AWS chief says enterprise AI is now delivering ROI

Amazon Web Services Chief Executive Officer Matt Garman says enterprise AI has moved beyond experimentation, with more organisations reporting measurable returns on investment as deployments shift into production.

Speaking in an interview on the Platformer podcast, Garman said customer demand for AI has strengthened significantly over the past year, influencing both enterprise technology strategies and Amazon's investment plans.

Garman said discussions with chief information officers indicate a marked change in how organisations view AI adoption.

"I was talking to a room full of CIOs just a couple of months ago," Garman said. "I asked, 'How many of you are either seeing materially positive ROI today or have a path in the next couple of months to really high ROI?' 90% of hands went up, which is totally different than a year before."

AI returns

According to Garman, enterprises are progressing from pilot projects to operational AI deployments that generate business value.

He said AI adoption is advancing more quickly than cloud computing did because organisations already have cloud infrastructure in place. That foundation allows businesses to deploy AI services without first making the large-scale infrastructure changes that cloud migration required over the past two decades.

Garman said this growth underpins Amazon's commitment to invest USD $200 billion in capital expenditure this year.

He described the investment approach as being based on customer demand and long-term infrastructure assets rather than speculative forecasts.

"If you really like the ROIC [Return on Invested Capital] of a business, you want the 'C' to be as high as possible. It's not speculative," Garman said.

He said investments in land and power retain value even if market demand changes, while server and semiconductor purchasing decisions are made only months in advance when customer demand is more visible.

"We have a lot of mitigations in there and we really think intentionally about how we can reduce risk," Garman said.

Managing costs

Garman also discussed how AWS is helping organisations improve returns from AI deployments by matching workloads with appropriate AI models.

He said many organisations increase costs by using the most advanced models for every task regardless of complexity.

AWS addresses this through Kiro, its agentic software development environment, which automatically selects different models depending on the work being performed. Simpler models can handle tasks such as code generation, while more advanced reasoning models are reserved for more complex requests.

"One of the things that's driven up a bunch of cost of AI is that people were trying to use the best model for every single thing," Garman said. "In Kiro, we do a lot of this for customers where we pick the right model and then we appropriately help customers budget to get the results they want faster, and less expensively."

Enterprise focus

Beyond model selection, Garman said organisations should evaluate AI based on business outcomes rather than measuring token usage or computing consumption.

He encouraged businesses to give employees greater autonomy over AI use while focusing on the value delivered by AI applications.

Garman also recommended expanding projects that demonstrate measurable returns while discontinuing initiatives that fail to produce results.

The interview also covered the future of AI agents, employment, infrastructure investment and enterprise adoption as organisations continue integrating AI into business operations.