Google invests $300M in OpenAI rival Anthropic

Microsoft barely had time to celebrate its fresh $10 billion bet on OpenAI before Google fired back with its own proxy.

According to a report from the Financial Times, Google quietly poured about $300 million into Anthropic in late 2022, securing a 10 percent stake in the high-flying AI startup. The cash injection pushes Anthropic's valuation to roughly $5 billion.

Alongside the funding news, Anthropic announced Google Cloud as its preferred cloud provider, with both companies teaming up to co-develop dedicated AI computing hardware.

The OpenAI defectors building Claude

If you haven't tracked Anthropic yet, its pedigree speaks for itself. The company was founded in 2021 by Dario Amodei, the former vice president of research at OpenAI, who brought along a squad of top researchers including Tom Brown, the lead engineer behind GPT-3.

The split was ideological. Amodei and his crew walked away after growing uncomfortable with OpenAI's aggressive commercial pivot following Microsoft's initial 2019 investment. While OpenAI rushed ChatGPT onto the open web, Anthropic took the public benefit corporation route, pitching AI that is reliable, interpretable, and steerable.

Their flagship product is Claude, a general-purpose conversational chatbot currently in closed beta via Slack integrations. Inside research circles, Claude has turned heads for occasionally outperforming ChatGPT.

Anthropic trains Claude using a technique it calls Constitutional AI. Instead of relying solely on messy human feedback, the process combines supervised learning and reinforcement learning against a set of core principles, teaching the model to explain why it cannot fulfill a harmful prompt rather than bluntly stonewalling or hallucinating.

Anthropic's ledger isn't entirely squeaky clean, though. A big chunk of its previous $580 million funding round came from now-disgraced FTX founder Sam Bankman-Fried, leaving lingering questions about whether bankruptcy trustees might try to claw back those funds.

Why it matters

For developers and power users, Google's backing turns Claude from an academic experiment into a viable, enterprise-grade alternative to ChatGPT. Once Claude moves past its private beta, having a competitive second frontier model will give builders real bargaining power over API pricing and infrastructure lock-in.

For the tech industry, this solidifies the new playbook of the generative AI era: cloud titans arming elite research labs with compute in exchange for equity and platform usage. Google has no shortage of internal language models, but seeing Microsoft sprint to bake ChatGPT into Bing forced Mountain View's hand.

We think this deal is a smart hedge for Google. For just $300 million, it locks in a massive cloud computing customer while securing a front-row seat to OpenAI's sharpest rival. The open question is whether Anthropic can hold onto its high-minded safety principles when its biggest backer expects fast commercial ammunition against Microsoft.