Meta lifted its 2026 capital‑expenditure outlook to $135‑$145 billion and introduced a non‑voting Class C share to fund AI projects, highlighting an AI capex strain that could pressure cash flow. The company posted third‑quarter revenue of $64.6 billion, up 26% year over year, but operating income fell 37% to $12.99 billion. CFO Susan Li warned that serving costs will climb sharply as compute intensity rises, signaling that the AI‑driven spending surge is beginning to strain cash flow.
Tesla beat earnings expectations, generating $1.53 billion in free cash flow versus a consensus forecast of a $4.99 billion shortfall. The automaker’s capex for the quarter was $4.73 billion, roughly $3 billion below analysts’ $7.77 billion estimate, reflecting a deliberate postponement of infrastructure spending. The company’s 2026 capex target of $25 billion remains viable only if it accelerates spending to about $12 billion in the final three months of the year.
The convergence of rising AI‑related capex and a highly concentrated corporate AI‑token market raises questions about the sustainability of the current investment boom. Ramp data show that the top 1% of companies account for 80% of OpenAI and Anthropic spend, with that elite group averaging $9,568 per employee in September 2026—a modest 3.1% dip from the July peak. Token prices have also slipped: Anthropic’s price fell to $0.83 per million tokens, down from over $1.70 in September, while OpenAI now charges $0.45 per million tokens. Lower pricing and slower spend growth could pressure AI labs to seek new revenue streams or further price cuts.
Data‑center owners that power these models are already feeling the pressure. Goldman Sachs reports that return on invested capital for Amazon, Alphabet, Meta, Microsoft and Oracle fell from 40% to 25% as the cost of scaling compute outpaces revenue growth. Meta’s CFO cited “more than 10‑fold” gains in compute‑cost efficiency, but investors remain uneasy about the scale of future spending. Tesla’s cash‑flow beat, while welcome, rests on deferred capex that may rebound sharply in the fourth quarter, creating a potential mismatch between short‑term earnings and longer‑term cash needs.
Public‑sector agencies and small businesses could feel the impact next. As firms try to recoup higher infrastructure costs, AI‑service prices may rise, narrowing access for organizations that rely on affordable models for public services or niche applications. Investors in Meta, Tesla and AI‑focused funds now face heightened valuation risk, with the disparity between cash‑flow beats and postponed spending underscoring the fragility of current growth narratives.
Meta’s new Class C stock carries no voting rights and is designed to raise capital for AI without diluting Mark Zuckerberg’s control. The move could set a precedent for other tech giants seeking to fund compute‑intensive projects while preserving founder influence, but it adds a layer of complexity for shareholders assessing the true cost of AI expansion.
The next few months will test whether the sector can sustain its spending pace. Analysts will watch Tesla’s fourth‑quarter capex run‑rate and Meta’s quarterly serving‑cost trajectory for signs of strain. At the same time, the concentration of AI spend among a tiny elite of venture‑backed firms leaves open the question of whether broader market competition can emerge to diffuse benefits and temper price pressures. Balancing ambitious AI roadmaps with cash‑flow realities is now the defining risk for the industry’s leading players.