Author: Ada, Shenchao TechFlow
On May 5, according to The Information, Anthropic pledged to pay $200 billion to Google Cloud over the next five years.
This multi-year agreement, starting in 2027, will account for more than 40% of Google Cloud's revenue backlog, a metric that reflects contractual commitments from enterprise customers.
An AI company that didn’t exist five years ago has secured nearly half of Google Cloud’s future revenue through a single contract.
On the day the news was released, Alphabet rose 2% after hours.
But more intriguing is another figure: Alphabet has also made a reverse investment of up to $40 billion in Anthropic.
Money leaves Google’s account, circulates, and returns to Google’s account. In between, an accounting line item labeled "Anthropic Compute Expense" appears.
So, is this the largest cloud computing order in history, or the most elegant financial trick ever?
An "exclusive commitment" not just for Google
To understand the nature of this transaction, first examine a set of data that is not isolated.
On April 20, Anthropic announced an expanded partnership with Amazon, committing to spend over $100 billion on AWS technologies over the next decade in exchange for up to 5 gigawatts of computing power. In return, Amazon is adding up to $25 billion to its existing $8 billion investment.
In November last year, Microsoft agreed to invest up to $5 billion in Anthropic, and Anthropic committed to purchasing $30 billion worth of Azure computing power.
In other words, Google: invested $40 billion, received $200 billion. Amazon: invested $33 billion, received over $100 billion. Microsoft: invested $5 billion, received $30 billion.
The three major cloud giants collectively invested approximately $78 billion in return for $330 billion in contractual commitments, resulting in a net inflow of $250 billion on paper.
The essence of this strategy is to reclassify capital expenditures as revenue. The money invested in Anthropic is recorded as cash flow from investing activities, while Anthropic’s payments for computing power are recorded as operating revenue. The same funds move from one pocket to another, resulting in a clean, attractive backlog on the financial statements.
Alphabet is funding Anthropic while counting Anthropic’s compute purchases as future revenue, creating a self-reinforcing loop in AI infrastructure growth.
Wall Street is the real winner in this game—as long as the backlog number is large enough, the price-to-earnings ratio can be sustained.
Advanced version of the flywheel
The story of adding positions at high levels isn't over yet—AI circles have amplified the same flywheel a thousandfold.
The strategy's logic is to raise funds by issuing stock, buy Bitcoin, use the rising price of Bitcoin to increase market capitalization, issue more stock, and buy more Bitcoin.
The cloud providers' logic is that they invest in AI companies, which then pay for computing power, driving revenue growth and stock price increases, prompting further investment from capital markets.
In contrast, Bitcoin is a scarce asset, with each coin corresponding to actual on-chain supply. Hash power is not. The “gigawatt-scale TPU capacity” set to launch in 2027 doesn’t even have racks installed today.
In other words, a significant portion of the $200 billion represents Anthropic’s advance commitment to purchase a batch of chips that have not yet been manufactured, allowing Google to use this commitment to persuade capital markets.
Isn’t this just a forward contract? The difference is that commodity futures have a delivery date and margin requirements, but this contract has neither. What if Anthropic can’t pay this amount in 2027—who bears the cost of default?
It’s not Google. It has already included the backlog in its earnings call slides. On April 29, Alphabet disclosed during its earnings call that Google Cloud revenue increased by 63% year-over-year, surpassing $20 billion, with cloud backlog reaching approximately $462 billion. This figure underpins Alphabet’s current market valuation.
It wouldn't be Anthropic either. It just needs to keep raising funds, since the next valuation is still rising.
The final buyers may still believe they are investing in the "AI shovel seller" story.
5 billion mobilizes 330 billion
Is Anthropic’s scale commensurate with this figure?
According to media reports, Anthropic's annual revenue increased from $1 billion to $5 billion in 2025.
A company with an annual revenue of only 50 billion signed three contracts totaling 330 billion: a 5-year contract worth 200 billion, a 10-year contract worth 100 billion, and an additional $30 billion.
Even if Anthropic's revenue were to increase tenfold, its cumulative revenue over five years would still fall short of $330 billion.
So, where does the money come from?
There’s only one path forward: secure additional funding.
And the largest potential investors are precisely these three cloud providers themselves.
That’s the entire secret of the cycle. Anthropic doesn’t need to actually make money—it just needs to stay in a state of continuous fundraising, treating each new round of funding as next year’s compute bill. As the valuation rises, it can raise even more.
Who does it sound like?
Strategy doesn't require Bitcoin to generate actual cash flow; it only needs to maintain the ability to continuously issue equity and debt. The only difference is that Strategy's balance sheet includes Bitcoin, a globally publicly priced asset.
The valuation logic of AI companies has become very similar to that of SaaS companies in 2021. Back then, everyone competed on ARR; today, they compete on compute commitments. At their core, both rely on discounting the future to value the present—the only question is whether the future will materialize.
What is OpenAI doing?
In the same 8-K filing in which Amazon invested in Anthropic, OpenAI also committed to consuming approximately 2 gigawatts of Trainium compute power via AWS infrastructure, ramping up starting in 2027.
Two months ago, Amazon invested $50 billion in OpenAI and signed a $100 billion cloud computing contract.
The script is exactly the same.
In other words, the three major cloud providers and the two leading model companies have played the same game multiple times, each time accompanied by headlines like “biggest in history,” “strategic partnership,” and “compute revolution.”
Behind every transaction is the same money circulating.
So, who will stop first?
It’s not the cloud providers—they currently rely on this narrative for their market valuation. Alphabet has raised its 2026 capital expenditure guidance to as much as $190 billion; such massive spending must be offset by revenue from companies like Anthropic and OpenAI, or Wall Street won’t accept it.
It also wouldn't be a model company—stopping means failing to secure the next round of funding, which means death.
The first to be eliminated may be the second-tier players who didn’t pick the right side.
Will the concert be canceled?
All of this vulnerability is hidden in the words “cash out.”
TPU will be launched in 2027. If Claude’s commercialization fails to keep pace with the expansion of computing power, what will Anthropic use to absorb this $200 billion?
If a contract is renegotiated, canceled, or split, Google Cloud’s $462 billion backlog immediately falls apart.
But today, no one is willing to be the first to speak up. CFOs are issuing guidance, analysts are publishing buy ratings, and CEOs are carefully choosing their words during earnings calls. Everyone is betting that they’ll be the first to secure a seat before the music stops.
It’s no longer a question of whether there’s a bubble—it’s a question of how to defuse it. Everyone knows this is a circular trading scheme, but everyone also knows that as long as the AI story continues, no one dares to short the backlog.
The contract is written on paper, money flows between three companies, and valuations swing between primary and secondary markets. Everyone holds a "promise of the future," and everyone treats that promise as a "current asset."
Until one day in the future, a company’s earnings report falls short of expectations. At that moment, $200 billion will suddenly have another name: liability.
And until that day arrives, the celebration will continue.
