In 2027, will AI capital expenditure reach 1 trillion dollars?

In 2027, will AI capital expenditure reach 1 trillion dollars?

```

The AI arms race is accelerating, and the bill is much bigger than the market anticipated.

According to Chasing Wind Trading Desk, on April 29, BofA Securities released a semiconductor industry research report. Analyst Vivek Arya’s team, after the four major U.S. tech giants—Google (Alphabet), Microsoft, Amazon (AWS), and Meta—announced their Q1 2026 financial results and updated capital expenditure outlooks, raised their forecast for global hyperscale cloud enterprise capital expenditures in 2026 to over $800 billion (up 67% year-on-year), and predicted that it will further surpass $1 trillion in 2027 (up about 25% year-on-year).

Four Giants Ramp Up Together

Google, Microsoft, Amazon, Meta—these four tech giants all raised their capital expenditure guidance for 2026, with Microsoft seeing the largest increase, about $24 billion higher than previous market expectations.

  • Google: 2026 capital expenditure about $185 billion, higher than the previous expectation of $178.3 billion; also previewed a "substantial increase" for 2027.
  • Microsoft: 2026 capital expenditure $190 billion, up 61% year-on-year, far exceeding previous market expectation of about $154 billion.
  • Amazon: Maintained 2026 capital expenditure guidance at $200 billion, up more than 50% year-on-year.
  • Meta: Raised the 2026 capital expenditure median to $135 billion, higher than the previous $125 billion.

Including Oracle ($59.6 billion), the report expects that total capital expenditure of the top five U.S. tech companies will reach about $769.6 billion in 2026. If Chinese leading cloud providers (Alibaba, Tencent, Baidu, totaling about $35 billion) are included, global hyperscale cloud providers’ total capital expenditure in 2026 will surpass $800 billion, up 67% year-on-year.

AI Sales Explode, Investment Returns Materializing

The core logic driving this round of capital expenditures is rapid realization of AI commercialization.

Google: Gemini's large model processes more than 1.6 billion tokens per minute via API, up 60% quarter-on-quarter; paid monthly active users for Gemini Enterprise Edition grew 40% quarter-on-quarter, becoming the main driver of Google Cloud’s quarterly growth; Google Search revenue grew 19%, the fastest pace in recent years, underpinned by rapidly growing AI search queries.

Microsoft: Annualized AI business revenues now exceed $37 billion, up 123% year-on-year; management expects that, despite ongoing supply tightness, Azure cloud services will achieve modest acceleration in the second half of 2026.

Amazon: AWS first quarter revenue grew 28% year-on-year to $37.6 billion, the fastest in more than three years, driven by AWS's workloads and Bedrock partnership with OpenAI, Anthropic and others.

Supply Constrained, Pricing Power Tilts Toward Sellers

BofA notably highlighted a key detail: computing power will remain supply-constrained throughout 2026, and hyperscale cloud provider expansion is supported by customer commitments, not just preemptive stockpiling.

More noteworthy is the cost transmission logic. Microsoft explicitly stated that about $25 billion of its $190 billion capital expenditure budget is due to price increases of memory, wafers, substrates and other components.

BofA believes this statement is a positive signal for AI semiconductor suppliers: major computing power and network device manufacturers are able to pass cost increases onto customers, pricing power and profit margins are likely to remain stable.

The report identifies AI semiconductor beneficiaries including: computing power (NVIDIA), storage, semiconductor equipment, power semiconductors and optical modules.

BofA also notes that most hyperscale cloud providers are pushing "heterogeneous deployment" equally—using both commercial GPUs and self-developed/custom chips in parallel, rather than betting exclusively on one approach. This trend means market space is not a zero-sum game and the entire AI infrastructure supply chain is poised to benefit.

At the current pace, BofA predicts global hyperscale cloud provider capital expenditure will surpass $1 trillion in 2027, up about 25% year-on-year. Google has clearly previewed a "substantial increase" in capex for 2027, and the other giants’ expansion intentions show no sign of abating.

 

~~~~~~~~~~~~~~~~~~~~~~~~

The above exciting content is from Chasing Wind Trading Desk.

For more in-depth interpretation, including real-time parsing, frontline research, etc., please join Chasing Wind Trading Desk · Annual Membership

Risk Disclosure & DisclaimerThe market involves risk, investment should be cautious. This article does not constitute personal investment advice, nor does it take into account an individual user's specific investment objectives, financial situation or needs. Users should consider whether any opinions, viewpoints or conclusions in this article fit their own situation. Any investment based on this is at the user's own risk. ```