AI markets have crystallized. Investor Elad Gil, who backed Harvey, Perplexity, and Character.AI in their early rounds four years ago, argues that the first wave of AI market winners has now solidified. The LLM layer is the clearest example: Anthropic, Google, Meta, Microsoft, Mistral, OpenAI, and X.AI are the named survivors, with three or four of those driving the majority of developer adoption and industry spend. Hyperscaler partnerships, specifically Amazon with Anthropic, Google GCP with Gemini, and Microsoft Azure with OpenAI, have created a funding dynamic that operates independently of whether these are good investments. Cloud AI spend has reached several billion dollars per quarter for the leading providers.
The interesting argument is not who won, but what winning means at this stage. Gil draws the Stripe and Facebook analogies deliberately: PayPal did not stop Stripe, Myspace did not stop Facebook. Current leaders can still be displaced, acquired, or killed. Newer entrants like SSI and Thinking Machine Labs remain wildcards, either finding novel approaches or becoming acquisition targets for incumbents buying talent. The code and software engineering vertical illustrates how fast the picture moved: in 2022 the category was obviously important, but Cursor did not launch until 2023, Windsurf reached general availability nine months ago, and Cognition launched Devin in limited release just over a year ago.
Gil's broader framework, covering not just LLMs but voice, image, video, biology, and materials models, is where the original piece earns a full read. He maps which adjacent markets have clarity now and which remain genuinely uncertain, and he does it from a position of having written the first checks. The question his analysis raises but does not fully close: if the LLM market is locked, where does the next four years of early-stage opportunity actually live.
[READ ORIGINAL →]