Meta Is Building a Cloud Business. Here's What That Does to Its Stock.

On July 1, 2026, a single Bloomberg report changed how the market was thinking about Meta Platforms. The story said Meta was developing plans for a cloud infrastructure business to sell AI computing power and hosted AI models to outside customers. On that news Meta's stock closed at $612.91, up 8.81%, marking one of its strongest single-day gains of 2026, and its best day since January, according to WSJ, on volume 159% above the three-month average, as per GuruFocus and Intellectia.AI. 

Meanwhile, a cluster of AI cloud companies got hit hard: CoreWeave (CRWV) closed at $85.69, down 13.92%, and Nebius Group (NBIS) slid 17.01% to $229.18, while Super Micro Computer (SMCI) declined 5.73%. IREN dropped approximately 5.27% in the session. Nvidia, Micron, Broadcom, AMD, and others also closed in the red. All of this because of a plan that has not officially launched yet.

Let's break down what Meta Compute actually is, why it hit the neocloud sector so hard, and what it could realistically mean for Meta's valuation from here.

What Is Meta Compute? Bloomberg’s Reported Meta AI Cloud Plan Explained

An internal organization called Meta Compute sits at the center of these efforts. Meta Compute is led by Santosh Janardhan, Meta's head of infrastructure; Daniel Gross, a leader inside the Meta Superintelligence Labs AI unit; and Meta President Dina Powell McCormick.

That combination of infrastructure, AI research, and senior business leadership matters. This is not a skunkworks experiment or a slide deck for the earnings call. It has named leadership with real organizational weight behind it.

The plans being considered involve two tracks. One possible plan would involve offering access to various AI models hosted on Meta's existing AI infrastructure, similar to what AWS does with its Bedrock. Meta would run the data centers and chips powering the models, charging developers to access them. The company is also weighing a plan to sell access to "raw" computing capacity, similar to what "neocloud businesses" like CoreWeave offer.

Think of Track 1 this way: imagine you want to use Meta's AI models, say, its Muse Spark line, but you don't want to build your own servers for it. Meta runs the servers, you pay per query. That's AWS Bedrock's model. Track 2 is more like renting the actual machinery itself; raw GPU compute by the hour, the way CoreWeave and Nebius have built their entire businesses.

No pricing, launch date, or customer pipeline has been confirmed publicly. A Meta spokesperson declined to comment. The plans could still change.

Why Meta’s $145 Billion AI Capex Matters for Investors

Meta has raised its 2026 capital expenditure forecast to between $125 billion and $145 billion, highlighting the company's commitment to AI infrastructure, prompting investors to assess whether this will yield sustainable financial returns.

For context, Meta's FY2025 capex was approximately $72 billion. That means 2026 spending nearly doubles in a single year. This had been a persistent overhang on the stock. Meta stock had its strongest single-day move of 2026 on July 1, and the catalyst was a single Bloomberg report that reframed how investors are thinking about the company's entire AI spending program. The question was never really "why is Meta spending so much?" It was always "what will Meta get back from it?"

Meta has guided to an enormous outlay on chips, land and power, and a cloud business is one of the few ways to turn idle capacity from that build-out into revenue instead of a sunk cost.

This is the core of what changed on July 1. Not the prospect of near-term cloud revenue, that's years away. What changed is the investment thesis. The market stopped asking "is this money wasted?" and started asking "how fast can the returns come?"

Here is a simple way to think about the shift. Before July 1, Meta's $145 billion capex was viewed almost entirely as a cost. After July 1, investors were forced to consider that a portion of it could become an asset that generates external revenue. That is a meaningful narrative change, even if the numbers are not there yet.

Is Meta Following Amazon’s AWS Playbook With AI Cloud?

This story has a precedent. Amazon built enormous server infrastructure to run its own e-commerce operations. In 2006, it started renting that capacity externally under the AWS brand. Today, AWS generates over $100 billion in annual revenue and is Amazon's most profitable business segment. Jefferies analyst Brent Thill said the move represents Meta following Amazon's AWS playbook by monetizing excess compute to lift utilization, improve return on invested capital, and boost cash flow to fund more capital expenditures.

Thill added that the overbuild narrative is backward and Meta is actually late to this cloud strategy, not ahead of it.

Meta CEO Mark Zuckerberg had hinted at this idea earlier, saying that outside companies had approached Meta about API access and buying compute capacity at a premium. Almost every week, he said, "there are different companies that come to us from outside asking us to both stand up an API service, or asking if we have compute that they could buy from us at some premium to what we've bought it at." He added that Meta had not acted on this yet because it still needed the compute internally, but acknowledged excess capacity would change that equation.

For context think about Reliance Industries and Jio. Reliance spent over $25 billion building Jio's nationwide fiber and cell tower infrastructure starting around 2016, mostly to launch a consumer telecom service. Over time, that same infrastructure became the backbone of JioCloud, JioFiber for enterprise, and various B2B cloud offerings. The initial capex looked like pure cost. With scale, it turned into a multi-revenue-stream asset. Meta is attempting a similar conversion: a capex pile that was built for one purpose (training and running Meta's own AI models) now potentially becoming a revenue-generating platform for others.

The key difference from Jio: Meta is entering a market already dominated by three very entrenched giants.

Why CoreWeave, Nebius and AI Cloud Stocks Fell After Meta’s News

The selloff in neocloud names was not random panic. It was a rational (if harsh) repricing of competitive risk.

Sources: Motley Fool, Globe and Mail, GuruFocus, Yahoo Finance (July 1, 2026).

The logic behind the CoreWeave and Nebius selloffs runs like this. CoreWeave's agreement with Meta reportedly totals around $21 billion, alongside separate partnerships involving OpenAI and Anthropic, making Meta one of its most significant commercial relationships. Nebius holds a reported $27 billion contingent deal with Meta, including $12 billion in guaranteed capacity starting in 2027, and Meta's new direction raises questions about how much of that arrangement will ultimately materialize.

The market reaction reflects a simple concern: Meta is not just a customer anymore, it could become a competitor.

Bernstein struck the most bearish tone on CoreWeave, calling the development "problematic" and warning that hyperscaler competition in the neocloud space was always inevitable.

This is the customer-to-competitor dynamic in its rawest form. CoreWeave and Nebius built massive businesses partly because Meta itself was willing to rent compute from them, as Meta had not yet built enough of its own. Now Meta is building exactly that. Even if the existing contracts hold, the market is re-assessing what these companies' backlogs will look like two or three years out.

How Meta’s AI Cloud Plan Could Change Its Stock Valuation

Before July 1, analysts were predominantly modeling Meta as an advertising business with a massive, unmonetized AI capex line. That framing justified a discount relative to, say, Microsoft or Alphabet, both of which have diversified revenue streams beyond advertising. Meta earns 97% of revenue from digital advertising, making an AI cloud business a transformative diversification into one of tech's most profitable markets. AWS, Azure, and Google Cloud generate a combined $300 billion annually, a market Meta could enter as the fourth hyperscale cloud provider.

The consensus analyst picture as of early July 2026, as per TipRanks (37 analysts, last 3 months): Meta Platforms has a consensus rating of Strong Buy based on 31 buy ratings, 6 hold ratings, and 0 sell ratings. The average price target is $815.82, with a high of $1,015 and a low of $622.25.

Analysts from BMO Capital reiterated a Market Perform rating on Meta with a price target of $720.00, acknowledging the positive market reaction to these developments. Jefferies reiterated its Buy rating. The broader consensus had not yet fully updated to incorporate the cloud angle, that re-modeling will likely play out in the July 28 earnings call, when Q2 results are reported and management has a chance to add color.

A Simple Valuation Framework: Can the Math Support Meta’s Rally?

Here is a framework that is anchoring on known numbers and layering in reasonable scenarios.

Step 1: What Is Meta Worth Without Cloud?

Meta's stock, at roughly $612 after the July 1 close, is trading close to a 20x multiple on ~$30 consensus EPS. That is cheap relative to its own history. Meta's five-year median P/E has been approximately 26x, as per GuruFocus. The market was applying a discount for the capex overhang and the absence of a clear return path. That discount was partial justification for the stock being down roughly 23% from its late-2025 high of ~$796 before this news hit.

Step 2: What Does Cloud Add to the Picture?

The hyperscaler cloud market (AWS, Azure, Google Cloud combined) is approximately $300 billion annually, as per 24/7 Wall St. Capturing even a small fraction of that market moves the needle on earnings modestly, but the bigger implication is what it does to the multiple. A company with both advertising and cloud revenue is a structurally different business than a company with 97% advertising dependence, and the market prices that difference.

These are estimates, not forecasts. They assume Meta builds out the offering, finds enterprise customers, and operates at margins comparable to what cloud providers have achieved over time. None of these is guaranteed.

Step 3: The Multiple Re-Rating

A diversified tech company with both advertising and cloud revenue should, logically, trade at a higher multiple than a pure-play social media company. This is not guaranteed either, it depends on execution, but it is the direction the market would move if cloud revenue materializes.

• At $30 EPS (FY2026) and a 25x multiple: $750.
• At $36-38 EPS in FY2027, depending on the starting FY2026 EPS base, a 25x multiple implies roughly $900-950.
• Add cloud EPS uplift of approximately $0.75 in the base case, at 25x, that is roughly $19 of additional stock value, getting you to around $970. The path to and beyond $1,000 is not primarily a cloud EPS story. It depends on whether the market re-rates Meta from 20x toward 26-28x as the business diversifies beyond advertising; the same re-rating that happened to Amazon once AWS became undeniable.

The analyst consensus target of ~$820-840 is essentially a 25-28x multiple on FY2026 earnings with no cloud revenue modeled in. If cloud becomes real, the ceiling moves higher. If it stays a press release, the core advertising business at 25x still supports the stock in the $750 range.

The math does not suggest Meta is cheap on a pure-multiples basis today, after the July 1 rally. At $612, it is fairly valued to slightly discounted relative to its own history. The cloud optionality is genuinely additive, but it needs to move from announcement to revenue to matter to the stock in the near term.

The Bear Case: Risks to Meta’s AI Cloud Business

Meta is entering a market with three incumbents that have spent decades building enterprise relationships, compliance frameworks, and developer ecosystems. AWS, Azure, and Google Cloud did not get to $300 billion in combined annual revenue by being easy to displace.

Some skeptics have warned the race to build out AI infrastructure is creating a bubble that leans heavily on rapidly depreciating chips. If AI demand cools, excess compute becomes harder to monetize for everyone, including Meta.

There is also the question of whether Meta's culture and sales DNA is suited to selling enterprise cloud services. The company has spent its entire history selling to advertisers and end consumers. Enterprise cloud is a different sales motion: longer cycles, more complex procurement, security and compliance requirements. This is not insurmountable, but it is not trivial.

This move may also signal Meta's flagging prospects when it comes to generating significant income with its own AI initiative. That is a bear-case reading: if Meta's internal AI projects (including its superintelligence ambitions) were consuming all the compute it was building, there would be no excess to sell. The existence of a cloud business plan implies there is more capacity than Meta can use internally, which raises questions about whether AI demand at Meta itself is meeting internal targets.

And Reality Labs continues to be a drag. The unit has accumulated over $83 billion in cumulative losses, with no clear profitability timeline disclosed.

Our Take: Should Investors Watch Meta Stock After the AI Cloud News?

We want to be clear about what this announcement actually is and is not. It is a plan, confirmed by Bloomberg sources, backed by named leadership, and telegraphed by Zuckerberg himself at Meta's May shareholder meeting. It is not a launched product, a signed enterprise contract, or a disclosed revenue number.

That said, the narrative shift is real. The central concern about Meta heading into mid-2026 was whether $145 billion in AI capex would ever generate a return beyond improved ad targeting. That concern has not been eliminated, but it has been partially answered. If Meta's infrastructure becomes the foundation of a cloud business, the spend calculus changes. The fixed costs stay roughly the same; the revenue potential expands.

The valuation picture at around $612 is interesting without being a screaming bargain. The advertising business, growing at 33% YoY as of Q1 2026, is intact and strong. Advertising revenue increased to $55.02 billion in Q1 2026, ad impressions rose by 19%, and the average price per ad increased by 12%. The core engine is not broken. The cloud initiative, if it materializes at even a base-case scale, is incrementally positive for earnings and for the multiple the market assigns to those earnings.

The risk is execution. Announcing a cloud business and building one that competes with AWS are two very different things, separated by years of enterprise sales, compliance work, and product development. 

Investors watching Meta should focus on what comes next: does management quantify the cloud opportunity on the July earnings call? Do they announce initial enterprise customers? The answers to those questions will matter more to the stock than the Bloomberg report itself.