Source: Anthropic plans to spend ~$200B on Google's Cloud and chips over five years, representing 40%+ of the “revenue backlog” that Google disclosed last week
When Google last month said it would supply Anthropic with an astonishing five gigawatts of server capacity …
The Information
Context & Ripple Effects
Related coverage traces a rapid expansion of the Anthropic–Google infrastructure relationship: discussions over high-tens-of-billions of dollars in compute became a partnership for 1 million TPUs and 1GW of 2026 capacity, followed by a reported five-year spending commitment of roughly $200B.
The reported commitment is unusually consequential because it accounts for more than 40% of Google’s disclosed revenue backlog. Later reporting that Anthropic pursued more than a dozen direct data-center leases suggests it is also seeking infrastructure options beyond a single cloud-procurement path.
First-order effects
Google Cloud gains a large, long-duration source of committed demand for its cloud services and chips, with Anthropic becoming a material contributor to the backlog Google has disclosed.
Anthropic secures a much larger supply of Google infrastructure, while taking on a major multi-year spending obligation tied to its ability to deploy and monetize AI capacity.
Second-order effects
The scale of Anthropic’s demand strengthens Google’s case for continued investment in TPUs and cloud capacity, while making execution—delivering power, data centers, and hardware—more central to converting backlog into revenue.
Anthropic’s reported direct leasing activity indicates that even a deep cloud partnership may not satisfy all of its capacity needs, increasing pressure on other infrastructure providers and data-center operators to compete for frontier-model workloads.
Third-order effects
If comparable commitments become common, frontier AI developers may increasingly function as anchor tenants for cloud and chip suppliers, concentrating a meaningful share of infrastructure demand in a small number of buyers and providers.
The combination of cloud commitments, strategic investment, and direct leasing points toward a more vertically intertwined AI infrastructure market, though Anthropic’s multi-sourcing efforts could limit dependence on any single provider.
The trend: This is one data point in the shift from ad hoc AI compute purchases to multi-year, capacity-reserving infrastructure agreements that bind model developers more closely to cloud and chip platforms.
The fact that this is seen as a bullish sign for Google is a sign our markets are controlled by people without object permanence. Anthropic *cannot afford to pay these bills* - Google just gave them $10 billion! - which means it's ultimately VC or Google feeding itself money
$GOOGL overtaking $NVDA is the most valuable company in the world after hours on reports that Anthropic will commit to spending $200B on cloud and chips @TimSeymour, @grassosteve , Julie Biel, and Empower Investments Chief Investment Strategist Marta Norton break down what it
News tonight with Anthropic and Google's deal announced is that it's 40% of Google's backlog and now The Information is saying those two frontier labs account for 50% of the backlog at all the major CSPs. Surely someone has done the math to compare estimates? Comparing apples t…
My prediction is Anthropic is making a big mistake by using Google's TPUs and not doing more with NVIDIA or Amazon. Why? 1. Google has 3 other massive internal customers fighting for these same TPUs: search, Youtube, and Gemini, which directly competes against Anthropic. [image]
ANTHROPIC COMMITS TO SPENDING $200B ON GOOGLE CLOUD AND CHIPS OVER THE NEXT 5 YEAR — THE INFORMATION Google $GOOGL is now the most valuable company on Planet Earth, overtaking $NVDA Nvidia. [image]
new: Anthropic and OpenAI spending commitments make up half of the $2 trillion revenue backlogs reported by Microsoft, Amazon, Google and Oracle. [image]
The Information: “But as part of the deal, which begins next year, Anthropic plans to spend about $200 billion with Google over five years, according to a person with knowledge of it. The commitment means Anthropic represents more than 40% of the “revenue backlog” Google
I repeat, the bubble is in the “e” not the “p” in today's PE ratios. The hucksters and talking heads will, as always, fail to realize until it's too late. But it's a very simple set up. Hyperscalers give OpenAI/Anthropic cash to pay for hyperscaler computing capacity. That
What are the chances that OpenAI and Anthropic turning on the hyperscalers and launching their own clouds? As Bezos said “Your margin is my oppurtunity”