NVIDIA’s print matters less as a one-quarter beat than as evidence the AI capex cycle is still expanding before customers have fully monetized it. That typically pulls through the hardest into the picks-and-shovels stack: HBM, advanced packaging, and networking should keep out-earning the headline GPU trade over the next 1-3 months, while cloud/software names that are funding the spend face near-term margin dilution and slower ROI scrutiny. The risk is not demand collapse; it is capex digestion if hyperscaler commentary turns from scarcity to optimization.

For Unity, the insider filing is mechanically irrelevant; the real signal is whether the current rerate can survive normalization. After a sharp run, the stock now trades on sustained ad-tech execution, so the next leg depends on whether growth converts into durable FCF rather than just better sentiment. Competitive pressure from larger ad platforms means any softness in ad budgets can hit the multiple first, with earnings revisions following later over the next 1-2 quarters.

Contrarian take: consensus is likely overconfident on the durability of the AI enthusiasm trade and underappreciating how crowded the winners have become. NVIDIA can still compound, but upside from here is more about relative share capture than a clean re-rating. Unity looks improved, but the market may already be discounting a lot of the turnaround before it is fully proven.