiplets scale the processor by combining multiple specialized dies in one package, whereas CXL scales a system by connecting processors, accelerators, and memory across the server.

Speaker #2: That's a very important shift for the AI infrastructure. We are moving away from scaling only through one large piece of silicon and toward scaling by connecting many specialized resources together.

Speaker #2: But this composition has security consequences. Every new die link, device controller, and firmware layer creates another place where trust has to be established. And that brings us to the next architecture.

Speaker #2: The same composability that gives us flexibility and scalability also creates a much larger attack surface. Let me separate this again into two architectural domains.

Felix Hsu: You could think of it as chiplets scale the processor by combining multiple specialized dies in one package, whereas CXL scales the system by connecting processors, accelerators, and memory across the server. That is a very important shift for the AI infrastructure. We are moving away from scaling only through one large piece of silicon and towards scaling by connecting many specialized resources together. But this composition has security consequences. Every new die link, device controller, and firmware layer creates another place where trust has to be established. That brings us to the next architecture. The same composability that gives us flexibility and scalability also creates a much larger attack surface. Let me separate this again to two architectural domains.

Speaker #2: On the CXL side, we now have multiple endpoints joining a fabric, with high-value memory moving across links. So it's just about controlling and deciding how pooled resources are located.

Speaker #2: That introduces several different classes of risk. For example, if the firmware of a highly privileged infrastructure component, such as the CXL switch, is compromised, the impact radius can be very large.

Speaker #2: Because now the attacker is not targeting only one endpoint. They may potentially manipulate connectivity, resource allocation, or the behavior of multiple devices in the fabric.

Speaker #2: Another important area is DMA, direct memory access, which is shown in panel five. DMA itself is a legitimate and very powerful capability. The security problem is when a compromised or malicious device receives more access than it should have.

Felix Hsu: On the CXL side, we now have multiple endpoints joining a fabric, high-value memory moving across links, switches controlling connectivity, and a fabric manager potentially deciding how pooled resources are located. That introduces several different classes of risk. For example, if the firmware of a highly privileged infrastructure component, such as a CXL switch, is compromised, the impact radius can be very large because now the attacker is not targeting only one endpoint. They may potentially manipulate connectivity, resource allocation, or the behavior of multiple devices in the fabric. Another important area is DMA, Direct Memory Access, which is in panel 5. DMA itself is a legitimate and very powerful capability. The security problem is when a compromised or malicious device receives more access than it should have.

Speaker #2: If those permissions are not properly controlled, a device can potentially read or modify memory outside its authorized region. And in an environment where memory is pooled or shared, the consequences can become especially serious.

Speaker #2: Data leakage across workloads or even between tenants. Inside the chiplet package, the security problem looks different, but the principle is the same. Now we're dealing with multiple active dies communicating across die-to-die interfaces.

Speaker #2: A counterfeit or rogue chiplet could introduce an unauthorized component into the package or supply chain, thus potentially affecting silicon identity and supply chain trust.

Speaker #2: A malicious or unauthorized die can introduce hardware Trojans, manipulate transactions, leak data, or undermine the assumptions of the entire package. Recent chiplet security research specifically highlights malicious chiplets and hardware Trojans as major system-level threats.

Felix Hsu: If those permissions are not properly controlled, a device can potentially read or modify memory outside its authorized region, and in an environment where memory is pooled or shared, the consequence can become especially serious. Data leakage across workloads or even between tenants. Inside the chiplet package, the security problem looks different, but the principle is the same. Now we are dealing with multiple active dies communicating across die-to-die interfaces. A counterfeit or rogue chiplet can introduce an unauthorized component into the package or supply chain, severely affecting silicon identity and supply chain trust. A malicious or unauthorized die can introduce hardware trojans, manipulate transactions, leak data, or undermine the assumptions of the entire package. Recent chiplet security research specifically highlights malicious chiplets and hardware trojans as major system-level threats. Chiplets may also be attacked leading to key extraction and secret leakage, as shown in panel 4.

Speaker #2: Chiplets may also be attacked, leading to key extraction and secret leakage, as shown in panel four. If the root key or device secret is compromised, the consequence is much greater than simply losing one piece of data.

Speaker #2: An attacker may be able to impersonate the device or undermine the trust relationship built on top of that identity. So, I don't expect everyone to remember these attacks, but the key point is the trust boundary has expanded.

Speaker #2: Security can no longer exist only at the board level or only around the CPU. What we increasingly need to know is this: is the device or die authentic?

Speaker #2: Did it boot trusted firmware? Can I trust the keys inside it? Can I trust the communication between components? And can I prove that trust to the rest of the system?

Speaker #2: This is why the root of trust increasingly has to move closer to individual silicon itself. And this is where architectures such as Calyptra become very important.

Speaker #2: Calyptra gives us an open-source security architecture for establishing a hardware root of trust in modern devices. But there is an important distinction here between a digital root of trust architecture and the physical security foundation underneath it.

Felix Hsu: If the root key or device secret is compromised, the consequence is much bigger than simply losing one piece of data. An attacker may be able to impersonate the device or undermine the trust relationship built on top of that identity. I do not expect everyone to remember these attacks, but the key point is the trust boundary has expanded. Security can no longer exist only at the board level or only around the CPU. What we increasingly need to know is this: Is the device or die authentic? Did it boot trusted firmware? Can I trust the keys inside it? Can I trust the communication between components? Can I prove that trust to the rest of the system? This is why the root of trust increasingly has to move closer to individual silicon itself. This is where architectures such as Caliptra become very important.

Speaker #2: On the left, we have the Calyptra subsystem. Within the subsystem, the Calyptra core includes things such as a RISC-V processor, cryptographic functions, SRAM and ROM, mailbox, and the firmware responsible for implementing the security architecture.

Speaker #2: But ultimately, all that digital security logic has to anchor itself to something physical in silicon. So, as you see on the right, there are three fundamental requirements for this physical hardware anchor.

Speaker #2: These are also called foundational security primitives. First, a Unique Device Secret (UDS). We need something unique to the individual piece of silicon that can establish device identity and support key derivation.

Felix Hsu: Caliptra gives us an open source security architecture for establishing a hardware root of trust in modern devices. There is an important distinction here between a digital root of trust architecture and the physical security foundation underneath it. On the left, we have the Caliptra subsystem. Within the subsystem, the Caliptra core includes things such as the RISC-V processor, cryptographic functions, SRAM and ROM, mailbox and the firmware responsible for implementing the security architecture. Ultimately, all that digital security logic has to anchor itself to something physical in silicon. As you see on the right, there are three fundamental requirements for this physical hardware anchor, which we also call foundational security primitives. First, a unique device secret, UDS. We need something unique to the individual piece of silicon that can establish device identity and support key derivation. Second is a secure non-volatile storage.

Speaker #2: Second is a secure, non-volatile storage. Security-critical information such as life cycle state, configuration, seeds, or other protective values needs to survive power cycles and remain protected from unauthorized modifications.

Speaker #2: Third, a trustworthy entropy. All of the cryptography above this layer ultimately depends on good randomness. If the entropy is weak or predictable, the keys derived from it can also become weak or predictable.

Speaker #2: So, Calyptra provides an architecture and framework for the root of trust, but the foundation of that trust still has to terminate in physical properties and protected state inside the silicon.

Speaker #2: That is the physical security anchor. And those requirements actually match very naturally to technologies of hardware security that eMemory and PUFsecurity have already been developing for many years.

Speaker #2: This is where we move from the architecture to the actual silicon implementation. For the UDS, unique device secret, we will use NeoPUF. Rather than simply programming the root or root secret into conventional memory, NeoPUF derives device-unique information from the intrinsic physical characteristics of each individual piece of silicon.

Felix Hsu: Security critical information such as life cycle state, configuration, seeds, or other protected value needs to survive power cycles and remain protected from unauthorized modifications. Third, a trustworthy entropy. All of the cryptography above this layer ultimately depends on good randomness. If the entropy is weak or predictable, the keys derived from it can also become weak or predictable. Caliptra provides the architecture and framework for the root of trust, but the foundation of that trust still has to terminate in physical properties and protected state inside the silicon. That is the physical security anchor, and those requirements actually match very naturally to technologies of hardware security that eMemory and PUFsecurity have already been developing for many years. This is where we move from the architecture to the actual silicon implementation. For the UDS unique device secret, we use NeoPUF.

Speaker #2: That gives us a silicon-bound foundation for device identity and key derivation. For secure non-volatile storage, we have NeoFuse, our OTP. NeoFuse provides protected storage for security-critical information such as life cycle values, seeds, configuration, or other persistent security assets.

Speaker #2: For entropy, we provide the physical noise source supporting the TRNG, or true random number generator, giving the cryptographic system the randomness required for secure key generation and operation.

Speaker #2: On top of that, we're extending this foundation to the post-quantum era through Pub PQC. Importantly, we view these IPs as different pieces of the same security foundation.

Speaker #2: Identity, protected storage, trusted entropy, and cryptographic agility—together, these are the types of silicon-level primitives that a framework such as Calyptra ultimately needs in order to establish trust.

Felix Hsu: Rather than simply programming the root secret into conventional memory, NeoPUF derives device unique information from the intrinsic physical characteristic of each individual piece of silicon. That gives us a silicon-bound foundation for device identity and key derivation. For secure non-volatile storage, we have NeoFuse, our OTP. NeoFuse provides protected storage for security critical information such as life cycle values, seeds, configuration, or other persistent security assets. For entropy, we provide the physical noise source supporting the TRNG, true random number generator, giving the cryptographic system the randomness required for secure key generation and operation. On top of that, we are extending this foundation to the post-quantum era through PUF-PQC. Importantly, we view these IPs as different pieces of the same security foundation. Identity, protected storage, trusted entropy, and cryptographic agility.

Speaker #2: By having the primitives, that's only part of the problem. For our customers, the next question is how to integrate all these pieces into a complete, validated security system.

Speaker #2: This is where our strategy goes beyond selling individual IPs. We start with Pub RT as the hardware anchor IP. Around that, we add post-quantum security capabilities.

Speaker #2: And importantly, we have experience in integrating with Calyptra. Because from a chip designer’s perspective, the difficult question isn’t just: can I license a Pub?

Speaker #2: The real question is: how do I turn the specification into silicon that actually works? How do I connect the hardware primitives to the Calyptra root-of-trust architecture?

Speaker #2: How do I validate the interfaces? How do I shorten integration and verification time? That is why we're moving toward a validated, subsystem approach.

Felix Hsu: Together, these are the types of silicon level primitives that a framework such as Caliptra ultimately needs in order to establish trust. But having the primitives is only part of the problem. For our customers, the next question is how to integrate all these pieces into a complete validated security system. This is where our strategy goes beyond selling individual IPs. We start with PUFrt as the hardware anchor IP. Around that, we add post-quantum security capabilities. Importantly, we have experience in integrating with Caliptra because from a chip designer's perspective, the difficult question isn't just can I license a PUF? The real question is: How do I turn the specification into silicon that actually works? How do I connect the hardware primitives to the Caliptra root of trust architecture? How do I validate the interfaces? How do I shorten integration and verification time?

Speaker #2: And once that foundation is available, the same security architecture can scale across accelerator boards, AI servers, rack-scale systems, and eventually data center infrastructure.

Speaker #2: What starts as a very small security block in silicon can ultimately become the trust foundation for a much larger computing platform. And that brings us back to where we started.

Speaker #2: This increasingly composed AI system—so the trust must extend across every layer of the AI infrastructure. AI infrastructure is becoming distributed across multiple layers.

Speaker #2: At the system level, CXL can protect data moving across a link through mechanisms for confidentiality, integrity, and replay protection. But protecting the link does not automatically mean that the device on the other end should be trusted.

Speaker #2: That end still needs integrity, trusted firmware, secure boot, and also needs to be able to prove its state through attestation. The multi-die package follows similar principles.

Felix Hsu: That is why we're moving towards a validated subsystem approach. Once that foundation is available, the same security architecture can scale across accelerator boards, AI servers, rack scale systems, and eventually data center infrastructure. What starts as a very small security block in silicon can ultimately become the trust foundation for a much larger computing platform. That brings us back to where we started, this increasingly composed AI system. The trust must extend across every layer of the AI infrastructure. AI infrastructure is becoming distributed across multiple layers. At the system level, CXL can protect data moving across a link through mechanisms for confidentiality, integrity, and replay protection. Protecting the link does not automatically mean that the device on the other end should be trusted.

Speaker #2: Each active chiplet may have its own identity keys, firmware lifecycle state, and relationships with other dies in the package. So trust increasingly has to extend from the data center, to the rack, to the server, to the device, and to the package.

Speaker #2: Ultimately, it's down to the individual silicon. And that's why our foundational IP is so important. NeoPUF, NeoFuse, and our entropy source and PUF PQC may be physically small pieces of an overall AI system.

Speaker #2: But they provide some of its most fundamental security properties: a unique identity, protected state, trustworthy entropy, and cryptographic trust. Pub RT brings those capabilities together as a physical hardware anchor.

Speaker #2: While Calyptra provides the architecture that builds a root of trust on top of them, as AI compute becomes more composable and more distributed, we believe security has to follow exactly the same direction.

Felix Hsu: The end still needs an integrity trusted firmware secure boot, and also needs to be able to prove its state through attestation. The multi-die package follows similar principles. Each active chiplet may have its own identity, keys, firmware, life cycle state, and relationships with the other dies in the package. Trust increasingly has to extend from the data center to the rack, to the server, to the device, to the package, ultimately down to the individual silicon. That's why our foundational IP is so important. NeoPUF, NeoFuse, and our entropy source and PUF-PQC may be physically small pieces of an overall AI system, but they provide some of its most fundamental security properties: a unique identity, protected state, trustworthy entropy, and cryptographic trust.

Speaker #2: Trust must also become distributed, but it must remain anchored in silicon. That is the role we see for eMemory and PUFsecurity's foundational IP—providing the silicon trust foundation that enables security, that enables secure chiplets, secure CXL-connected devices, and ultimately the next generation of AI infrastructure.

Speaker #2: Thank you for your time.

Speaker #1: This concludes our prepared statement. Next, we will enter the Q&A section. We will now begin the Q&A session. Please submit your questions in the input box on the webcast window.

Speaker #1: For our questions, we'll follow the format of answering the Chinese version first, followed by the English version. We will now collect the questions and begin our Q&A section.

Felix Hsu: PUFrt brings those capabilities together as a physical hardware anchor, while Caliptra provides the architecture that builds a root of trust on top of them. As AI compute becomes more composable and more distributed, we believe security has to follow exactly the same direction. Trust must also become distributed, but it must remain anchored in silicon. That is the role we see for eMemory and PUFsecurity's foundational IP, providing the silicon trust foundation that enables secure chiplets, secure CXL connected device, and ultimately the next generation of AI infrastructure. Thank you for your time.

Speaker #3: 这一题应该大家都很关心。这个上一季有说下半年权利金会明显成长。那现在对这个展望是不是更有信心呢？Joseph请回答。

Speaker #2: 是的，我们还是持续的很有信心。因为我们目前已经看到蛮多像这个新制程的产品，包含是像ADAS AI的加速器，那SSD网通以及其他的新产品都已经开始进入量产。那加上晶圆代工的这个报价其实持续的在涨价，那我们认为下半年的成长动能一定是会加速。谢谢。

Li Ching-Xiang: This concludes our prepared statement. Next, we will enter the Q&A section. We will now begin the Q&A section. Please submit your questions in the input box on the webcast window. All of our questions will follow the format of answering the Chinese version first, followed by the English version. We will now collect the questions and begin our Q&A section.

Operator: This concludes our prepared statement. Next, we will enter the Q&A section. We will now begin the Q&A section. Please submit your questions in the input box on the webcast window. All of our questions will follow the format of answering the Chinese version first, followed by the English version. We will now collect the questions and begin our Q&A section.

Speaker #1: Our first question is: In the previous earnings call, the company mentioned that royalty revenue was expected to grow in the second half of this year.

Speaker #1: Are you now confident in that outlook? Joseph, please.

Speaker #2: Yes, our outlook remains the same. We have already seen several events and known products, including ADAS, AI accelerators, SSDs, networking, and other new products. They have begun to enter the mass production phase.

Qing-Xiang Hsu: 这一题应该大家都很关心。上季有说下半年权利金会明显成长，那现在对这个展望是不是更有信心呢？Joseph，请回答。

Li-Jeng Chen: [Foreign language]

Speaker #2: In addition, with foundry wafer prices gradually increasing, we believe the growth momentum will continue to accelerate. Thank you.

Joseph Hsia: 是的，我们还是持续地很有信心，因为我们目前已经看到蛮多像新制程的产品，包含像是ADAS、AI的加速器、SSD、网通以及其他的新产品，都已经开始进入量产。加上晶圆代工的报价其实持续地在涨价，我们认为下半年的成长动能已经是会加速。谢谢。

Joseph Hsia: [Foreign language]

Speaker #1: 公司提到已经有不少3奈米客户产品完全授权。那请问今年下半年是否有机会开始贡献权利金收入？Joseph。

Speaker #2: 我们目前已经看到客户这边在3奈米已经有几百片的一个量产。那这表示他们的产品已经成功了。那明年这个量以及权利金贡献会更加的明显。

Speaker #1: The company has mentioned that the number of customers using 3-nanometer technology have already been licensed, because some of them will be contributing to royalty revenue in the second half of this year.

Li Ching-Xiang: Our first question is: In the previous earnings call, the company mentioned that royalty revenue was expected to grow in the H2 of this year. Are you now confident in that outlook? Joseph, please.

Li-Jeng Chen: Our first question is: In the previous earnings call, the company mentioned that royalty revenue was expected to grow in the H2 of this year. Are you now confident in that outlook? Joseph, please.

Speaker #1: Joseph, please.

Joseph Hsia: Yes, our outlook remains the same. We have already seen several advanced node products, including ADAS, AI accelerators, SSDs, networking, and other new products. They begin to enter the mass production phase. In addition, with foundry wafer prices gradually increasing, I would believe the growth momentum will continue to accelerate. Thank you.

Speaker #2: Yes, we have already seen customers moving into mass production with several hundred wafers. This indicates that their products have been successfully validated, and the contribution is expected to become much more meaningful next year.

Speaker #1: AI资料中心带动高压电源架构的升级。比如说400伏、800伏的HCDC更新一代的Power IC。请问公司是否已经开始受惠？Joseph。

Qing-Xiang Hsu: 公司提到已经有不少3奈米客户产品完全授权。那请问今年下半年是否有机会开始贡献权利金收入？Joseph。

Li-Jeng Chen: [Foreign language]

Speaker #2: 那AI中心的这个电源架构的一个升级对力量而言，那我们的IP是会被整合到，例如说像电源管理、电源控制以及相关的类比混合讯号的晶片中。那公司就会因此而受惠。那随着电压以及功率密度的提升，Power IC对于这个精准的校正参数的补偿，那装置的一个识别可靠度的一个管理以及资安功能的要求都会大幅度的增加。那力量也可以提供相关的IP，例如说OTP可以用在这个类比参数的一个校正Trimming跟这个电源控制设定以及产品的一个识别。那MTP的话可以用在这个需要多次更新的系统参数以及组态的一个设定。那目前客户已经陆续进入这个量产的阶段。那未来这个权利金的贡献会持续的扩大。

Joseph Hsia: 我們目前已經看到客戶這邊在3奈米已經有幾百片的一個量產，這表示他們的產品已經成功了。那明年這個量以及權利金貢獻會更加的明顯。

Joseph Hsia: [Foreign language]

Li Ching-Xiang: The company has mentioned that the number of customers using 3-nanometer technology have already been licensed or taped out. Could some of them be contributing to royalty revenue in the H2 of this year? Joseph, please.

Li-Jeng Chen: The company has mentioned that the number of customers using 3-nanometer technology have already been licensed or taped out. Could some of them be contributing to royalty revenue in the H2 of this year? Joseph, please.

Joseph Hsia: Yes, we have already seen customers moving into mass production with several hundred wafers, and this indicates that their products have been successfully validated and the contribution expected to become much more meaningful next year.

Qing-Xiang Hsu: AI 资料中心带动高压电源架构的升级，譬如说 400 伏、800 伏的 HVDC 跟新一代的 Power IC。请问公司是否已经开始受惠？Joseph。

Li-Jeng Chen: [Foreign language]

Speaker #1: AI data centers are driving upgrades in high-voltage power architecture, such as 400 volts and 800 volts HVDC, as well as next-generation power ICs. Has the company started to benefit from this trend? Which product lines would be the main beneficiaries?

Joseph Hsia: AI中心的电源架构的一个升级，对力旺而言，我们的IP是会被整合到，例如说像电源管理、电源控制以及相关的类比混合讯号的晶片中，公司就会因此而受惠。随着电压以及功率密度的提升，Power IC对于精准的校正、参数的补偿、装置的识别、可靠度的管理，以及治安功能的要求都会大幅度地增加。力旺也可以提供相关的IP，例如说OTP，可以用在这个类比参数的一个校正trimming，跟电源控制设定以及产品的一个识别。那MTP的话，可以用在需要多次更新的系统参数，以及组它的一个设定。目前客户已经陆续进入量产的阶段，未来权利金的贡献会持续地扩大。

Joseph Hsia: [Foreign language]

Speaker #1: Joseph, please.

Speaker #2: Yes. For eMemory, the upgrade of AI data center power architectures creates opportunities through power management, power control, and related analog mixed-signal chips.

Speaker #2: And as voltage levels and power density continue to increase for power ICs, they require much more precise calibration, parameter compensation, device identification, and also reliability management.

Speaker #2: And of course, the security functions. The related IPs we can provide mainly include OTP, which is primarily used for analog parameter calibration, trimming, power control setting, and, of course, product identification.

Li Ching-Xiang: AI data centers are driving upgrades in high voltage power architecture such as 400 volts and 800 volts HVDC as well as next generation Power ICs. Has the company started to benefit from this trend? Which product lines will be the main beneficiaries? Joseph, please.

Li-Jeng Chen: AI data centers are driving upgrades in high voltage power architecture such as 400 volts and 800 volts HVDC as well as next generation Power ICs. Has the company started to benefit from this trend? Which product lines will be the main beneficiaries? Joseph, please.

Speaker #2: And for NTP, they can support system parameters and configuration settings that need to be updated multiple times. Our customers have already started to move into mass production, and we expect the contribution to expand further in the coming quarters.

Speaker #2: Thank you.

Speaker #1: 可不可以分享权利金当中有多少来自先进制程？未来几年是否预期这个比例会提升？Joseph。

Joseph Hsia: Yes. For eMemory, the upgrade of AI data center power architectures creates opportunities through our

Joseph Hsia: Yes. For eMemory, the upgrade of AI data center power architectures creates opportunities through our IP being integrated into power management, power control and related analog mixed signal chips. As voltage levels and power density continue to increase, for Power ICs, they require much more precise calibration, parameter compensation, device identification and also reliability management, and of course, the security functions. The related IPs we can provide mainly includes OTP, which is mainly being used for analog parameter calibration, trimming, power control setting, and of course product identification. For MTP, they can support system parameters and configuration settings that need to be updated multiple times. Our customers have already started to move into mass production, and we expect the contribution to expand further in the coming quarters. Thank you.

Speaker #2: 我们以授权金的NIE占比来看，目前这个先进制程的部分已经超过了50%。那这是因为我们单一的授权金的单价其实是数倍高于平均的这个license的单价。那未来权利金的贡献也可望这个明显的提升。

Joseph Hsia: IP being integrated into power management, power control and related analog mixed signal chips. As voltage levels and power density continue to increase, for Power ICs, they require much more precise calibration, parameter compensation, device identification and also reliability management, and of course, the security functions. The related IPs we can provide mainly includes OTP, which is mainly being used for analog parameter calibration, trimming, power control setting, and of course product identification. For MTP, they can support system parameters and configuration settings that need to be updated multiple times. Our customers have already started to move into mass production, and we expect the contribution to expand further in the coming quarters. Thank you.

Speaker #1: Could you share how investors should think about the contribution from advanced known products to your royalty revenue today, and how you expect that mix to evolve over the next few years?

Speaker #1: Joseph, please.

Speaker #2: In terms of contribution, advanced-known projects already accounted for more than 50%. And this is mainly because the licensing fee for a single advanced-known project is several times higher than the average project licensing price.

Speaker #2: So, looking ahead, as these projects gradually move into mass production, we expect the royalty contribution to increase meaningfully as well.

Li Ching-Xiang: 可不可以分享權利金當中有多少來自先進制程？未來幾年是否預期這個比例會提升？Joseph。

Li-Jeng Chen: [Foreign language]

Speaker #1: Q2权利金收入相较Q1下滑，可否说明主要原因？谢谢。

Joseph Hsia: 我們以授權金的NRE占比來看呢，目前先進制程部分已經超過50%。這是因為我們單一的授權金的單價其實是數倍高於平均的license的單價，未來權利金的貢獻也可望有明顯的提升。

Joseph Hsia: [Foreign language]

Speaker #2: 我们第二季的权利金主要是对应代工厂在第一季的出货。那第一季的部分普遍成熟制程的产能利用率仍属于一个低档。那加上部分客户有这个季节性的库存调整。那在第二季的部分，我们已经看到这个整体代工厂的产能利用率来提升。那加上我们新应用已经开始持续进入这个量产阶段。那后续还有这个代工价的一个上涨来带动。那我们预期Q3以及下半年的权利金的成长可以可以加速。谢谢。

Li Ching-Xiang: Could you share how investors should think about the contribution from advanced node product to your royalty revenue today? How you expect that mix to evolve over the next few years? Joseph, please.

Li-Jeng Chen: Could you share how investors should think about the contribution from advanced node product to your royalty revenue today? How you expect that mix to evolve over the next few years? Joseph, please.

Joseph Hsia: In terms of licensing NRE contribution, advanced node projects already accounted for more than 50%. This is mainly because the licensing fee for a single advanced node project is several times higher than average project licensing price. Looking ahead, as these projects gradually move into mass production, we expect the royalty contribution to increase meaningfully as well.

Speaker #1: Royalty revenue declined in the second quarter compared with the first quarter. Could you please help explain the main reasons behind this decrease? Joseph, please.

Speaker #2: So for our Q2 royalty revenue, it corresponds to the foundry production in the first quarter. And overall, mature foundry capacity utilization remained at a relatively low level in Q1.

Speaker #2: And some customers are also going through we're also going through seasonal inventory adjustment. And starting from the second quarter, overall foundry capacity utilization began to improve while new applications started to enter mass production.

Li Ching-Xiang: Q2 权利金收入相较 Q1 下滑，可否说明主要原因？谢谢。

Li-Jeng Chen: [Foreign language]

Joseph Hsia: 我们第二季的权利金主要是对应代工厂在第一季的出货。第一季的部分，普遍成熟制程的产能利用率仍属于一个低档，加上部分客户有季节性的库存调整。在第二季的部分，我们已经看到整体代工厂的产能利用率提升，加上我们新应用已经开始持续进入量产阶段，后续还有代工价的一个上涨来带动。我们预期 Q3 以及下半年的权利金的成长可以加速。谢谢。

Joseph Hsia: [Foreign language]

Speaker #2: And together with the subsequent impact from foundry price increases, we believe royalty revenue is expected to accelerate in the second half of 2026. Thank you.

Speaker #1: 随着2奈米跟Go around制程逐步进入量产，请问Neopath跟Neofuse在先进制程上竞争优势是否进一步提升？那目前客户导入的速度是否也比过去来得快？Charles。

Speaker #2: 随着制程进入2奈米跟这个GAA晶片设计跟这个制造成本成本持续提高，客户面对面积功耗可靠度以及符合安全标准的要求也更加严格。在这个中情况下，那个Neofuse跟那个Neopath base的security IP的竞争优势会更加明明显。Neopath可以在晶片底层建立每颗晶片独有的硬体身份，支援金钥产生跟这个硬体信任跟那Neofuse则可以提供这可靠的OTP的储存用于晶片的设计设定，还有这个call patch，还有SRAM repair等等。这些功能都能与先进制程整合，符合先进制程SOC对安全性可靠性跟这高良率的需求，与以往相比我们确实看到更多的客户更早开始讨论这个安全的架构，然后参与评估的这个晶片应用数量也增也都在增，也都在增加。

Li Ching-Xiang: Royalty revenue declined in Q2 compared with Q1. Could you please help explain the main reasons behind the decrease? Joseph, please.

Li-Jeng Chen: Royalty revenue declined in Q2 compared with Q1. Could you please help explain the main reasons behind the decrease? Joseph, please.

Joseph Hsia: Our Q2 royalty revenue corresponds to the foundry production in Q1. Overall, mature foundry capacity utilization remained at a relatively low level in Q1. Some customers were also going through seasonal inventory adjustment. Starting from Q2, overall foundry capacity utilization began to improve, while new applications started to enter mass production. Together with subsequent impact from foundry price increases, we believe royalty revenue is expected to accelerate in H2 2026. Thank you.

Li Ching-Xiang: 随着2奈米跟Gate-all-around制程逐步进入量产，请问NeoPUF跟NeoFuse在先进制程上的竞争优势是否进一步提升？目前客户导入的速度是否也比过去来得快？Charles。

Li-Jeng Chen: [Foreign language]

Charles Hsu: 随着制程进入2奈米跟GAA，晶片设计跟制造成本持续提高，客户面对面积、功耗、可靠度，以及符合安全标准的要求也更加严格。在这个情况下，NeoFuse跟NeoPUF based security IP的竞争优势会更加明显。NeoPUF可以在晶片底层建立每颗晶片独有的硬体身份，支援金钥产生跟硬体信任根。那NeoFuse则可以提供可靠的OTP的储存，用于晶片的设计设定，还有code patching，还有SRAM repair等等。这些功能都能与先进制程整合，符合先进制程SoC对安全性、可靠性跟高良率的需求。与以往相比，我们确实看到更多的客户更早开始讨论安全的架构，参与评估的晶片应用数量也都在增加。

Charles Hsu: [Foreign language]

Speaker #1: First, as two-nanometer and GAA processes gradually move toward mass production, the competitive advantage of NeoPUF and NeoFuse in advanced nodes becomes even stronger.

Speaker #1: Are customers also adopting this technology faster than they did last year? Charles, please.

Speaker #2: As the process technologies move to 2-nanometer and GAA chip design and the manufacturing costs continue to rise, customers are placing even stricter requirements on area, power consumption, reliability, and compliance with security standards in this environment.

Speaker #2: The competitive advantage of Neofuse and the Neopath based security IP become even more apparent. Neopath can create an unique hardware identity at the silicon level for each chip, supporting key generation and hardware rule of the trust and the Neofuse provides reliable OTP storage and for the chip configuration call patch call patching and SRAM repair and other functions.

Li Ching-Xiang: As 2 nanometer and GAA processes gradually move toward mass production, have the competitive advantage of NeoPUF and NeoFuse in advanced nodes become even stronger? Are customers also adopting this technology faster than they did last year? Charles, please.

Li-Jeng Chen: As 2 nanometer and GAA processes gradually move toward mass production, have the competitive advantage of NeoPUF and NeoFuse in advanced nodes become even stronger? Are customers also adopting this technology faster than they did last year? Charles, please.

Speaker #2: This capability can be integrated with advanced logic processes and support the security, reliability, and the high heat and high year requirements of advanced SoC, compared with the past.

Charles Hsu: As process technologies move from 2 nanometer and GAA, chip design and manufacturing costs continue to rise, and customers are placing even stricter requirements on area, power consumption, reliability, and compliance with security standards. In this environment, the competitive advantage of NeoFuse and the NeoPUF-based security IP become even more apparent. NeoPUF can create a unique hardware identity at the silicon level for each chip, supporting key generation and the hardware root of trust. NeoFuse provides reliable OTP storage for chip configuration, code patching, and SRAM repair, and other functions. This capability can be integrated with advanced logic processes and support the security, reliability, and high requirement of advanced SoC. Compared with the past, we are indeed seeing customers start discussion on the security architectures earlier, and the number of applications under evaluation is also increasing.

Speaker #2: We are indeed seeing customers start discussions on the security architectures earlier, and the number of applications under evaluation is also increasing.

Speaker #1: 越来越多Path客户晶片应用进入量产，那公司如何看未来授权金跟权利金比例结构的变化？Charles。

Speaker #2: 近年来的这个营运表现来看，我们的Path在授权金跟权利金方面的成长非常强劲，因为这个大部分都是这个advanced know用在security上面的需求。那我们觉得这个当AI开始，现在AI才那个的成长非常快，那我们认为这个security的趋势才刚开始。

Speaker #1: As more Path licenses enter mass production, how do you expect the mix between licensing and royalty revenue to evolve? Charles, please.

Speaker #2: Based on our operating performance in recent year, our Path business has delivered very strong growth in both licensing and the royalty revenue. We believe this trend is be still in the early stage because the due to the AI booming very fast and now there are a lot of demands to secure the AI's applications.

Li Ching-Xiang: 越来越多PUF客户晶片应用进入量产，那公司如何看待未来授权金跟权利金比例结构的变化？Charles。

Li-Jeng Chen: [Foreign language]

Speaker #2: So I believe that we believe that the license fee and also royalty due to the application of Path will be increased.

Charles Hsu: 近年来的营运表现来看，我们的PUF在授权金跟权利金方面的成长非常强劲，因为大部分都是advanced node用在security上面的需求。那我们觉得当AI开始，现在AI的成长非常快，我们认为这个security的趋势才刚开始。

Charles Hsu: [Foreign language]

Speaker #3: 客户目前的进展如何？是否有机会在未来一年左右进入量产？

Speaker #2: Okay. 目前那个1T的flash与多家晶圆代工厂合作进行平台开发跟验证，各平台正依照开发计划持续进行，有机会在明年看到客户开始量产。

Li Ching-Xiang: As more PUF licenses enter mass production, how do you expect the mix between licensing and royalty revenue to evolve? Charles, please.

Li-Jeng Chen: As more PUF licenses enter mass production, how do you expect the mix between licensing and royalty revenue to evolve? Charles, please.

Speaker #1: Could you share the latest progress with customers for 1T flash? Is there a possibility that it could start moving into mass production within the next year or so?

Charles Hsu: Based on our operating performance in recent years, our PUF business has delivered very strong growth in both licensing and royalty revenue. We believe this trend is still in the early stage, because due to the AI booming very fast, there are a lot of demands to secure the AI's applications. So we believe that the licensee and also royalty, due to the application of PUF, will be increased.

Speaker #1: Charles, please.

Speaker #2: For 1T Flash, we are currently working with several foundry partners on platform development and verification. Each platform is progressing according to its development plan, and we may see customers begin mass production next year.

Speaker #1: 公司security IP过去已经导入了AI CPU BMC等应用，今年又进一步扩展到AI data center SD跟CSL，是否代表公司security IP正从AI server中的单一晶片逐步扩展到更多关键晶片？未来还有哪些应用具有较大的成长机会？Felix。

[Analyst]: 请问 1T Flash 客户目前的进展如何？是否有机会在未来一年左右进入量产？

Li-Jeng Chen: [Foreign language]

Charles Hsu: 目前1T Flash已与多家晶圆代工厂合作进行平台开发跟验证，各平台正依照开发计划持续进行，有机会在明年看到客户开始量产。

Charles Hsu: [Foreign language]

Speaker #2: 是的，随着AI server架构持续升级，我们看到硬体安全的需求已经从特定功能逐步延伸到运算储存资料传输及系统管理等不同环节。这也让公司的security IP有机会进入更多不同类型的晶片。从今年上半年的授权状况也可以看到这个趋势，相关的应用已涵盖CPU储存装置、光通讯IC还有FPGA及高速界面晶片等多个领域。这对我们来说呢，未来的成长不只是来自于更多客户，而是security IP在整个AI infrastructure中的应用范围持续扩大。同时我们在每颗晶片中能够提供的安全功能也在增加。随着hardware root of trust资料保护及系统及安全成为AI infrastructure的重要需求，我们认为security IP在AI与data center的市场会有非常大的成长空间。

Li Ching-Xiang: Could you share the latest progress with customers for 1T Flash? Is there a possibility that it could start moving into mass production within the next year or so? Charles, please.

Li-Jeng Chen: Could you share the latest progress with customers for 1T Flash? Is there a possibility that it could start moving into mass production within the next year or so? Charles, please.

Charles Hsu: For 1T Flash, we are currently working with several foundries partners on platform development and verification. Each platform is progressing according to its development plan. We may see customers begin mass production next year.

Li Ching-Xiang: 公司 Security IP 过去已经导入了 AI CPU、BMC 等应用，今年又进一步扩展到 AI data center SSD 跟 CXL，是否代表公司 Security IP 正从 AI server 中的单一晶片逐步扩展到更多关键晶片？未来还有哪些应用具有较大的成长机会？Felix。

Li-Jeng Chen: [Foreign language]

Speaker #1: Your security IP has expanded from AI CPUs and BMCs into AI data center SSDs and CXL-related applications. Does this suggest that eMemory security IP is moving from a single chip to multiple critical chips within an AI server?

Felix Hsu: 是的。随着AI server架构持续升级，我们看到硬体安全的需求已经从特定功能逐步延伸到运算、储存、资料传输及系统管理等不同环节，这让公司的Security IP有机会进入更多不同类型的晶片。从今年上半年的授权状况也可以看到这个趋势。相关的应用已涵盖CPU、储存装置、光通讯IC，还有FPGA及高速界面晶片等多个领域。这对我们来说，未来的成长不只是来自于更多客户，而是Security IP在整个AI infrastructure中的应用范围持续扩大。同时我们在每颗晶片中能够提供的安全功能也在增加。随着hardware root of trust、资料保护及系统级安全成为AI infrastructure的重要需求，我们认为Security IP在AI与data center的市场会有非常大的成长空间。

Felix Hsu: [Foreign language]

Speaker #1: Where do you see the next major opportunities? Felix, please.

Speaker #2: Yes. So as AI server architectures continue to evolve, we are seeing hardware security requirements extend across different parts of the system. These include computing, storage, connectivity, and system management.

Speaker #2: This gives us—this gives our security IP opportunities to address a much broader range of chips. Our licensing activity in the first half also reflects this trend.

Speaker #2: With applications across CPUs, storage devices, optical communication ICs, FPGAs, and high-speed interface chips, for us, the opportunity is not simply about adding more customers.

Speaker #2: It's about expanding our security footprint across AI infrastructure, while also providing more security functionality within each chip. As hardware root of trust, data protection, and system-level security become very important for AI infrastructure, we believe there is significant room for our security IP business to continue to grow across AI and data center applications.

Li Ching-Xiang: Your security IP has expanded from AI CPUs and BMCs into AI data center SSDs and CXL related applications. Does this suggest that eMemory's security IP is moving from a single chip to multiple critical chips within an AI server? Where do you see the next major opportunities? Felix, please.

Li-Jeng Chen: Your security IP has expanded from AI CPUs and BMCs into AI data center SSDs and CXL related applications. Does this suggest that eMemory's security IP is moving from a single chip to multiple critical chips within an AI server? Where do you see the next major opportunities? Felix, please.

Speaker #1: 你有发觉是否会与enterprise consumer SD使用的NAND flash？像SLC、TLC的竞争还是属于不同的市场？Charles。

Felix Hsu: Yes. As AI server architectures continue to evolve, we are seeing hardware security requirements extend across different parts of the system. These include computing, storage, connectivity, and system management. This gives our security IP opportunities to address a much broader range of chips. Our licensing activity in the H1 also reflects this trend with applications across CPUs, storage devices, optical communication ICs, FPGAs, and high-speed interface chips. For us, the opportunity is not simply about adding more customers. It is about expanding our security footprint across AI infrastructure while also providing more security functionality within each chip. As hardware root of trust, data protection, and system level security become very important for the AI infrastructure, we believe that there is significant room for our security IP business to continue to grow across AI and data center applications.

Speaker #2: 嗯，Salon flash的开发首先首要的我们是以这个no flash为主啊，所以跟这个NAND flash处于不同定位的记忆体技术。因为两者两者主要解决不同的应用需求，no flash呢注重于在于快速开机啊，低延迟啊，随机拖读读取，然后手这个execute execution in place啊，XIP及高可可靠度啊，如此因此呢主要用用于储存这个firmware及系统的韧体。但是那个NAND flash呢则是以这个高容量啊，低成本为主要的优势，适合大型资料的存存取等应用。然而那个现在那个待等到这个代工厂顺利量产这个stand alone no flash之后，那未来我们在才会再才会将这1T的flash呃拓拓展到这个NAND flash的架构。

Li Ching-Xiang: Neo Flash 是否会与 enterprise consumer SSD 使用的 NAND flash，像 SLC、TLC、QLC 竞争，还是属于不同的市场？Charles。

Li-Jeng Chen: [Foreign Language]

Speaker #1: Will Neo flash compete with NAND flash used in enterprise and consumer SSDs such as SLC, TLC, or QLC, or does it address a different market?

Charles Hsu: Standalone Flash的開發，首要的我們是以NOR Flash為主，所以跟NAND Flash處於不同定位的集體技術，因為兩者主要解決不同的應用需求。NOR Flash主重於在於快速開機、低延遲、隨機讀取，然後也有execution in place，XIP，及高可靠度。因此主要用於儲存firmware及系統的韌體。但是NAND Flash則是以高容量、低成本為主要的優勢，適合大型資料的存取等應用。然而現在等到代工廠順利量產standalone NOR Flash之後，未來我們才會將1T Flash拓展到NAND Flash的架構。

Charles Hsu: [Foreign language]

Speaker #1: And Charles, please.

Speaker #2: For stand-alone applications, our initial development focus is mainly on the no-flash, and no-flash and NAND flash are two different memory technologies with different positioning.

Speaker #2: And they mainly address different application needs. NoFlash is focused on fast boot, low latency random read and execution in place, or XIP, and high reliability.

Speaker #2: Therefore, it is mainly used to store firmware and system core. Flash, on the other hand, is mainly designed for high capacity and low cost, making it more suitable for large-scale data storage applications. However, after our foundry partner successfully moves standalone NOR flash into mass production, we may further extend 1T flash into NAND flash architecture in the future.

Li Ching-Xiang: Will NeoFlash compete with NAND Flash used in enterprise and consumer SSD such as SLC, TLC, or QLC? Or does it address a different market? Charles, please.

Li-Jeng Chen: Will NeoFlash compete with NAND Flash used in enterprise and consumer SSD such as SLC, TLC, or QLC? Or does it address a different market? Charles, please.

Speaker #1: 2026年9月欧盟的网际韧性法案即将全面强制执行，那治安法规呃在呃要在落地的背景下那公司有没有看到这个PUF在先进制程SOC中的采用趋势？Felix。

Charles Hsu: For standalone application, our initial development focus is mainly on the NOR Flash. The NOR Flash and the NAND Flash are two different memory technologies with different positioning, and they mainly address different application needs. NOR Flash is focused on fast boot, low latency, random read, and execution in place, or XIP, and high reliabilities. Therefore, it is mainly used to store firmware and system call. NAND Flash, on the other hand, is mainly designed for high capacity and low cost, making it more suitable for the larger scale data storage applications. However, after foundry partners successfully move standalone NOR Flash into mass production, we may further extend 1T Flash into NAND Flash architecture in the future.

Speaker #2: 随着这个网际韧性法案法规将正式强制实施，违规将面临高额的罚款，且适用范围涵盖所有在欧洲市场销售的联网产品，市场对治安合规的重视程度持续提升。同时secure by design与secure by default的理念也逐渐成为产业的共识。在这样的趋势下，PUF作为硬体信任跟的关键技术之一，与装置身份识别与金钥保护中扮演核心角色。随着法规要求日益严格，我们预期相关的安全机制的导入需求将持续上升。

Speaker #1: As cybersecurity regulations such as the EU Cyber Resilience Act, or CRA, gradually come into force, and as we approach September this year, how does the company view the adoption trend of PUF in advanced-node SoCs?

Li Ching-Xiang: 2026年9月，欧盟的网际韧性法案即将全面强制执行。在资安法规要落地的背景下，公司有没有看到PUF在先进制程SoC中的采用趋势？Felix。

Li-Jeng Chen: [Foreign language]

Speaker #1: Felix, please.

Speaker #2: With the CRA expected to become mandatory, non-compliance may result in significant penalties. The regulation also applies broadly to all connected products sold in the European market.

Felix Hsu: 随着网际韧性法案法规将正式强制实施，违规将面临高额的罚款，且适用范围涵盖所有在欧洲市场销售的联网产品，市场对资安合规的重视程度持续提升。同时，secure by design 与 secure by default 的理念也逐渐成为产业的共识。在这样的趋势下，PUF 作为硬体信任根的关键技术之一，于装置身份识别与金钥保护中扮演核心角色。随着法规要求日益严格，我们预期相关的安全机制的导入需求将持续上升。

Felix Hsu: [Foreign language]

Speaker #2: So, as a result, the market is placing great emphasis on cybersecurity compliance. At the same time, the concepts of 'secure by design' and 'secure by default' are increasingly becoming industry consensus.

Speaker #2: In such a context, PUF, as a key technology for hardware root of trust, plays an important role in device identity and key protection. As regulatory requirements become more stringent, we expect demand for related security mechanisms to continue to increase.

Speaker #1: 请问公司IP有没有切入一些太空通讯相关应用场景，呃请Joseph回答。

Li Ching-Xiang: As cybersecurity regulations such as the EU Cyber Resilience Act, or CRA, gradually come into force at September this year, how does the company view the adoption trend of PUF in advanced-node SoCs? Felix, please.

Li-Jeng Chen: As cybersecurity regulations such as the EU Cyber Resilience Act, or CRA, gradually come into force at September this year, how does the company view the adoption trend of PUF in advanced-node SoCs? Felix, please.

Speaker #2: 嗯，这方面我们的技术其实已经有被美国客户用在低轨卫星相关的一些应用上。那由于这个低轨卫星是在非常严苛的这个太空环境下来运作，而且发射后基本上不太可能去回收去做这个repair，所以OTP在这一类应用中主要呃有几个机制来提供相关的保障。那首先是抗辐射性以及这个高的可靠性。那太空中的这个强力辐射会使得传统的memory solution，例如说这个flash或是double eprom产生这个位元的翻转，那会导致资料错误以及卫星的失效。那我们的OTP是透过这个物理性的机制来改变电路的结构，那进而去储存这个资料。所以写入就不会受这种辐射的影响。那其实是非常好用来存放这种关键的一些程式码，例如说这个book code，那最安全的一个地方。那第二部分是安全通信以及这个密钥储存。那第一低轨卫星作为这个庞大的网络节点，那治安当然是非常重要。那OTP可以呃用于储存这个加密的金钥以及数位的凭证，那他们具备这个无法被远端篡改或是抹除的一个特性。那其实是帮这个卫星以及这个地面站之间的通讯建立一个非常牢固的硬体信任根，那有效的去防止骇客的一个攻击。那第三个部分就是这个硬体的身份识别，那面对成千上万颗的这个卫星的一个信息，其实每一颗呃晶片在出厂的时候都需要去写入唯一的一个硬体ID。那这对于这个信息的一个管理啊，故障的追踪以及这个频谱的授权啊，这个自动化的验证都是非常重要。最后就是极端环境下的一个参数补偿，其实面对这个太空剧烈的一个温差，OTP储存这个精密感测器的校准以及补偿参数，那可以确保卫星在极端的这个温度温差的环境运作下，各项的这个电子讯号依然能够保持一个良好的一个精准运作。谢谢。

Felix Hsu: With the CRA expected to become mandatory, non-compliance may result in significant penalties. The regulation also applies broadly to all connected products sold in the European market. As a result, the market is placing great emphasis on cybersecurity compliance. At the same time, the concepts of secure by design and secure by default are increasingly becoming industry consensus. In such a context, PUF, as a key technology for hardware root of trust, plays an important role in device identity and key protection. As regulatory requirements become more stringent, we expect demand for related security mechanisms to continue to increase.

Li Ching-Xiang: 请问公司IP有没有切入一些太空通讯相关应用场景？请Joseph回答。

Li-Jeng Chen: [Foreign language]

Joseph Hsia: 这方面我们的技术其实已经有被美国客户用在LEO相关的一些应用上。由于LEO是在非常严苛的太空环境下来运作，而且发射后基本上不太可能去回收，去做repair。所以OTP在这一类应用中，主要有几个机制来提供相关的保障。首先是抗辐射性，以及高的可靠性。太空中的强力辐射会使得传统的memory solution，例如说Flash或是EEPROM产生位元的翻转，会导致资料错误以及卫星的失效。我们的OTP是透过物理性的机制来改变电路的结构，进而去储存资料，所以一旦写入就不会受这种辐射的影响。其实是非常好用来存放关键的程式码，例如说boot code，最安全的地方。第二部分是安全通信以及密钥储存。低轨卫星作为庞大的网络节点，资安当然是非常重要。OTP可以用于储存加密的密钥以及数位的凭证，它们具备无法被远端篡改或是抹除的一个特性。其实是帮卫星以及地面站之间的通讯建立一个非常牢固的信任根，有效地防止骇客的攻击。第三个部分就是硬体的身份识别。面对成千上万颗的卫星的星系，其实每一颗芯片在出厂的时候都需要去写入唯一的一个硬体ID。这对于星际的管理、故障的追踪、芯片的授权、自动化的验证都是非常重要。最后就是极端环境下的参数补偿。面对太空剧烈的温差，OTP储存精密感测器的校准以及补偿参数，可以确保卫星在极端的温度、温差的环境运作下，各项的电子讯号依然能够保持一个良好的精准运作。谢谢。

Joseph Hsia: [Foreign language]

Speaker #1: Has eMemory’s IP been adopted in applications related to low Earth orbit, or LEO satellites, and space communications? Joseph, please.

Speaker #2: Yes, our technology has already been adopted by U.S. customers for low Earth orbit satellite applications. And given that satellites operate in extreme environments and are non-recoverable once launched, our OTP provides four critical safeguards.

Speaker #2: First of all, it's radiation hardening and high reliability. Because in space, intense radiation can cause bit flips, and traditional flash or double EPROM solutions can lead to data errors and mission failure.

Speaker #2: Our OTP, on the other hand, stores data by permanently altering the physical circuit structure. So, once written, it remains immune to radiation, making it the most secure place to store critical boot code.

Speaker #2: And second, it's for secure communication and key storage. As satellites, they function as massive network nodes, so cybersecurity is critical. Our OTP stores encryption keys and digital certificates that cannot be remotely tampered with or erased.

Speaker #2: And this establishes a robust hardware root of trust, protecting the communication between satellites and also the ground stations from hacking attempts. And third is how we're doing identification.

Li Ching-Xiang: Has eMemory's IP been adopted in applications related to low Earth orbit, or LEO, satellites and space communications? Joseph, please.

Li-Jeng Chen: Has eMemory's IP been adopted in applications related to low Earth orbit, or LEO, satellites and space communications? Joseph, please.

Speaker #2: For constellations with thousands of satellites, precise management is vital. Each chip should be programmed with a unique ID during production, which is essential for fleet management, fault tracking, and automated spectrum licensing verification.

Joseph Hsia: Yes, our technology has already been adopted by US customers for low Earth orbit satellite applications. Given that satellites, they operate in extreme environments and are non-recoverable once launched, our OTP provides four critical safeguards. First of all is radiation hardening and high reliability. Because in space, intense radiation can cause bit flips in traditional Flash or EEPROM solutions, and they can lead to data errors and mission failure. Our OTP, on the other hand, they store data by permanently altering the physical circuit structure, so once written, they remain immune to radiation, making it the most secure place to store critical boot code. Second is for secure communication and key storage. As satellites, they function as massive network nodes. Cybersecurity is critical, and our OTP stores encryption keys and digital certificates that cannot be remotely tampered with or erased.

Speaker #2: And lastly, is parameter compensation for extreme environments. Because, as you know, to withstand drastic temperature fluctuations in space, our OTP stores calibration and compensation parameters for precision sensors.

Speaker #2: And this ensures that electronic signals can remain accurate regardless of the extreme thermal conditions. Thank you.

Speaker #1: In the interest of time, we will begin the last questions. 我想这这一个呃昨天早上呃公告就是我们公司是被这个MCI标准指数剔除。那加上这个公司的外资持股比例在台湾可能就正常的呃就是譬如真正的外资应该是仅次于台积电呃呃大概六七percent。那我相信大家投资人都会很关心这个被指数剔除会不会造成大量的外资卖出。那公司是不是有什么因应策略？好，那我就自己来回答。那这个事情那个因为我们都知道今今年就是呃AI浪潮，很多股票是喷涨，在很短的时间涨了呃数倍。那我们不去judge这个fundamental。那因为它标准指数的成分股是以市值排名，好像在赛马一样，所以每一段时间就是你跑输了就被踢掉。那其实是跟公司基本面无关的，有人上去一年内又掉下来。那我我我们呃应该这样讲，我在十年前我们在十年前不在MSCI标准指数的时候，我们外资持股就已经超过50percent了，而大部分外资呢到现在都还持有，甚至在今年7月底呃就7月最后一天的时候，我们的外资持股是历史新高。那询问我们可是它就是已经发生了嘛。OK，因为你你没办法你没办法控制别人涨这种排名的事。那我们能做就是我们去问几个我们持股比例呃高的这个大股东。那我们问他说哎，那你你如果我们在标准指数会不会去影响你的呃这个持股决策？那他们的回回复呢都说不会，因为他们虽然benchmark and MCI，但他们是主动判断的。OK，那对于这个被动基金就是ETF调节。那根据过去的经验，这类的指数成分股的变动，因为这个是游戏嘛，就是有剂量方法是可以预测的。所以年初以来就呃外资券商或者是内资的broker就有预估说啊，我们或其他公司会被剔除。所以呢，事实上他们都会事先反应啊，不会到最后。所以你可以看到呃年初以来到今天呃外资呃是在卖方啊，有的是借券，借券就是等着最后一天去block trade给呃要卖的被动。那另外他呃我们是还有一个小型基金嘛，我们是用标准指数放到小型，小型他也会有个买盘。所以呢，我们应该认为说他已经在发生了。OK，那你如果问我们说他会调节到什么时候，其实我们也不清楚，因为没人知道。好，那对我们来讲，一个客户来讲，我们能够做的可能我们以前做的也是不够，所以我们会更积极呢跟资本市场沟通啊，特别是内资啊，因为我们发现我们的外资收长线，他买了之后不动，我们就出现了一个叫流动性的问题。所以这也不见得是坏事，因为这个被动出来就是强迫呃增加liquidity。那我会呃非常积极的去参加法说，甚至于OTC针对散户的8月有一场我也去，如果是个人股东的话也欢迎你当场去去询问问题。那另外一个很重要的就是增加这个外资的报告呃，在很很难懂呃也很难写，但是就尽量呃让更多人cover。所以你在information的时候，你就不就不会觉得说哎发生什么事，你不知道发生什么事，会不动他们，因为过去的波动主要是这个流动性不好，有人买有人卖，那个波动就很可怕。那最最重要呢，我们非常非常深具信心，这个是十年对我们来讲最好的十年。所以我们未来会用这个基本面跟数字来证明公司的价值。以上。 Being removed from the MSCI index coupled with our relatively high foreign ownership may may raise concerns about potential heavy foreign sell-off and how the company plans to response.

Joseph Hsia: This establishes a robust hardware root of trust, protecting the communication between satellites and also the ground stations from hacking attempts. Third is hardware identification. For a constellation with thousands of satellites, precise management is vital, and each chip should be programmed with a unique ID during production, which is essential for fleet management, fault tracking, and ultimately spectrum licensing verification. Lastly is a parameter compensation for extreme environments. As you know, to withstand drastic temperature fluctuation in space, our OTP stores calibration and compensation parameters for precision sensors, and this ensures that electronic signals, they can remain accurate regardless of the extreme thermal conditions. Thank you.

Li Ching-Xiang: In the interest of time, we will begin the last questions.

Li-Jeng Chen: In the interest of time, we will begin the last questions.

Qing-Xiang Hsu: 昨天早上公告，我们公司被MSCI标准指数剔除，加上这个公司的外资持股比例，在台湾真正的外资应该是仅次于台积电，大概六七%。我相信大家投资人都会很关心，被指数剔除会不会造成大量的外资卖出，公司是不是有什么因应策略？那我就自己来回答。这个事情，因为我们都知道今年是AI浪潮，所以很多股票是喷涨，在很短的时间涨了数倍。那我们不去judge这个fundamental，因为它标准指数的成分股是以市值排名，好像在赛马一样，所以每段时间你跑输了就被踢掉，其实是跟公司基本面无关的。有人上去一年内又掉下来。我们应该这样讲，我们在10年前不在MSCI标准指数的时候，我们外资持股就已经超过50%了，大部分外资到现在都还持有。甚至在今年7月底，就7月最后一天的时候，我们的外资持股是历史新高。可是它就是已经发生了，OK，因为你没办法控制别人涨这种排名的事。那我们能做的就是，我们去问了几个我们持股比例高的大股东，我们问他说："如果我们在标准指数，会不会去影响你的持股决策？"他们的回复都说不会，因为他们虽然benchmark MSCI，但他们是主动判断的。那至于被动基金，就是ETF的调节，根据过去的经验，这类的指数成分股的变动，因为这个是个游戏，有计量方法是可以预测的。所以年初以来，外资券商或者是类似的broker就有预估说我们或其他公司会被剔除。所以事实上他们都会事先反应，不会到最后。所以你可以看到年初以来到今天，外资是在卖方，有的是借券，借券就是等了最后一天去block trade给要卖的被动。那另外我们是还有一个小型基金，我们是从标准普尔降到小型，小型它也会有一个买盘。所以我们应该认为说它已经在发生了。那你如果问我们说它会调节到什么时候，其实我们也不清楚，因为没人知道。那对我们一个Corp来讲，我们能够做的，可能我们以前做得也是不够，所以我们会更积极跟资本市场沟通，特别是内资。因为我们发现我们的外资都长线，他买了之后不动，我们就出现了一个叫流动性的问题。所以这也不见得是坏事，因为这个被动出来就是强迫增加liquidity。那我会非常积极地去参加法说，甚至于OTC针对散户的，八月有一场我会去，如果是个人股东的话，也欢迎你当场去询问问题。那另外一个很重要的就是增加外资的报告。虽然很难懂，也很难写，但是就尽量让更多人cover。所以你在information的时候，你就不会觉得说发生什么事，你不知道发生什么事，波动很大。因为过去的波动主要是流动性不好，所以有人买，有人卖，那个波动就很可怕。那最最重要的，我们非常地深具信心，这个是十年，对我们来讲最好的十年。所以我们未来会用基本面跟数字来证明公司的价值。以上。

Li-Jeng Chen: [Foreign language]

Speaker #1: MSCI index inclusion is primarily based on market capitalization ranking, and it's not a reflection of company fundamentals. In fact, 10 years ago, when we were not included in the MSCI standard index, our foreign ownership of our shares was already above 50%.

Speaker #1: And at the end of July this year, our foreign shareholding actually reached a record high of about 67% to 68%, just slightly lower than the TSMC foreign shareholding.

Speaker #1: We have reached out to our major foreign shareholder, and their response is that the index adjustment will not affect their investment decision. As for passive fund rebalance, experience shows that the index constituent changes are largely predictable during quantitative modeling, so the market usually discounts and absorbs the impact in advance.

Speaker #1: The foreign selling pressure observed since early August likely reflects this passive fund adjustment. To mitigate significant stock price volatility, we will proactively enhance our communication with the capital market, especially local institutional investors.

Li Ching-Xiang: Being removed from the MSCI index, coupled with our relatively high foreign ownership, may raise concerns about potential heavy foreign sale off and how the company plans to respond.

Li-Jeng Chen: Being removed from the MSCI index, coupled with our relatively high foreign ownership, may raise concerns about potential heavy foreign sale off and how the company plans to respond. MSCI index inclusion is primarily based on market capitalization ranking, and it's not the reflection of company fundamental. In fact, 10 years ago, when we are not including in the MSCI standard index, our foreign ownership of our shares already above 50%. At the end of July this year, our foreign shareholder actually is a record high like 67% to 68%, just slightly lower than the shareholding of the TSMC foreign shareholding. We have reached out to our major foreign shareholder, and their response that the index adjustment will not affect their investment decision. As for passive fund rebalance, experiences show that the index comes to change are largely predictable during quantitative model.

Speaker #1: We plan to participate in more investor conferences and expand analyst research coverage to reflect the company's true value through our strong fundamentals and operating results.

Qing-Xiang Hsu: MSCI index inclusion is primarily based on market capitalization ranking, and it's not the reflection of company fundamental. In fact, 10 years ago, when we are not including in the MSCI standard index, our foreign ownership of our shares already above 50%. At the end of July this year, our foreign shareholder actually is a record high like 67% to 68%, just slightly lower than the shareholding of the TSMC foreign shareholding. We have reached out to our major foreign shareholder, and their response that the index adjustment will not affect their investment decision. As for passive fund rebalance, experiences show that the index comes to change are largely predictable during quantitative model. So the market usually discount and absorb the impact in advance. The foreign selling pressure observed since early August likely reflect this passive fund adjustment.

Speaker #2: And next, we will begin the closing comments. Please proceed.

Speaker #3: Okay, thank you for attending our investor conference. For more information about our top-based security IP and technology, we encourage you to visit our PUFsecurity website and check out our articles and other materials. Thank you once again for your patience and for the support for eMemory. We will continue to work hard on technology and IP innovation, and top-based hardware security solutions for our customers, bringing higher returns for our shareholders. Thank you.

Speaker #1: Thank you, ladies and gentlemen. Please be advised that the conference recording will be accessible within the next three hours. Thank you, everyone, for joining us today.

Li-Jeng Chen: So the market usually discount and absorb the impact in advance. The foreign selling pressure observed since early August likely reflect this passive fund adjustment. To mitigate significant stock price volatility, we will proactively enhance our communication with the capital market, especially local institution investor. We plan to participate in more investor conference and expand analyst research coverage, proving the company to value through our strong fundamental and operating result.

Qing-Xiang Hsu: To mitigate significant stock price volatility, we will proactively enhance our communication with the capital market, especially local institution investor. We plan to participate in more investor conference and expand analyst research coverage, proving the company to value through our strong fundamental and operating result.

Li Ching-Xiang: Next, we will begin the closing comments. Chair, please proceed.

Operator: Next, we will begin the closing comments. Chair, please proceed.

Charles Hsu: Thank you for attending our investor conference. For more information about our PUF-based security IP and technology, we encourage you to visit our PUFsecurity website and check out our articles and other materials. Thank you once again for your patience and the support for eMemory. We will continue to work hard on technology and IP innovation and PUF-based hardware security solutions for our customers and bring in higher returns for our shareholders. Thank you.

Li Ching-Xiang: Thank you, ladies and gentlemen. Please be advised that the conference recording will be accessible within the next 3 hours. Thank you everyone for joining us today. We hope you will join us again next quarter. You may now disconnect. Goodbye and have a good day.

Li-Jeng Chen: Thank you, ladies and gentlemen. Please be advised that the conference recording will be accessible within the next 3 hours. Thank you everyone for joining us today. We hope you will join us again next quarter. You may now disconnect. Goodbye and have a good day.

Charles Hsu: Mm-hmm.

Q2 2026 eMemory Technology Inc Earnings Call

Request TrialDemo

eMemory Technology

Earnings