A new report from Morgan Stanley suggests the massive spending on AI by big tech is entering a 'new era' of financial risk. This isn't just another spending cycle; it's an infrastructure build-out so large that its intensity could surpass the dot-com boom.
The numbers are staggering. A handful of tech giants, often called 'hyperscalers'—like Amazon, Google, Microsoft, and Meta—are projected to spend over $2 trillion on AI infrastructure between 2026 and 2028. This spending is so concentrated that the investment decisions of these few companies are now driving the entire market. This creates a fragile situation where a slowdown from just one or two players could have a ripple effect.
So, what’s the core problem? It's what we can call a 'monetization gap.' First, the spending is happening now, and it's enormous. Companies like Amazon and Alphabet have already announced record-breaking capital expenditure (capex) plans for 2026, totaling hundreds of billions. Second, this spending immediately creates fixed costs, like the depreciation of expensive servers and interest on debt, which will weigh on profits for years. Third, the revenue from these AI investments, however, is still more of a promise than a reality. The spending curve is far steeper than the revenue curve, making their earnings extremely sensitive. Even a small disappointment in revenue growth could lead to a big hit to their bottom line.
Adding to this financial pressure is a new, very real physical constraint: electricity. Building and running these massive AI data centers requires an incredible amount of power. A recent report from the Electric Power Research Institute (EPRI) dramatically increased its forecast, estimating that data centers could consume up to 17% of all U.S. electricity by 2030. The power grid is becoming a bottleneck, which means securing energy is now a major cost and a potential source of delays.
Ultimately, 2026 is shaping up to be a 'prove-it' year. The question is no longer whether these companies will invest in AI, but whether they can make those investments pay off before the immense costs overwhelm their financial results. While suppliers like NVIDIA are clear winners for now, the tech giants themselves are taking on all the risk. They must now prove that their AI ambitions can generate real, substantial profits—and soon.
- Hyperscaler: A term for large cloud service providers that operate enormous data centers, such as Amazon Web Services (AWS), Google Cloud, and Microsoft Azure.
- Capex (Capital Expenditure): Funds used by a company to acquire, upgrade, and maintain physical assets like property, buildings, or equipment.
- Monetization: The process of converting something (like an asset or service) into money or revenue.