Even if investors crash the stock price, $200 billion must be "burned"—Amazon aims to "regroup"

Amazon is launching the largest capital expenditure plan in its history, attempting to regain momentum in the artificial intelligence sector through massive investments and defend its leading position in cloud computing. Facing strong challenges from Microsoft and Google, this tech giant is trying to reverse the tide by expanding data centers, developing chips, and building models.

According to the Financial Times, Amazon CEO Andy Jassy announced last week that the company’s capital expenditure this year will rise to $200 billion, a figure that surpasses the spending levels of Google and Microsoft. The majority of this huge investment will focus on computing infrastructure, with approximately three-quarters of the budget allocated to its cloud computing division, AWS. At the same time, Jassy has not only integrated the chip, model, and advanced research teams to unify the AI strategy but also cut costs by eliminating about 30,000 corporate positions.

However, investors are uneasy about the scale of this bold gamble. Due to concerns that the return on this massive spending may be too slow, Amazon's stock price has dropped more than 20% from its high last November. Despite facing market punishment, Jassy remains committed to this strategy, stating that the company has deep experience in understanding AWS demand signals and converting capacity into capital returns, and he is confident in this investment.

AWS employees have revealed that the company's aggressive moves reflect internal anxiety that Amazon has not fully leveraged its first-mover advantage in cloud computing. Especially after OpenAI launched ChatGPT in 2022, Amazon was slower than its competitors in vying for major AI vendor contracts.

The Largest Capital Expenditure in History

According to public documents, the vast majority of Amazon’s planned $200 billion capital expenditure will be used for AWS expansion. In comparison, the total spending of Microsoft, Google, and Oracle this year is expected to approach $400 billion. Jassy stated that Amazon plans to significantly increase its data center capacity this year. In 2025, the company will have added nearly 4 gigawatts of capacity, equivalent to the annual energy consumption of more than 3.2 million American households, and plans to double that capacity by 2027.

To support this expansion, Jassy adjusted the organizational structure last December, integrating the group’s chip, model, and advanced research teams under a unified leadership structure. Jassy emphasized that the company will continue to operate like “the world’s largest startup,” maintaining a customer-centric, lean, and fast-moving state.

Competitive Anxiety Facing AWS

Although AWS generated nearly $130 billion in sales last year and contributed more than 60% of Amazon’s total profit, remaining the world’s largest cloud service provider, analysts predict that as demand for AI-driven cloud services surges, Microsoft’s cloud business will surpass AWS within the next three years.

According to more than a dozen current and former senior employees, AWS is losing ground in vying for enterprise AI contracts. A former senior AWS employee admitted, “We were totally unprepared for things to move this fast.”

This lag is evident in key deals. As one of OpenAI's earliest investors, Microsoft secured an exclusive cloud computing contract with the ChatGPT creator early on. Amazon did not sign a $38 billion cloud computing agreement with the startup until OpenAI allowed for a company restructure last year. However, this agreement pales in comparison to OpenAI’s $250 billion contract with Microsoft and its $300 billion deal with Oracle. In addition, although Amazon invested $8 billion in Anthropic and built data centers for it, this investment only came after Google had already backed the startup.

Chip Strategy and Computing Power Breakthrough

To reduce its reliance on Nvidia products and improve profit margins, Amazon is vigorously promoting its self-developed chips. The company claims that sales of its Graviton chips, used for general cloud computing, and Trainium chips, used for AI training, are expected to generate over $10 billion in annual revenue. Last December, Amazon launched the latest generation of Trainium chips, promising significant performance improvements.

Sources have revealed to the media that Amazon is in negotiations to join OpenAI’s latest multi-billion dollar financing round, partly to ensure that the ChatGPT creator adopts its semiconductor products. However, competition remains fierce. Google has successfully attracted Anthropic to purchase its one million TPUs (Tensor Processing Units), with the deal valued at tens of billions of dollars.

Ben Bajarin of tech consultancy Creative Strategies expressed skepticism about whether leading AI startups would adopt Amazon chips. He pointed out that although Amazon emphasizes cost-effectiveness, some users require the utmost performance, even if Amazon’s solution is cheaper to run than Nvidia’s.

Self-Developed Model Mocked as “Amazon Basics”

In terms of model development, Amazon is also investing in its AI model named “Nova,” positioning it as a low-cost alternative to competitor models. However, independent benchmarking shows that Nova’s performance lags behind the state-of-the-art models produced by OpenAI, Google, Meta, and Anthropic.

Sources have told the media that some AWS employees privately refer to Nova as “Amazon Basics,” a term usually used to describe the group’s generic, low-cost household products, which has annoyed executives. Although the company is encouraging employees to use its own AI tools and has set a goal for 80% of developers to use AI for coding at least once a week, several company engineers said they prefer using Anthropic’s Claude model over Nova in their coding work. One AWS engineer bluntly stated, “I don’t even know we’re supposed to have a model.”

The pressure to regain ground in the AI race is weighing heavily on employees. Some are concerned that Amazon may slip into what founder Jeff Bezos described in 2018 as “Day Two”—a state of business stagnation followed by “excruciating, painful decline.” A senior AWS engineer said, “The culture has changed, and the world around us has changed. We have to prove our value.”