By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
Panelists are neutral to bearish on Meta's cloud expansion plans, citing high execution risk, thin margins, and dependence on internal AI workloads to justify massive capex. They agree that Meta's core ad business growth is the primary return driver.
Risk: High execution risk in pivoting to a B2B cloud provider, with lack of legacy sales force, support infrastructure, and developer ecosystem.
Opportunity: Potential moat from proprietary Llama models becoming the industry standard for enterprise AI inference engines.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
For months, the knock on Meta Platforms (NASDAQ: META) hasn't been its business. It has been the bill. When management raised its 2026 capital expenditures guidance in April to a range of $125 billion to $145 billion, shares sank on the news. And heading into Wednesday, the stock was down nearly 15% for the year, sitting well below its 52-week high of $796.25 -- even as the company reported accelerating growth.
Then investors got a look at what could become the other side of that spending story. Bloomberg reported Wednesday that Meta is developing plans for a cloud business that would sell access to artificial intelligence (AI) computing power and models, putting the social media company in competition with the cloud units of Amazon and Microsoft. Shares jumped 8.8% to $612.91.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now, when you join Stock Advisor. See the stocks »
The reaction makes sense. But can renting out computing capacity actually change the return math on one of the biggest capital spending programs in corporate history?
According to Bloomberg, the effort is internally called Meta Compute, and the company is debating two approaches: giving developers access to AI models hosted on Meta's infrastructure, or selling raw computing power. The report said the plans are still in development and could change. And it's worth emphasizing that Meta hasn't announced anything.
Still, the idea isn't coming out of nowhere. CEO Mark Zuckerberg said in May that selling excess computing capacity was "definitely on the table" if Meta ends up building more data center capacity than it needs, according to the report.
Overbuilding is precisely the worry that has weighed on the stock. Meta's spending plan for 2026 -- raised in April from a prior range of $115 billion to $135 billion -- compares to $72.2 billion in capital expenditures in 2025. In other words, spending could double this year.
And unlike Amazon, Microsoft, and Alphabet, Meta has no cloud computing business renting its infrastructure to outside customers. Every dollar of return on those data centers has to come from Meta's own products, mainly advertising. If the company builds more capacity than its apps and AI ambitions need, the excess earns nothing. A cloud business would change that equation, turning idle capacity into revenue -- and giving Meta a source of sales beyond the ad market.
What's easy to miss in the debate over Meta's spending is that the core business is accelerating -- a sign its AI investments are already generating returns. First-quarter revenue rose 33% year over year to $56.3 billion -- a step up from 24% growth in the fourth quarter of 2025 and 22% growth for the full year.
In addition, the company is making progress on its efforts to build a superintelligence.
"We're on track to deliver personal superintelligence to billions of people," said Zuckerberg in the company's first-quarter earnings release.
None of this makes the reported cloud plan a sure thing -- or even a near-term one. Building an enterprise cloud business requires sales teams, support operations, and reliability commitments that Meta would be starting mostly from scratch, while Amazon and Microsoft have spent nearly two decades building exactly that. And selling raw computing capacity tends to carry lower margins than Meta's advertising business. So even if the reported plans turn into a product, it could take years for the revenue to matter.
In the meantime, the stock's valuation looks undemanding. After Wednesday's jump, Meta trades at about 21 times earnings -- a multiple that arguably still reflects the market's skepticism about all that spending rather than the company's growth.
To me, the stock is a buy -- just not because of Wednesday's report. An unconfirmed plan shouldn't be anyone's investment thesis. The better reason is the combination of an accelerating core business and a modest valuation. The reported cloud business is best viewed as a free option: If it launches, Meta gains a second way for its AI infrastructure to pay off. If it never does, buyers today still own one of the market's fastest-growing big tech companies at a reasonable price.
There are risks. Spending could climb even higher, and advertising demand can turn quickly in a weak economy. But among big tech's AI spenders, Meta now offers a rare combination: accelerating growth, a modest multiple, and if the report proves right, a new way to get paid for all those data centers.
Before you buy stock in Meta Platforms, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Meta Platforms wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $418,761! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,195,804!
Now, it’s worth noting Stock Advisor’s total average return is 918% — a market-crushing outperformance compared to 208% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of July 3, 2026. *
Daniel Sparks and his clients have no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, and Microsoft. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Meta's cloud plans are too preliminary and operationally distant to alter the risk-reward on its unprecedented capex program."
Meta's reported Meta Compute initiative remains vaporware with no product, sales force, or enterprise track record, while 2026 capex guidance of $125-145B already doubles 2025 levels. Selling raw GPU capacity or hosted Llama models pits Meta against AWS and Azure's decade-plus moats in compliance, SLAs, and account management. Revenue acceleration to 33% YoY is real but still 100% ad-driven, leaving the stock exposed if macro weakens or if excess capacity never materializes as sellable. The 21x earnings multiple embeds skepticism, not optimism.
Accelerating ad growth already validates the AI spend, so any cloud revenue would be pure upside without needing to displace hyperscalers.
"Meta's pivot to cloud infrastructure is a defensive reaction to high Capex that risks diluting their high-margin advertising model with lower-margin, capital-intensive enterprise services."
The market is celebrating a 'free option' on cloud revenue, but it is ignoring the massive operational friction of pivoting from a consumer-facing ad juggernaut to a B2B enterprise cloud provider. Meta lacks the legacy sales force, support infrastructure, and developer ecosystem that make AWS and Azure sticky. If they enter this space, they are essentially commoditizing their own infrastructure to compete with their primary GPU suppliers and partners. While the 21x forward P/E is attractive, the $125B+ Capex guidance suggests a margin profile that may compress if they are forced to compete on price in a crowded cloud market. I am neutral on META until we see if these margins can be defended.
If Meta successfully abstracts its massive Llama model ecosystem into a 'model-as-a-service' layer, they could bypass the need for a traditional cloud sales force and capture high-margin software revenue rather than low-margin raw compute.
"Meta's core business acceleration is real and underpriced at 21x forward earnings, but the cloud business is a speculative option being used to justify capex that must be justified by ad-business returns alone."
The article conflates two separate stories: Meta's core business acceleration (33% YoY revenue growth, real) and an unconfirmed cloud venture (speculative). The 8.8% pop on Wednesday reflects relief that management has *thought about* monetizing excess capacity—not evidence the business will exist or be material. Meta's capex doubling to $125-145B in 2026 remains a real cash drag. The cloud narrative is a valuation salve, not a thesis. What matters: does Meta's core ad business sustain 30%+ growth, or does it normalize toward 15-20%? That's the actual return driver. A cloud business, if it materializes in 2027-2028, would be gravy—but the article treats it as the justification for current spending, which inverts causality.
Meta has zero enterprise sales infrastructure, no SLA track record, and faces entrenched competitors with 15+ year head starts; even if 'Meta Compute' launches, it could cannibalize margin-rich ad revenue while generating commodity-priced compute revenue—a value-destructive trade-off the article never quantifies.
"Even with a potential cloud monetization, Meta’s huge capex and thin-margined external compute business likely leave a modest near-term ROI impact, making the cloud plan more of a long-horizon optionality than a near-term catalyst."
Meta’s 2026 capex guide of $125-145B implies a major commitment that could redefine the ROI of its AI bet, even before any cloud revenue materializes. The Bloomberg cloud plan—whether hosting models or selling raw compute—adds optionality, but the execution risk is high: cloud sales require enterprise sales, SLAs, and support ecosystems that Meta has not built, and margins on externally sold compute tend to be thinner than advertising. Overbuilding data centers risks a long payback, while usage shocks or weaker ad demand could tighten the free cash flow path. The stock already trades around 21x earnings, leaving little margin for missteps on a multi-year, capital-intensive pivot.
The countercase is that Meta could monetize idle capacity at disciplined prices, leveraging its AI-infrastructure to attract enterprise customers and achieve meaningful incremental ROIC if utilization ramps and margins stabilize. In that scenario, the cloud move becomes a valuable multiplier rather than a sunk cost.
"Internal AI demand may crowd out sellable cloud capacity, making external revenue secondary to Meta's own utilization rates."
Grok flags the ad-driven nature of Meta's growth but misses how sustained 30%+ revenue gains could internally justify the capex even if external cloud sales stay minimal. The overlooked linkage is utilization: if Meta's own AI workloads consume most new capacity through 2027, the 'excess' available for monetization shrinks, turning the 21x multiple into a bet on internal ROI rather than third-party revenue.
"Meta's capex is an investment in model dominance, not just commodity cloud compute, which creates a proprietary moat AWS cannot replicate."
Claude is right about the causality inversion, but everyone is missing the 'Llama' moat. Meta isn't building a generic cloud; they are building an inference engine for their own proprietary models. If Llama 4 or 5 becomes the industry standard for enterprise agents, Meta doesn't need a traditional sales force—they own the software stack. The capex isn't just for 'compute'; it's for training the foundational intelligence that will eventually gate-keep the entire enterprise AI ecosystem.
"Llama dominance solves the software problem, not the infrastructure-as-a-service problem—Meta still needs to build cloud operations from scratch."
Gemini's Llama-as-moat argument is seductive but conflates two different markets. Yes, if Llama becomes the standard, Meta owns software pricing power. But that's a *model* business, not a *cloud* business. The $125B capex is infrastructure spend—data centers, GPUs, cooling. Even if Llama dominates, Meta still competes on inference latency and availability against Azure/AWS running the same Llama weights. The capex doesn't automatically translate to defensible cloud margins unless Meta also builds the sales, support, and compliance machinery everyone else flagged. Gemini is right that the software layer matters, but wrong that it eliminates the need for enterprise infrastructure competence.
"Utilization visibility determines ROI on 125-145B capex; without strong internal demand or credible cloud monetization, the investment risks becoming a speculative drag rather than a moat."
Claude's warning about cloud margins is valid, but it understates the risk that Meta's vast capex is only justifiable if internal AI workloads soak up most capacity. If ad growth decelerates or cloud demand never materializes, the company faces a prolonged payback and heavier debt/FCF compression. The real lever is utilization visibility—without it, 125-145B becomes a speculative asset, not a strategic moat.
Panelists are neutral to bearish on Meta's cloud expansion plans, citing high execution risk, thin margins, and dependence on internal AI workloads to justify massive capex. They agree that Meta's core ad business growth is the primary return driver.
Potential moat from proprietary Llama models becoming the industry standard for enterprise AI inference engines.
High execution risk in pivoting to a B2B cloud provider, with lack of legacy sales force, support infrastructure, and developer ecosystem.