SAN FRANCISCO — The American artificial intelligence industry spent the week doing what it does best — fighting itself — as hundreds of OpenAI and Google employees filed a legal brief backing rival Anthropic against the U.S. government, OpenAI's own boardroom drama threatened to torpedo its planned stock offering, and analysts warned the whole mess could hand the crown to Google on a silver platter.
Start with the strange bedfellows. Rank-and-file engineers at OpenAI and Google put their names on an amicus brief siding with Anthropic in a fight with the Pentagon. Workers at competing shops don't usually lock arms. They did this time. The brief signals that the people building these systems believe something bigger than quarterly revenue is at stake — namely, who gets to decide how AI tools serve the defense establishment.
Meanwhile, the corner office at OpenAI can't stop generating headlines that have nothing to do with large language models. The company's revolving door of executives, public spats, and governance circus have Wall Street types asking a pointed question: Can this outfit hold it together long enough to ring the bell? An IPO requires investor confidence. Confidence requires stability. Stability is not the word anyone uses to describe OpenAI in 2025.
Here's the kicker. Axios reports the knife fight between OpenAI and Anthropic could end up being the best thing that ever happened to Google. While the two startups bloody each other over talent, contracts, and public trust, the search giant sits on more compute, more data, and more distribution than both of them combined. Every dollar OpenAI and Anthropic spend on lawyers and lobby shops is a dollar not spent beating Gemini.
And if the domestic brawl weren't enough, there's a gate-crasher. China's DeepSeek claims it trained high-performing AI models on the cheap — without access to the most advanced American chips that export controls were supposed to keep out of reach. If those claims hold water, the entire theory that U.S. hardware dominance guarantees U.S. AI dominance needs rewriting.
The arithmetic is brutal. Three American companies fighting on three fronts — against each other, against the government, and against their own employees — while a Chinese upstart rewrites the cost curve from the other side of the Pacific.
For outfits like Trilogy International's portfolio companies that depend on enterprise AI tooling to run lean operations across 130-plus countries, the question isn't which lab wins the cage match. The question is whether the cage match slows down the technology everybody else needs to ship product.
Nobody on Sand Hill Road had a good answer for that one today.
