Meta shares slide 7% on massive spending plans

• Third-quarter revenue of $64.6 billion compared with the expected $63.3 billion
• Fourth-quarter revenue guidance of $73–76 billion versus the $73.7 billion consensus
• Third-quarter capital expenditure reached $34.9 billion, and free cash flow turned negative
• 2026 capital expenditure forecast of $135–145 billion, with the floor raised from $130 billion
• $825 billion in future spending commitments
• New nonvoting Class C shares issued via a 2-for-1 stock split

Despite topping revenue and guidance estimates, Meta's stock declined 7%, a worrying indicator for the AI sector. The shares are also below their level before the Muse app launch.

Intraday, shares fell to $651.61 and had dropped 7.3% by midday. The stock slid about 3% in premarket trading to $677.50, with selling pressure increasing as the day progressed.

The cash flow picture

Advertising operations are performing well. Revenue increased 26% to $64.6 billion, easily surpassing projections. The fourth-quarter guidance of $73–76 billion indicates a midpoint of $74.5 billion, exceeding the $73.7 billion consensus, which represents approximately 24% growth. Excluding legal expenses, earnings were stronger. GAAP earnings per share came in at $3.83 compared to expectations of $3.56, and would have reached $7.40 absent the legal charge.

A year ago, such revenue figures would have suggested a 10% stock increase.

The issue lies beneath the revenue line. Third-quarter capital expenditure reached $34.9 billion, rising from $31 billion in the prior quarter and $19.4 billion a year earlier. This drove free cash flow to negative $2.1 billion, the first such quarter since 2012.

Operating expenses jumped 68% year-over-year to $51.6 billion, driven by infrastructure spending, third-party AI token costs, and a $9.9 billion charge from an August settlement with multiple state attorneys general.

The figure that will linger is the future spending commitments. A regulatory filing revealed roughly $825 billion in planned spending on AI data centers, energy, and computing, an increase from nearly $700 billion as of July. This comprises $356 billion in leases not yet begun and $455 billion in non-cancelable contracts, with $121 billion due in 2027. The two main categories expanded from about $139 billion to $810 billion in one year, representing roughly a sixfold increase.

Class C as a signal

Funding challenges are mounting across the sector. Meta plans to introduce a nonvoting Class C share, effective January 19, 2027, issued as a dividend to all shareholders. This effectively functions as a two-for-one stock split.

Meta stated the move aims to provide the company with the means to fund its expansion and maintain its competitive edge globally. Investors view the new class as a dilutive currency.

A direct parallel is Alphabet's 2014 recapitalization. The Class C shares give Meta a currency for employee compensation, acquisitions, and capital raising. With free cash flow in the red, buybacks halted, and AI investment accelerating, the structure allows management to issue equity without affecting Zuckerberg's voting power.

The bears had already calculated the numbers. Oppenheimer estimates roughly $135 billion in free cash flow losses related to AI and Muse by 2028, exceeding Meta's approximate $90 billion in cash and securities. The firm had highlighted the risk of an equity offering, and Meta has now created the mechanism to do so. The buyback era will be reversed swiftly.

The optimistic thesis remains

Analysts are not abandoning their positive views. Barclays maintained its Overweight rating and increased its price target to $880 from $780.

The primary bull case rests on Muse adoption. Zuckerberg noted that over 10 million users in the US and Canada created Muse personal agents within the first six weeks, with about a quarter using them daily. He repeatedly emphasized this metric during the call. Meta also stated that Reality Labs losses are expected to decrease to roughly $15 billion by 2027, moving toward profitability thereafter.

The capital outlays are contractual and concentrated upfront. Agent-related revenue shows promise but remains nascent.

Wider market impact

The selloff is spreading to the wider AI space. Technology is the worst-performing sector, with declines in Nvidia and AMD, and memory manufacturers Micron and SK Hynix also dropping. Lam Research slid as much as 10% after citing supply chain issues extending to 2027. The Nasdaq fell 2.3%, and the prevailing view is that capital is insufficient to fund all the AI infrastructure hyperscalers desire, forcing either higher costs or reduced plans.

The final point carries more weight than often acknowledged. Meta, financing a $825 billion expansion, is turning to equity markets with the risk-free rate at 5%. All hurdle rates increase. Tolerance for deferred ROI diminishes.

The conclusion is straightforward. Meta's core advertising business is at its strongest point. However, for the first time in over ten years, it is not generating sufficient cash to cover its aspirations, and management has outlined its funding strategy. Without concrete returns from Muse, the market will value Meta as a capital-intensive AI infrastructure firm instead of a cash-generating ad platform, resulting in a lower valuation multiple.

With a market cap of roughly $1.65 trillion and a trailing price-to-earnings ratio of 22, Meta appears inexpensive based on earnings. But the market is applying a cash flow-based valuation.