Google parent Alphabet may invest up to $40 billion in Anthropic

Google parent Alphabet (GOOGL, GOOG) is putting up to $40 billion into Anthropic, even though Anthropic’s Claude sits right across the table from Gemini in the AI model fight.

On paper, that looks like a company handing rockets to its rival, but allow us to familiarize you with the mind of Alphabet CEO Sundararajan Pichai.

What this guy is trying to do here is get control of Anthropic’s computing infrastructure while also keeping Claude tied to Google’s cloud pipes.

On Friday, Google committed $10 billion in cash at a $350 billion valuation, with the rest ($30 billion) coming only after Anthropic hits agreed-upon targets. Interesting, isn’t it?

Google turns Claude traffic into cloud revenue while Gemini stays under pressure

Sure, Anthropic fights Google on AI models, but Claude’s API traffic can still run on Google Cloud, which means Google can earn money when customers use Claude, even if those customers are not using Gemini.

Every Claude request on Google’s infrastructure becomes cloud income, so this is just a tenant getting locked into the landlord’s building.

The Anthropic-Google deal comes with about 5 gigawatts of TPU compute tied to Google’s stack for the next five years, which matters for all companies already using Claude in real products.

One example in the notes is Fortuna, which uses Claude agents for e-commerce customer service. For users like that, Claude is no longer backed only by Anthropic’s own balance sheet. It now has Alphabet’s financial weight behind the compute it needs.

Gemini and DeepMind teams still need more GPU and TPU resources for their own work, so I bet watching a competitor get huge compute support doesn’t feel so cute.

Some of the staff were allegedly told the company needs to make sure Google wins “in any scenario.” Several current and former Google executives and employees have said businesses are starting to use AI tools that can build products from simple chatbot prompts, but Google does not yet have one clean product answer for that demand.

Gemini is spread across different tools, names, and workflows, which honestly slows teams down and makes the product story harder to follow.

Pichai puts Gemini into products, coding work, and defense talks

This is the kind of chess master Sundar has always been, long before he became Google CEO in 2015. As Cryptopolitan reported in our Op-ed on Sundar last year, the CEO has spent years turning research into products used by billions of people all over the world.

After OpenAI made ChatGPT the loudest name in AI in 2022, Sundar had to answer questions about whether it was still leading. So what did he do? Well, he combined Google Brain and DeepMind, built Gemini, and put it into Search, Android, Chrome, Google Cloud, and its own app.

Google is also back in U.S. defense AI talks. The Pentagon is reviewing how much it depends on current vendors and is looking at more partners after tensions around systems like Claude, according to a report from The Information, citing two people briefed on them.

The possible agreement would allow the U.S. Department of Defense to use Gemini for classified and other lawful work.

The military wants faster AI tools inside daily operations, especially for decisions and battlefield awareness.

A Pentagon official allegedly told reporters: “The Pentagon will continue to rapidly deploy frontier AI capabilities to the warfighter through strong industry partnerships across all classification levels.”

The department is testing several AI platforms while building its own rules for handling them.

Inside Google, Sundar also says AI now writes a huge share of code. In a post on The Keyword, he said: “Today, 75 per cent of all new code at Google is now AI-generated and approved by engineers, up from 50 per cent last fall.”

He also said one complex code migration was finished six times faster with AI agents and engineers than human engineers could do one year earlier.

Sundar added: “We’re now shifting to truly agentic workflows. Our engineers are orchestrating fully autonomous digital task forces, firing off agents and accomplishing incredible things.”