[DigitalToday reporter Yoonseo Lee] Amazon CEO Andy Jassy (앤디 재시) defined artificial intelligence as Amazon’s core growth opportunity and unveiled an expansion strategy led by Amazon Web Services and its own AI chips.

On April 11, TechRadar reported that Jassy, in a blog post released alongside an annual letter to shareholders, set out six key views on the future of AI.

Jassy first dismissed claims that AI is overheated. He said every customer experience will be reshaped by AI and that many new experiences will emerge that are possible only because of AI. Debate continues over whether AI is overvalued, in a bubble phase and sufficiently profitable, but Amazon’s view is “no, no, yes,” he said.

He also stressed the pace of AI adoption. “We have never seen a technology adopted faster than AI,” Jassy said, citing the rapid growth of ChatGPT in recent years. He said AI’s impact could be compared with the spread of electricity, adding that if it took 40 years for electricity to take hold, AI appears to be moving 10 times faster.

Amazon’s plan is to place AWS at the centre of that change. Jassy said AWS AI revenue exceeded $15 billion on an annualised basis as of the first quarter of 2026 and that growth is rapid. He said AWS offers broader capabilities than rivals and, through its global network, can process customers’ inference work as close as possible to where they are. He also cited an integrated ecosystem with non-AI services, security and operational performance as key strengths.

He also said customer demand is already outpacing the speed of infrastructure expansion. AWS posted annualised revenue of $142 billion as of the fourth quarter of 2025, up 24 percent from a year earlier, and added 3.9 gigawatts of new power capacity in 2025. Even so, Amazon plans to double its total power capacity by the end of 2027. Jassy said “the absolute scale of growth is very large,” while adding that unmet demand remains. He said customers are seeking more computing resources as AI demand expands.

Amazon also positioned its in-house chip business as a key pillar of its AI strategy. AWS has expanded its chip business since launching Graviton in 2018 and announced fourth-generation hardware in December 2025. Jassy said the AI market has effectively revolved around Nvidia chips so far, but that a different flow has now begun. He said Amazon will continue to work with Nvidia, but customers want better price performance.

He said “demand for Trainium is surging,” underscoring the economics of its in-house AI chips. Jassy said securing high-demand AI chips internally can lower customer costs while improving AWS’s profit structure. If Trainium becomes firmly established, he said it could reduce annual capital expenditure costs by tens of billions of dollars and deliver an advantage of several hundred basis points in operating profit margin compared with relying on outside suppliers for inference chips.

He also reaffirmed a policy of large upfront investment. AWS and Amazon maintain a structure of investing first in land, power, buildings, chips, servers and network equipment, and then monetising them. Jassy said such proactive capital investment allows the company to start new investment more quickly later. He said AWS went through the same cycle during its first major growth phase and the results were also satisfactory.

He also expressed confidence about the future demand base. Jassy cited a recent case of OpenAI committing more than $100 billion and said there are also multiple other customer contracts that have been signed but not yet disclosed, or are in final procedures. If big customer wins and undisclosed contracts continue, the pace of AWS AI infrastructure expansion could accelerate.

Jassy concluded by saying, “AI is unprecedented even with its current growth momentum, and the room for growth ahead is a much bigger, once-in-a-lifetime opportunity.” Amazon is moving to expand its AI lead by linking AWS’s infrastructure competitiveness and customer base with its in-house chip business as new growth engines.