Google was already a big investor in Anthropic with an estimated $14 billion stake and significant TPU deals, but it turns out that this relationships goes even deeper.
According to a report by the Financial Times, Demis Hassabis — the CEO of Google DeepMind and one of the most influential figures in AI — was a personal, early investor in Anthropic. The report also said that Anthropic CEO Dario Amodei considers Hassabis a “role model”. The revelation adds a striking new layer to the already intertwined world of frontier AI, and reveals just how closely Google and Anthropic are aligned in the race to build transformative AI.
OpenAI Was Built to Stop Google
To understand why this matters, you have to go back to OpenAI’s origins. Elon Musk co-founded OpenAI partly to prevent Google from controlling AGI — an explicit check on what he saw as the most dangerous concentration of AI power imaginable. The founding philosophy was simple: if AGI was coming, it couldn’t be left in the hands of a single corporation. The non-profit, open-research model was meant to be the antidote to Google’s dominance.
That rivalry was personal from the start. The battle to recruit Ilya Sutskever away from Google to OpenAI reportedly ended Musk’s friendship with Google co-founder Larry Page. It was, in Musk’s own words, mostly a fight between him and Demis Hassabis — who was trying to keep Sutskever inside the Google ecosystem. Musk won that round. But the broader war between the two camps has only intensified since.
Anthropic’s Break from OpenAI
Anthropic was founded in 2021 when Dario Amodei, Daniela Amodei, and roughly a dozen other researchers walked out of OpenAI over AI safety concerns. The view inside that group was that OpenAI, increasingly commercialized and under pressure, wasn’t serious enough about the dangers of the technology it was building. Anthropic was the supposed corrective — a safety-first lab that would build powerful AI but not at the expense of alignment research.
It seems that Anthropic, born out of distrust for OpenAI’s direction, had no such distrust of Google. Hassabis investing personally in Anthropic in its early days — while simultaneously running Google’s own frontier AI lab — suggests the relationship between the two organizations was warm from the very beginning. OpenAI had removed “safe AI” from its mission statement; Anthropic, it seems, found a more natural ally in the company it was supposedly competing against.
Google and Anthropic: Closer Than the Org Chart Suggest
On paper, Google and Anthropic are competitors. Google DeepMind is building Gemini. Anthropic is building Claude. They are fighting for the same enterprise customers, the same developer ecosystems, and ultimately the same shot at AGI.
But the financial and personal ties run deep. Google has poured billions into Anthropic, making it one of its largest backers alongside Amazon. Anthropic runs significant workloads on Google Cloud. And now it emerges that Hassabis was an early believer in the company — not as a corporate bet, but as a personal investment.
This is a different kind of relationship than what exists between, say, Microsoft and OpenAI — a partnership often described as fraught with tension and complicated by contractual clauses about what happens when AGI is achieved. The Google-Anthropic alignment, by contrast, looks more like genuine ideological proximity. Hassabis has long been concerned about safety in AI development — he was arguing against reckless open-sourcing of AI as far back as 2016, in a pointed email to Elon Musk. Anthropic, founded on precisely those concerns, was a natural place for him to put his money.
Implications for the AI Race
The frontier AI landscape is now essentially a three-way contest: Google, Anthropic, and OpenAI. xAI is emerging as a fourth, and Meta is a wildcard with its open-source strategy. But at the very top, the race is being run by those three labs.
What’s becoming clear is that this isn’t a clean three-way split. Google and Anthropic are structurally, financially, and — it now appears — personally entwined. That has implications for how the competitive dynamics play out. If Anthropic is both a Google partner and a company whose leadership philosophy aligns with Hassabis’s own views on safety, the two organizations are more likely to coordinate on norms, share infrastructure, and present a unified front on regulatory questions.
OpenAI, meanwhile, sits increasingly alone. Its relationship with Microsoft is complicated, its governance has been turbulent, and 8 of its 11 co-founders have departed. Anthropic, by contrast, is the only major AI lab to have retained every single one of its co-founders — a striking stability that may reflect the coherence of its founding mission.
The original story of AI’s frontier was framed as a race to stop Google. It now looks rather different. Google’s CEO is personally invested in the lab that split from OpenAI over safety concerns. The company that OpenAI was built to counter has become Anthropic’s closest ally. And OpenAI — the institution created to supposedly democratize AI — is racing largely on its own.
The allegiances in this race are not what they first appeared to be.