Amazon Raises AWS Ambition to $1 Trillion as AI Demand Accelerates

Key Takeaways

• AWS generated $42.2 billion in Q2 2026 revenue, rising 37% year over year and reaching a $169 billion annualized run rate.
• Amazon CEO Andy Jassy said AWS could eventually become a $1 trillion annual revenue business as cloud migration and AI adoption reinforce each other.
• The PGA Tour shows how moving core data to AWS can create a foundation for automated broadcasts, personalized content, and agentic AI applications.

Not long ago, the PGA Tour moved racks of servers and electronic scoreboards between golf courses by truck. Employees assembled the equipment for each tournament, operated the data systems, then dismantled everything and started again in another city.

“The PGA Tour is a traveling circus,” the vice president of digital architecture at the PGA Tour said. “We don’t play inside a set of white lines.”

That traveling infrastructure offers a compact example of the opportunity Amazon CEO Andy Jassy now sees for Amazon Web Services. During Amazon’s second-quarter earnings call, Jassy said approximately 85% of global IT spending remains tied to on-premises systems. He expects that balance to flip over the next 10 to 20 years.

AWS is already benefiting from the shift. Revenue reached $42.2 billion in Q2 2026, up about 37% from the prior year (source). It was AWS’s fastest growth rate in 18 quarters and its fifth consecutive quarter of acceleration, according to Fortune.

The cloud division experienced significant sequential revenue growth. AWS is now operating at an annualized revenue rate of approximately $169 billion, while its backlog has reached nearly $500 billion. Amazon shares rose more than 15% on Friday following the results.

Those numbers help explain why Jassy told analysts that AWS could “very possibly” become a $1 trillion annual revenue business over time. The forecast is not based only on hosting large AI laboratories. It also assumes millions of existing business applications will migrate to the cloud and begin using AI inference.

AI adoption often increases conventional cloud consumption. Data needs to be stored, governed, moved, and processed before a model can generate useful output. Amazon CFO Brian Olsavsky described that relationship directly, saying customers pursuing AI are accelerating cloud transitions and increasing their consumption of core services.

The PGA Tour followed that sequence. When it opened its global headquarters about five years ago, it did not install fixed IT infrastructure. Instead, the PGA Tour migrated ShotLink, its media archive, website, mobile application, and data-distribution systems to AWS.

AI came later. Using Amazon Bedrock, the PGA Tour can now produce roughly 150 betting profiles in about nine minutes. It also generates thousands of content items each week. At the 2025 Players Championship, AI-generated commentary supplied a two-sentence summary for every shot taken during the tournament.

Building on this foundation, the PGA Tour introduced “agentic production,” a fully automated television broadcast in which ShotLink data directs camera selection and graphics. Its mobile application also presents AI-generated performance insights when players such as Scottie Scheffler complete a round. Could those products have been built while servers were still traveling from course to course? Perhaps, but the operational burden would have been considerably higher.

The broader market supports Amazon’s aggressive investment. Reuters has tracked the expanding infrastructure requirements surrounding AI and cloud services, while technology investor and publisher Tom Tunguz has examined AWS’s position in the cloud race. Amazon now expects 2026 capital expenditures to reach $220 billion, up from previous estimates.

That spending still carries execution risk. Data centers, chips, networking equipment, memory, and power capacity require substantial upfront commitments, and AWS competes with Microsoft Azure and Google Cloud for enterprise workloads. Customers are also weighing model accuracy, latency, governance, and cost rather than adopting AI solely for novelty.

Amazon Bedrock is central to AWS’s response. Customers spent more on Bedrock during Q2 2026 than in all previous quarters combined since its 2023 launch. More customers joined the service in the past six months than during its first two years, while AWS's AI business continues to scale rapidly (specific AI revenue run rate metrics were not disclosed).

Jassy describes current AI demand as a barbell: large AI laboratories and breakout applications occupy one end, while enterprises using automation to reduce costs occupy the other. The largely undeveloped middle consists of production workloads that could incorporate inference but have not yet done so.

For business technology leaders, the PGA Tour offers a practical lesson. The IT leadership stated the organization no longer pursues isolated “AI projects.” It pursues business projects that may use AI, while accepting that some should not. That distinction matters. AWS’s path toward Jassy’s $1 trillion ambition depends less on experimentation for its own sake and more on customers finding repeatable uses that generate enough value to keep consuming cloud infrastructure.