Original Author: Ada, TechFlow

On May 5th, according to The Information, Anthropic has committed to paying Google Cloud $200 billion over the next five years.

Starting in 2027, this multi-year agreement will account for over 40% of Google Cloud's revenue backlog, an indicator that reflects contractual commitments from enterprise clients.

An AI company that didn't even exist five years ago, with a single contract, has consumed nearly half of Google Cloud's future revenue.

On the day the news broke, Alphabet's stock rose 2% in after-hours trading.

But another number is more intriguing. Alphabet simultaneously invested up to $400 billion back into Anthropic.

Money leaves Google's books, makes a full circle, and returns to Google's books, adding a new accounting line item labeled "Anthropic computing expenditure" in between.

So, is this the largest cloud computing order in history, or the most elegant accounting magic trick ever performed?

An "Exclusive Commitment" Not Just for Google

To understand the essence of this deal, we first need to look at a set of non-isolated data points.

On April 20th, Anthropic announced an expanded partnership with Amazon, committing over $100 billion in AWS technology spending over the next 10 years in exchange for up to 5 gigawatts of computing power. In return, Amazon added up to $25 billion in new investment on top of its existing $8 billion investment.

Last November, Microsoft agreed to invest up to $5 billion in Anthropic, while Anthropic committed to purchasing $30 billion worth of Azure computing power.

So, the tally is: Google: Invests $40 billion, receives $200 billion. Amazon: Invests $33 billion, receives over $100 billion. Microsoft: Invests $5 billion, receives $30 billion.

Collectively, the three cloud giants have dished out about $78 billion in exchange for $330 billion in "contractual commitments," resulting in a net book inflow of $250 billion.

The essence of this strategy is to wash capital expenditure as revenue. The investment in Anthropic is recorded under investing cash flow, while the computing fees paid by Anthropic are counted as main operating revenue. It’s the same money moving from the left pocket to the right, creating a beautiful backlog on the financial statements.

Alphabet, on one hand, infuses capital into Anthropic, and on the other hand, books Anthropic's computing procurement as future revenue, creating a self-reinforcing closed loop for the AI infrastructure boom.

Wall Street is the true winner in this game. As long as the backlog numbers are large enough, price-to-earnings ratios can be sustained.

The Advanced Version of the Flywheel

Before the story of Strategy doubling down at market peaks is even over, the AI circle has scaled the same flywheel a thousand times over.

Strategy's logic was: issue stocks to raise capital, buy Bitcoin, Bitcoin rises boosting market cap, issue more stocks, buy more Bitcoin.

The logic for the cloud vendors is: invest in AI companies, AI companies pay for computing power, revenue grows, stock prices rise, capital markets add bets, continue investing in AI companies.

The difference is that Bitcoin is a scarce asset; every coin corresponds to a real supply on the chain. Computing power is not. The "multi-gigawatt TPU capacity," which won't be operational until 2027, doesn't even have its server racks installed today.

In other words, a significant portion of that $200 billion represents Anthropic's advance commitment to purchase a batch of yet-to-be-manufactured chips, and Google uses this commitment to convince the capital markets.

Isn't this essentially a forward contract? The difference is that commodity futures have delivery dates and margin requirements, but this contract does not. If Anthropic truly cannot pay this amount in 2027, who bears the cost of default?

It won't be Google. It has already written the backlog into its earnings call presentations. During Alphabet's earnings call on April 29th, it disclosed that Google Cloud revenue grew 63% year-over-year, exceeding $20 billion, with the cloud business backlog reaching approximately $462 billion. This number props up Alphabet's current market capitalization.

It also won't be Anthropic. It can just continue raising funds; after all, its valuation in the next round will likely still be rising.

Ultimately, it might be the retail investors who think they're buying into the "AI shovel sellers" story who end up footing the bill.

$5 Billion Leveraging $330 Billion

Does Anthropic's own scale justify these numbers?

According to media reports, Anthropic's annualized revenue grew from $1 billion to $5 billion in 2025.

A company with an annualized revenue of only $5 billion signed contracts for $200 billion over 5 years, $100 billion over 10 years, plus another $30 billion, totaling $330 billion across three contracts.

Even if Anthropic's revenue were to multiply tenfold again, its cumulative earnings over 5 years wouldn't reach $330 billion.

So where does the money come from?

There is only one route: continue raising funds.

And the largest potential investors happen to be these three cloud vendors themselves.

This is the entire secret of the cycle. Anthropic doesn't necessarily need to be profitable; it just needs to maintain a state of "always fundraising," using the new funds from each round to pay for the next year's computing bills. As its fundraising valuation increases, it can raise even more.

Who does that sound like?

Strategy. It also doesn't need Bitcoin to generate cash flow; it only needs to maintain the state of being "constantly able to issue stocks and bonds." The only difference is that Strategy's balance sheet still holds an asset like Bitcoin, which has a global, publicly determined price.

The valuation logic for AI companies has become very similar to that of SaaS companies in 2021. Back then, everyone was competing on ARR; today, the competition is over computing commitments. Essentially, both use the future to discount the present. The only question is whether that future will materialize.

What is OpenAI Doing?

In the same 8-K filing where Amazon increased its bet on Anthropic, OpenAI also committed to consuming approximately 2 gigawatts of Trainium computing 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 identical.

In other words, three major cloud providers and two major model companies – five players have played the same game multiple times. Each time is accompanied by headlines like "Biggest in History," "Strategic Cooperation," and "Computing Revolution."

Behind each instance, it's the same money circulating.

So, who will be the first to stop?

It won't be the cloud vendors. Their current market valuations depend on this narrative. Alphabet has raised its 2026 capital expenditure guidance to up to $190 billion. Expenditures on this scale absolutely require Anthropic and OpenAI to "hedge" them into revenue; otherwise, Wall Street would be the first to object.

Nor will it be the model companies. Stopping means failing to secure the next round of funding, and consequently, death.

The ones most likely to be kicked out first are the second-tier players who haven't aligned themselves properly.

Will the Music Stop?

The fragility of it all is hidden in the word "materialization."

TPUs will go live in 2027. If Claude's commercialization fails to keep pace with the expansion of computing power by then, how will Anthropic digest that $200 billion?

If a contract gets renegotiated, canceled, or distributed, the facade of Google Cloud's $462 billion backlog will immediately reveal itself.

But today, no one wants to be the first to pop the bubble. CFOs are writing guidance, analysts are writing buy ratings, and CEOs are choosing their words carefully on earnings calls. Everyone is betting that they will be standing closest to the chair when the music stops.

The issue now isn't whether it's a bubble, but how to dismantle the bubble. Everyone knows this is circular trading, but everyone also knows that as long as the AI story continues, no one dares to short the backlog.

Contracts are on paper, money circulates among the three companies, and valuations bounce between primary and secondary markets. Everyone holds a "future promise," and everyone treats that promise as a "current asset."

Until one day in the future, when a company's earnings fail to meet expectations. At that moment, that $200 billion will suddenly acquire another name: contingent liability.

And until that day arrives, the party will continue.