Major U.S. technology companies are pressing ahead with large investments in AI infrastructure—especially data centers. The Verge’s compilation shows that Amazon, Microsoft, Alphabet and Meta plan combined capital expenditures of at least $720 billion this year, and have committed or prepared at least another $1 trillion in spending for future years.
Revenue growth but even faster rising costs
Quarterly results indicate the core businesses are still growing: Amazon reported a level of growth not seen in five years, Meta’s ad revenue is rising strongly, and Google Cloud’s revenue more than doubled year‑on‑year. However, The Verge argues that this revenue growth may not be sufficient to justify the capital intensity of AI investments, since costs—particularly compute capacity and data‑center construction—can rise faster than revenues.
Alphabet: monetization challenges for AI queries
At Alphabet, Philipp Schindler, the company’s Chief Business Officer, acknowledged that the growth in AI model queries has affected revenues because those services have shown lower monetization. Google’s search ad revenue grew 12 percent in the third quarter versus 17 percent in the same quarter a year earlier, placing pressure on the search business that had previously provided a stable revenue base while AI‑based search’s business model remains unproven.
Executives signal more spending, not restraint
Company leaders are not signalling a slowdown. Alphabet expects $195–205 billion in capital expenditures this year and its CFO indicated significant further increases could come in 2027. Amazon plans $220 billion in capex this year and CEO Andy Jassy said they would spend meaningfully more in 2027. Meta forecasts $135–145 billion of capex for 2026, and major financial institutions expect additional increases next year.
Mark Zuckerberg has warned that Muse, Meta’s agentic AI program, can only become profitable if the company drives down compute costs—he said Meta expects a tenfold efficiency improvement next year, which would require substantial investment.
Financing shifts: Microsoft bond sale and outside capital
Microsoft’s quarterly capex was $45.2 billion, below the $50 billion analysts expected, but The Verge notes this shortfall stemmed from third‑party capacity delays rather than a deliberate pullback. Microsoft also issued a $25 billion bond offering in ten tranches, signaling that even the strongest‑balance tech companies are increasingly tapping external financing to fund the infrastructure race.
Cost pressure extends beyond the biggest platforms
Cost pressures affect startups as well. Anthropic’s updated SEC filings show quarterly revenue of $17 billion while compute costs were $17.5 billion. The company disclosed nearly $900 billion of future compute and infrastructure commitments and plans to build its own data centers. The Verge notes that some of these commitments could translate into TPU purchases from Google, which would not produce immediate cash revenue for Google while Google must finance the hardware buildout in advance.
Risks: investor patience, politics and physical constraints
Wall Street’s patience—which had been relatively long over the past three years—is fraying as investors press for clearer evidence of when and how these investments will pay off. Tech stocks came under pressure last week as markets began to price in the costs and uncertain returns of AI spending; Alphabet’s shares plunged about 13 percent in a single day, one of the largest one‑day losses in the company’s history.
If companies do not slow voluntarily, external brakes could appear: regulators, local community opposition to data centers, grid capacity limits or ultimately capital markets could constrain the pace. The Verge also highlights the political angle: an uptick in data‑center‑antagonistic policies after U.S. midterm elections could act as an external brake, even though election risk has not been a prominent topic on corporate earnings calls.
Conclusion
Big Tech views AI as a historic opportunity and is committing immense capital to realize it. At the same time, the timing and scale of returns remain unclear. That ambiguity, together with mounting investor scrutiny, potential political resistance and infrastructural limits, could increasingly determine how long and how intensely these companies can finance the race for AI infrastructure.
