Author: Ada, Deep Tide TechFlow

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

This multi-year agreement starting in 2027 will account for more than 40% of the revenue backlog for Google Cloud, a metric that reflects contract commitments from enterprise clients.

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

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

But what is more interesting is another number. Alphabet also made a reverse investment in Anthropic of up to $40 billion.

Money flows out of Google's accounts, makes a round, and comes back to Google's accounts, with an additional accounting line for "Anthropic's computing expenses."

So, is this the largest cloud computing order in history, or the most dignified financial magic in history?

A "Exclusive Commitment" Not Just for Google

To understand the essence of this deal, let's first look at a set of not isolated data.

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

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

This means, Google: invest $40 billion, receive $200 billion. Amazon: invest $33 billion, receive $100 billion+. Microsoft: invest $5 billion, receive $30 billion.

The three cloud giants collectively forked out about $78 billion in exchange for $330 billion in "contract commitments," resulting in a net inflow of $250 billion.

The essence of this play is to wash capital expenditures into revenue. The money invested in Anthropic is counted as cash flow from investing activities, while the fees paid by Anthropic for computing power are counted as main revenue. The same amount of money moves from the left pocket out and comes back into the right pocket, resulting in a nice backlog on the financial statements.

While Alphabet provides lifeblood to Anthropic, it simultaneously counts Anthropic's computing purchases as future revenue, thus creating a self-reinforcing closed loop of AI infrastructure prosperity.

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.

The Advanced Version of the Flywheel

The story of Strategy's high-level investment is not over, while the AI circle has amplified the same flywheel by a thousand times.

The logic of Strategy is to issue stocks to raise money, buy Bitcoin, and as the price rises, increase market capitalization, issue more stocks, and buy more coins.

On the other hand, the logic of cloud vendors is to invest in AI companies; the AI companies pay for computing power, revenue grows, stock prices rise, the capital market increases its bets, and continues to invest in AI companies.

The difference is, Bitcoin is a scarce asset, each one corresponds to a real supply on the chain. Computing power is not. The "multi-gigawatt TPU capacity" that will go online in 2027 hasn't even been installed in cabinets today.

This means that a significant portion of the $200 billion consists of Anthropic's advance commitment to purchase a batch of chips that have not yet been manufactured, and Google uses this commitment to persuade the capital market.

Isn’t this just a forward contract? The difference is that commodity futures have a delivery date and margin, but this contract does not. What if Anthropic can't pay this amount in 2027? Who bears the cost of default?

It won’t be Google. It has already included the backlog in the PPT for the earnings call. During the earnings call on April 29, Alphabet revealed that Google Cloud revenue grew by 63% year-over-year, exceeding $20 billion, and the cloud business backlog reached around $462 billion. This number supports Alphabet's current market value.

It won’t be Anthropic either. It just needs to continue fundraising, as the next round of valuations will still be increasing.

The ones left holding the bag may be retail investors who think they have bought into the "AI shovel sellers" story.

$5 billion Levers $330 billion

Is Anthropic's own scale worthy of this number?

According to media reports, Anthropic's annualized revenue is expected to rise from $1 billion to $5 billion by 2025.

A company with an annualized revenue of only $5 billion has signed contracts totaling $330 billion: a $200 billion contract over 5 years, a $100 billion contract over 10 years, plus a $30 billion contract.

Even if Anthropic's revenue were to increase tenfold, it wouldn't reach $330 billion in five years.

So, where is the money coming from?

There is only one path: continue to raise funds.

The largest potential investors are precisely these three cloud vendors themselves.

This is the secret to the cycle. Anthropic doesn't need to actually make money; it just needs to maintain the state of "always raising funds," treating every new round of funding as the next year's computing bill. As fundraising valuations rise, it can raise even more funds.

Who does this sound like?

Strategy. It also does not need Bitcoin to actually generate cash flow; it just needs to maintain the state of "always able to issue stocks and bonds." The only difference is that Strategy's balance sheet includes Bitcoin, a globally publicly priced asset.

The valuation logic of AI companies resembles that of SaaS companies in 2021. Back then, everyone was focused on ARR; today, the focus is on computing commitments. The essence is the same: discounting the future to the present, and the only question is whether the future will be fulfilled.

What is OpenAI Doing

In the same 8-K document where Amazon increased its investment in Anthropic, OpenAI also committed to consuming approximately 2 gigawatts of Trainium computing power through AWS infrastructure, beginning its ramp-up 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.

This means that the three major cloud vendors and the two major model companies have played this same game several times. Each time comes with titles like "the largest ever," "strategic cooperation," "computing power revolution."

Each time, the same money is circulating.

So, who will stop first?

It won’t be the cloud vendors; their current market value relies on this narrative. Alphabet has raised its capital expenditure guidance for 2026 to up to $190 billion; expenditures of this scale must have Anthropic and OpenAI to "hedge" into revenue, or else Wall Street will be the first to object.

It won’t be the model companies either; stopping would mean no more funding for the next round and ultimately, death.

The first to be kicked out may be those secondary players that haven't aligned themselves properly.

Will the Music Stop?

The fragility of all of this lies in the two words "fulfillment."

TPUs will launch in 2027. If Claude's commercialization cannot keep pace with the expansion of computing power, how will Anthropic digest this $200 billion?

If a contract is renegotiated, canceled, or distributed, Google Cloud's $462 billion backlog will immediately reveal its flaws.

But today, no one wants to be the first to poke the bubble. CFOs are writing guidance, analysts are writing buy ratings, and CEOs are carefully choosing their words during earnings calls. Everyone is betting that before the music stops, they will have positioned themselves closest to a chair.

This is no longer a question of whether there is a bubble, but rather how to dismantle the bubble. Everyone knows this is a circular trade, but everyone also knows that as long as the AI narrative continues, no one dares to short the backlog.

Contracts are written on paper, money flows among the three companies, and valuations circulate between primary and secondary markets. Everyone has taken a "future commitment," and everyone treats this commitment as "current assets."

Until one day in the future, one company's financial report fails to meet expectations. At that moment, the $200 billion will suddenly have another name, perhaps as a liability.

And before that day arrives, the revelry will continue.