Generated by Codex with GPT-5
What this article is about
This summary covers The Economist’s March 28th, 2026 Business article listed in the contents as Amazon and published under the Schumpeter column with the headline The Bet-Everything Store.
The article argues that Amazon is making the biggest spending bet in its history in order to regain momentum in artificial intelligence. The core question is whether this is an undisciplined burst of spending or a calculated attempt to protect AWS, its most valuable business. The piece comes down on the side of cautious optimism: the bill is enormous, but Amazon has real strategic reasons for taking it on.
Why Amazon is spending so heavily
Amazon plans to spend about 200bn on capital expenditure this year, mostly to build the data centres, chips and power capacity needed for AI. It has also signalled that it may put as much as 50bn into OpenAI, far more than Microsoft originally committed when it backed the company.
The article says this is partly about genuine demand. AWS is still selling cloud capacity as quickly as it can build it. But it is also about competition. Since ChatGPT arrived in 2022, Microsoft and Google have grown faster in cloud and AI, eating into the aura of AWS as the unquestioned leader. Amazon is now trying to regain the initiative rather than accept a slower drift into second place.
Why the article thinks the gamble could pay off
The piece presents Amazon’s spending spree as a recognizable company pattern rather than a sudden break with its past. Amazon has a history of making giant bets, some of which looked reckless at the time and later proved highly profitable. Its expansion in logistics and AWS in 2016 and 2017 improved margins for years afterward, even though its pandemic-era warehouse binge was a clear overbuild.
This time the article thinks Amazon has several advantages:
- AWS still has the broadest customer base in cloud computing
- Amazon now has relationships with both OpenAI and Anthropic
- AWS can offer customers a wide menu of models and chips rather than forcing them into one stack
- Amazon has more reason than Microsoft or Google to prioritise cloud customers, because AWS is far more profitable than its retail business
That last point matters. Microsoft and Google must divide scarce AI resources between their cloud customers and their own lucrative software and search products. Amazon, by contrast, has a stronger incentive to keep AWS at the centre of the queue.
Why investors are uneasy
The article does not treat the spending as obviously safe. Investors worry that the returns may arrive too slowly, while depreciation and infrastructure costs show up immediately. Even if AI demand stays strong, the scale of the build-out could compress margins.
There is also a second fear: AI agents could weaken Amazon’s main shopping business by standing between consumers and Amazon’s marketplace. If customers begin delegating purchases to bots, Amazon risks losing some of its direct relationship with shoppers and advertisers.
Still, the piece says Amazon is not as exposed as sceptics assume. Its deal with OpenAI appears to have helped cool plans for a competing shopping service, and Rufus, Amazon’s own shopping assistant, reportedly generated 12bn in incremental annualised sales last year. In other words, Amazon is trying to turn the agent threat into another reason to spend more aggressively.
The takeaway
The article’s main message is that Amazon’s AI binge is not just corporate excess. It is a deliberate effort to stop AWS from losing strategic ground in the most important infrastructure race in tech.
The risk is obvious: if enterprise AI adoption stays slower than expected, Amazon could be left with a very expensive overbuild. But if AI demand broadens over the next year or two, the company may look less like a careless spender and more like the tech giant that moved early enough to keep its lead.