admin
America’s four largest AI spenders are pushing capital outlays towards a record level this year, betting that demand for computing power will justify one of the most expensive investment cycles in corporate history.
Microsoft, Amazon, Alphabet and Meta are preparing to spend roughly $680 billion to more than $700 billion on data centres, chips, servers, networking equipment and related infrastructure in 2026, after quarterly earnings showed that artificial intelligence is now central to growth plans across cloud computing, advertising, software and consumer platforms. The scale of the commitment has reassured some investors that demand remains strong, while raising concern that returns may take longer to materialise than markets expect.
Microsoft has emerged as one of the clearest examples of the trade-off. Its cloud business delivered another strong quarter, with Microsoft Cloud revenue exceeding $54 billion and Azure and other cloud services growing about 40 per cent. The company’s AI business has crossed an annualised revenue run rate of more than $37 billion, a figure that underlines how quickly enterprise demand has shifted from experimentation to deployment. Yet management also signalled that capacity constraints could continue through 2026, implying that heavy spending may remain necessary even after a sharp increase in data-centre investment.
Amazon’s case rests on the breadth of its cloud and chip strategy. Amazon Web Services generated $37.6 billion in quarterly revenue, up 28 per cent, its fastest pace of growth in several quarters. The company is relying heavily on Trainium, its custom AI chip, to reduce dependence on third-party processors and win large workloads from AI companies. Commitments tied to Trainium have reached hundreds of billions of dollars, helping support Amazon’s plan for about $200 billion in capital expenditure this year. Even so, free cash flow pressure remains a key concern as the company balances cloud expansion, logistics investment and retail margin discipline.
Alphabet has drawn stronger market approval because its spending is being matched by faster cloud growth. Google Cloud revenue rose 63 per cent to about $20 billion, helped by enterprise adoption of AI tools, custom chips and infrastructure services. Alphabet raised its capital expenditure guidance to about $180 billion to $190 billion, citing higher demand and capacity needs. Its integrated approach, combining cloud infrastructure, proprietary tensor processing units, search, advertising and Gemini models, has strengthened the perception that it may be converting AI investment into revenue more efficiently than some rivals.
Meta’s position is more divisive. The company lifted its 2026 capital spending forecast to $125 billion to $145 billion, reflecting higher component costs and accelerated data-centre expansion. Revenue growth remains strong, with first-quarter sales rising 33 per cent to $56.3 billion, supported by better advertising performance and AI-driven targeting. But investors have questioned whether the scale of spending will deliver returns quickly enough, particularly as Meta continues to carry losses from its virtual and augmented reality division. A $25 billion bond sale also showed how the funding model for AI infrastructure is widening beyond internal cash generation.
The investment wave is being driven by a shortage of advanced computing capacity. Large language models, enterprise AI agents, image and video generation tools, coding assistants and inference workloads all require dense clusters of graphics processors or custom accelerators. Demand is also shifting from model training to everyday use, meaning companies need infrastructure that can handle millions of queries and business processes at low latency.
That shift has widened the race beyond chips. Power availability, land, cooling systems, fibre links, grid connections and construction timelines have become strategic constraints. Data-centre operators are signing long-term electricity agreements, exploring nuclear and renewable power arrangements, and seeking locations where permitting and energy supply can support rapid build-outs. Memory prices and supply-chain bottlenecks have added further pressure, particularly for companies trying to secure high-bandwidth memory used in AI accelerators.
Investors are no longer judging AI spending by ambition alone. Earnings reactions showed a sharper distinction between companies that can link infrastructure investment to visible revenue gains and those asking markets to wait for future monetisation. Alphabet and Amazon benefited from stronger cloud signals, while Meta faced pressure after lifting spending guidance. Microsoft’s cloud performance remained strong, but the prospect of sustained capacity constraints and higher outlays kept attention fixed on future margins.
