Who Will Acquire OpenAI and Anthropic?
Frontier LLMs are money losers.
Companies mentioned: OpenAI, Anthropic, xAI, Google (GOOGL), Meta (META)
I am not referring to the previous generation of LLMs as money losers, but rather the bleeding edge models that will always be money losers. I am aware that OpenAI is trying to build its version of the AppStore and screw its developer partners by way of making partner applications serve GPT’s chat feature in a homogenized fashion. I am also aware that OpenAI and Anthropic may fold Ads into their respective free tiers, which means less privacy than a Google search. The potential success of these initiatives is 50/50 in my view.
Stay current with T2D Pulse - our Technology news aggregation site
OpenAI and Anthropic desperately require server capacity (e.g., “compute”, “cloud resources”), but can’t afford it. Therefore, each company is perpetually raising money. What if that equity pulls back? Debt investors won’t extend credit to OpenAI nor Anthropic unless they bring their cloud partners into the deal. If OpenAI and Anthropic can’t stand on their own two feet, how much longer will this game of musical chairs last before valuations come in dramatically?
OpenAI and Anthropic will be acquired in the end. Better to sell now, even if this is not an absolute market top versus trying to sell after a 90% haircut. However, I’m not sure OpenAI’s or Anthropic’s partners would be buyers. Those partners know this game can’t last and don’t want to pay top dollar valuations.
IPO as an exit? Maybe, but how long will lofty valuations last in the public market when every successive quarter consists of larger and larger losses?
Google and xAI have the benefit of real balance sheets that can offset LLM losses. Meta does as well, but it is now transitioning from an opensource LLM strategy to proprietary, which I believe is a mistake. A proprietary, frontier LLM strategy requires that the company in question be “all in”, something META is not.
The only save I can think of with respect to OpenAI and Anthropic maintaining independence over the long-term would be if the U.S. Government decides to subsidize OpenAI and Anthropic. Doing so would be incredibly stupid.
