Meta Enters the Market to Sell Computing Power, Cloud Computing Landscape Shifts

Author: TechFlow Research

On July 1, Bloomberg broke a major story: Meta is assembling a cloud computing business unit, preparing to sell its excess AI computing power to external customers.

After the news broke, the market reaction was immediate, but the divergence was extremely sharp. Meta itself surged over 8% in pre-market trading, while the two flagships of the "new cloud computing" track, CoreWeave and Nebius, plummeted 6% and 10% respectively. Amazon also turned down in pre-market trading in response. On one side was revelry, on the other panic; this news was like a scalpel, precisely cutting open the interest dividing line on the AI computing power industry chain.

The most ironic part is: Meta recently signed a $21 billion computing power procurement agreement with CoreWeave, and another cooperation deal with Nebius worth up to $27 billion. Now it is turning around to compete for business with its own suppliers.

What is exactly going on?

According to Bloomberg citing informed sources, a department within Meta named "Meta Compute" is leading this plan. This department is jointly led by three executives: Meta Head of Infrastructure Santosh Janardhan, Daniel Gross of Meta Superintelligence Labs, and Meta President Dina Powell McCormick.

There are two paths for the business model currently under consideration:

The first is "Model-as-a-Service" (Model-as-a-Service), allowing external developers to pay to call AI models hosted on Meta infrastructure, including Meta's self-developed Muse Spark model. The benchmark product for this path is AWS's Bedrock service; essentially, it opens up the AI model inference capability that Meta built at an astronomical cost as a paid API.

The second is more aggressive, directly renting out bare GPU computing power. This is exactly what CoreWeave and Nebius are doing, and it is precisely this path that caused the stocks of these two Neocloud companies to crash instantly. When your biggest customer announces they are going to do the exact same business as you, your investors will likely run for the hills first.

Meta has not yet issued an official comment on this matter.

A $145 Billion Computing Power Gamble Needs an "Insurance Policy"

To understand the core logic of this news, one must first look at a set of numbers.

In 2026, Meta's AI-related capital expenditure guidance is $125 billion to $145 billion, raised again from the previous guidance of $115 billion to $135 billion. What concept is this number? It is only slightly lower than Google parent company Alphabet's $175 billion to $185 billion, Microsoft's $190 billion, and Amazon's $200 billion. The combined expenditure of the four major tech giants on AI infrastructure this year will exceed $700 billion.

More importantly is Meta's unique situation. Among these four companies, Amazon has AWS, Microsoft has Azure, and Google has Google Cloud; they all have mature cloud computing businesses to digest AI infrastructure investments and can directly sell computing power to customers. Only Meta does not. All its data centers and GPU clusters, theoretically, serve only its own social platforms, advertising systems, and AI R&D.

This creates a huge risk exposure: if Meta's internal demand growth for AI computing power falls short of expectations, that $145 billion in capital expenditure becomes sunk cost. Data centers are built, GPUs are bought, long-term power contracts are signed; these are "rigid" investments that cannot be reduced at any time like adjusting a marketing budget.

Zuckerberg personally responded to this concern at the shareholder meeting on May 27. He said the cloud computing business is "definitely on the table," and revealed that "almost every week external companies come to us, either asking if we can open API services, or asking if we can sell computing power to them at a premium."

To translate Zuckerberg's subtext: We are not afraid of spending so much money. If all AI computing power is used, the return on this investment is reflected through our products; if there is excess capacity, we can also sell it to make money. We can't lose either way.

What Wall Street has always worried about most is "Meta spending too much money on AI but seeing no return"; the cloud computing business equals giving investors a safety cushion: the $145 billion in capital expenditure is no longer purely risk investment, but has become a two-sided bet that can advance to attack or retreat to defend.

The Survival Crisis of Neocloud

But the revelry of Meta investors is a nightmare for CoreWeave and Nebius.

To understand this relationship, first look at a key fact: the entire business model of Neocloud companies is to manage GPU computing power on behalf of tech companies that do not do cloud computing themselves. The core logic why CoreWeave and Nebius were able to secure sky-high contracts is "Meta has huge AI computing power demand, but does not have its own cloud business to digest excess capacity, so it needs to rent from external sources."

Now Meta says, I am preparing to do it myself.

This impact is structural. CoreWeave currently has nearly $100 billion in revenue backlog orders, a significant portion of which comes from Meta and other large AI companies. The $27 billion contract Nebius signed with Meta includes reserving $12 billion worth of GPU capacity for Meta starting from early 2027. If Meta's self-built capacity begins to replace external leasing, the conversion rate of these orders becomes questionable.

The deeper problem lies in the fact that the position of Neocloud companies in the AI industry chain was already fragile. What they provide is essentially a "computing power intermediary" service, buying GPUs from Nvidia, building data centers, and then selling them to AI companies at a markup. This model can make huge profits when GPUs are in short supply, but when supply bottlenecks ease and large customers start building their own, the value of the intermediary will be rapidly compressed.

CoreWeave's Q1 revenue this year was $2.078 billion, up 168% year-over-year, but net loss was $740 million, with the loss amount doubling year-over-year. Its total debt has exceeded $25 billion. Nebius Q1 revenue was $399 million, surging 684% year-over-year, performance was bright, but it also lost over $100 million, with total debt exceeding $9.5 billion. Both companies are still using high leverage to exchange for high growth. If the biggest customer becomes a competitor, this "borrowing money to expand production" model becomes exceptionally dangerous.

The market is already voting with its feet. In June, CoreWeave's short interest ratio reached 14%, and Nebius was even higher at 20%. Investor confidence in the Neocloud track is shaking.

The Traditional Three Cloud Giants Cannot Rest Easy Either

Meta entering cloud computing is also not good news for AWS, Azure, and Google Cloud.

The global cloud infrastructure market reached a quarterly scale of $129 billion in Q1 2026, with an annualized growth of 35%, moving towards an annual revenue of $500 billion. This market has long been divided among AWS, Azure, and Google Cloud, collectively occupying over 60% of the market share.

If Meta officially enters the field, it will break this tripartite pattern. Moreover, Meta has several unique advantages: it operates one of the largest social networking platforms in the world, has massive practical experience in AI models and application scenarios; it has already established a strong developer ecosystem in the open-source AI field (Llama series models); its AI infrastructure investment scale is already close to the three major cloud giants.

Of course, cloud computing is not just hardware. The reason AWS dominates the market is not just because it has data centers, but because it spent nearly 20 years building a complete product system: from computing, storage, databases to machine learning, security, IoT, there are over 200 services. For Meta to build a product matrix of the same level from scratch, it needs to invest huge engineering resources and time.

But Meta's strategy may not be to fully replicate AWS. A more realistic path is to focus on the AI computing power segment, which is the hottest and fastest-growing. If Meta only does these two things "AI Computing + Model Services," its starting barrier is much lower, and it cuts precisely into the most profitable part of the cloud market.

Seeking Alpha noted that after the Bloomberg news broke, Amazon's stock price turned from rising to falling in pre-market trading. AWS is Amazon's most profitable business, with Q1 2026 cloud revenue growing 28% year-over-year, the fastest growth rate in 15 quarters. Any new entrant dividing the cloud market will make AWS investors nervous.

Selling $600 Billion "Surplus" Too: The Deep Logic of Meta's Computing Power Strategy

There is a detail worth savoring repeatedly.

Meta announced a "Meta Compute" plan this January, aiming to accumulate "tens of GW" of computing power capacity within this decade, and long-term looking at "hundreds of GW or even more." Currently Meta operates over 30 data centers, and the AI-optimized facility capacity under construction ranges from 1GW to 5GW. In June, it also signed a 1.6GW computing power procurement contract with data center company Crusoe.

What do these numbers add up to mean? Meta is building AI infrastructure on the scale of a "supercomputing nation."

This brings out another broader background: the real bottleneck of the AI industry in 2026 is neither chips nor capital, but electricity. Just a few days ago, reports pointed out that Google had to limit Meta's access to its models because it could not provide enough Gemini computing power. Google Cloud itself has over $460 billion in signed but undelivered contract backlog. Even the richest tech companies on Earth cannot buy enough computing power, not because of lack of money or chips, but because of lack of electricity.

Against this background, Meta acting as both buyer and seller has another layer of strategic meaning: whoever locks in electricity and data center capacity first possesses a structural advantage in the AI race. The infrastructure Meta built spending $145 billion is both its weapon to catch up to AI superintelligence, and a "strategic reserve" that can be monetized externally.

Several Key Judgments

The market impact of this news will continue to ferment over the next few months, and there are several dimensions worth paying attention to:

For Meta itself, if the cloud computing business lands, it will open a brand new revenue source for it, reducing excessive reliance on the advertising business. Currently, over 99% of Meta's revenue comes from advertising. Even if the cloud business initially accounts for only a few percentage points of revenue, its symbolic significance and valuation effect cannot be underestimated.

The uncertainty facing the Neocloud track has risen significantly. The investment logic of CoreWeave and Nebius is built on the premise that "large tech companies need to rent external GPU computing power." If Meta's self-built cloud business succeeds, other tech giants may also follow suit, and the long-term survival space of Neocloud will be compressed. Of course, in the short term, the supply and demand gap for AI computing power is still huge, and the signed contracts of Neocloud companies also provide certain revenue certainty. But valuations need more margin of safety.

The bigger picture is: the AI industry is moving from the stage of "frantically burning money to build infrastructure" into the stage of "how to make infrastructure investments generate returns." Meta selling computing power, Open Standard issuing OUSD, major banks deploying stablecoins, behind these seemingly unrelated events points to the same logic: when the investment scale reaches a certain level, capital itself will seek all possible monetization paths.

For the AI arms race, Meta's move is actually telling the world: $145 billion is not a gamble, it is infrastructure investment. The characteristic of infrastructure is that once built, you can charge everyone.

Disclaimer: This article is for information reference only and does not constitute investment advice. Tech stocks and related investments carry high risks, please judge for yourself.