StoneX reiterated its "Hold" rating on Meta Platforms on October 19, expecting the company to raise its full-year expense and capital expenditure guidance in its third-quarter earnings report scheduled for release on October 21.
The firm believes that, due to factors including the U.S. state attorneys general settlement agreement and increased research, development, and marketing spending on the AI product Muse, Meta's existing 2026 total expense guidance of $165 billion to $167 billion could be raised by at least $10 billion.
At the same time, with the release of the Wonder model and intensifying competition in frontier AI, the company's full-year capital expenditure forecast may also be further increased, placing greater pressure on 2027 profit margins and free cash flow.
Full-Year Expense Guidance May Be Raised by at Least $10 Billion, Muse Further Increases Operating Costs
StoneX expects Meta to raise its 2026 total expense guidance from the current $165 billion to $167 billion by at least $10 billion, mainly reflecting additional expenses from the U.S. state attorneys general settlement agreement reached during the third quarter.
Meanwhile, Meta's increased marketing and R&D spending in the second half of this year around the AI agent product Muse could also push expense expectations even higher.
The firm specifically notes that Muse has already moved from the R&D stage into actual product delivery, meaning the related spending is no longer limited to research and development expenses.
As the user base expands, ongoing costs such as AI inference, per-user virtual machine resources, and back-end processing will be recorded in operating costs, potentially putting pressure on future gross margins.
StoneX expects Meta to provide a general direction for 2027 expense growth in its third-quarter earnings report, but formal full-year expense guidance may not be announced until the fourth-quarter earnings release.
Wonder Model Release Drives AI Competition Higher, Capital Expenditure May Exceed Previous Expectations
On the capital expenditure side, Meta currently expects 2026 spending of $130 billion to $145 billion.
StoneX believes that with the release of the Wonder model and the company's increased investment in frontier AI research and development, management may raise its full-year capital expenditure forecast and further narrow the guidance range toward the higher end.
Meta is continuing to expand its AI training and inference infrastructure while strengthening its procurement of external computing resources.
StoneX believes these investments could drive up equipment depreciation, infrastructure operating costs, and computing resource procurement costs over the coming years, while increasing cash flow pressure in the near term.
The firm expects that Meta may only provide a general direction for 2027 capital expenditure when it releases third-quarter results, with formal quantitative guidance likely to be disclosed in the fourth-quarter earnings report.
Currently, StoneX's forecasts for Meta's 2027 total expenses and capital expenditure are 3.5% and 3.0% above market consensus, respectively.
CoreWeave and Nebius Computing Contracts May Further Compress 2027 Gross Margins
StoneX is particularly focused on Meta's new contracts with AI cloud computing companies such as CoreWeave and Nebius, believing these agreements could increase the company's future cash operating costs and put additional pressure on 2027 gross margins.
Although Meta's gross margin over the past 12 months still reached 81.75%, as AI products move into actual operation, inference computing, user-dedicated resources, and back-end processing fees will continue to be incurred as service usage grows.
In addition, depreciation and amortization expenses arising from large-scale capital expenditure may also gradually be reflected in future income statements.
StoneX expects that the gross margin pressure Meta faces in 2027 may be greater than the market currently anticipates.
The firm will also focus on whether the company accelerates payments to new cloud computing service providers and whether these expenditures further compress short-term free cash flow.
Guggenheim and Morgan Stanley Raise Target Prices, Bullish on AI Product Commercialization Prospects
Unlike StoneX's caution on expense growth, some Wall Street institutions have recently become more optimistic about the growth potential of Meta's artificial intelligence business.
Guggenheim raised its target price on Meta from $700 to $850, maintaining a "Buy" rating, mainly based on the rapid growth in daily active users of the AI agent product Muse, which demonstrates strong market acceptance.
Morgan Stanley raised its target price on Meta from $775 to $820, maintaining an "Overweight" rating.
The firm is bullish on Meta's future AI business growth and has correspondingly raised its 2027 and 2028 revenue forecasts, while slightly increasing its U.S. GAAP earnings per share expectations.
The market's main disagreement over Meta centers on the relationship between AI investment and commercial returns.
Some institutions believe that user growth in AI products such as Muse will drive future revenue expansion, while StoneX is more focused on the impact of new computing demand, R&D investment, and ongoing operating costs on profitability.
With Meta set to release its third-quarter earnings report on October 21, the market will focus on whether full-year expense and capital expenditure guidance is raised, as well as management's latest assessment of 2027 AI infrastructure investment, gross margins, and free cash flow.
Comments