• by Laurie Sullivan , Staff Writer, 10 hours ago

Anthropic confirmed an investment from Google -- $10 billion in cash, with the expectation that another $30 billion could potentially follow if it hits specific performance targets -- the company said Friday.

The reports come as Anthropic sees demand for the company’s coding agent, Claude Code, and a similar $25 billion commitment from Amazon on April 20. The initial new investment from Amazon is $5 billion, with possibly another $20 billion in the future. This builds on the $8 billion that Amazon previously invested.

The investment commitment from Google comes the same day it announced an update about eliminating political bias in Claude. The company wants to ensure that people can ask questions and receive honest answers about political parties, candidates and issues during election time, as well as answering simpler questions such as when, where, and how to vote.

Anthropic is working to ensure that Claude can remain a reliable source ahead of the U.S. midterms and other major elections worldwide, even without advertising.

To ensure fair and unbiased answers, the company trained Claude to treat different “political viewpoints with equal depth, engagement, and analytical rigor — a principle set out in Claude’s constitution.”

This has been built into the model through character training, which actually rewards the model for producing responses that reflect a set of values and traits, which are then reinforced through system prompts, which carry explicit instructions on political neutrality into every conversation on Claude.ai.

AI models pick up bias from reading mounds of information across the internet, and subsequently tilt information a particular way.

It’s not clear whether these types of learnings and teachings for its models have prompted additional separate investments from Google and Amazon, but it makes sense that the company has waited to build out an advertising ecosystem and first focus on unbiased responses.

Imagine that an unbiased AI model funded by advertising would prioritize transparency, fairness and user empowerment over subtle manipulation or undisclosed sponsorships. In this ideal state, the AI acts as a neutral agent, balancing commercial interests with objective user assistance.

Anthropic also released its latest model, Mythos, to a limited group of partners this month, claiming it is the company’s most powerful model to date and has significant cybersecurity applications.

Along with Claude, Anthropic requires increased computing power, and has faced widespread complaints about its use limits.  

Earlier this month, Anthropic struck a deal with cloud computing provider CoreWeave for data-center capacity. As part of its agreement with Amazon, Anthropic is expected to gain computing power.

Although they are competitors, Anthropic relies heavily on Google Cloud for chips and infrastructure, including access to its tensor processing units (TPUs) that specialize in chips designed for AI workloads.

It also relies on chips from Broadcom through a strategic partnership that primarily involves Google’s hardware.

All supports advertisers in a roundabout way. While Anthropic does not run an advertising business, its reliance on Google, Broadcom and others supports the broader advertising ecosystem through a "circular economy" of infrastructure and funding.