3 Sources
[1]
AWS installs an AI accelerator every 14 seconds and eyes $1 trillion annual revenue
Before long, we could be writing about a cloud company with a retail side hustle Amazon Web Services will become a trillion-dollar business in five to ten years, and thanks to the AI boom Amazon CEO Andy Jassy thinks it should be able to grow by 40 percent or more in 2027 or 2028. The cloud giant
[2]
Amazon Q3 earnings: customers 'gobbling' up AI capacity, Jassy says
Andy Jassy says Amazon could have two businesses with $1tn in annual revenue within a decade. The Amazon Q3 earnings also showed its highest operating margin yet, 13.9%. Amazon's free cash flow fell to minus $42.4bn over the 12 months to September. Spending on AI drove most of the drop, the
[3]
Amazon Q3 Earnings: CEO Jassy Says Bedrock Could Become As Big As EC2
Twenty-plus years in technology has taught Amazon CEO Andy Jassy 'that there is not one tool to rule the world,' he says. Amazon Web Services' Bedrock artificial intelligence inference service could become as large as its core Elastic Compute Cloud web service as enterprises adopt multiple models
Share
Copy Link
Amazon Web Services is installing AI accelerators every 14 seconds to meet surging demand, CEO Andy Jassy revealed in Q3 earnings. AWS hit $184 billion annual revenue with 39.3% growth, while Amazon spent $75.9 billion on infrastructure last quarter. Jassy projects AWS could reach $1 trillion revenue in 5-10 years as Bedrock becomes the fastest-growing service in company history.
Amazon Web Services achieved 39.3% year-over-year growth in Q3, pushing quarterly revenue to just under $46 billion and establishing an annual revenue run rate of $184 billion
1
2
. This marks AWS revenue growth at its fastest rate in almost five years, with operating income climbing to $17.8 billion from $11.4 billion2
. CEO Andy Jassy projects AWS could grow by more than 40% in both 2027 and 2028, potentially reaching $1 trillion in annual revenue within 5-10 years1
2
. Overall Amazon earnings showed revenue of $180.17 billion for the quarter, with operating margins reaching a record 13.9%2
.
Source: CRN
AWS is installing a new AI accelerator approximately every 14 seconds, with customers consuming capacity almost as quickly as it becomes available, Jassy revealed on the earnings call
1
2
. This AI capacity expansion required massive capital expenditure, with Amazon spending $75.9 billion in Q3 alone, up from $35.1 billion a year earlier2
. The company expects approximately $220 billion in capital expenditure for 2026 and plans to spend meaningfully more in 20271
2
. Amazon raised $48.1 billion in long-term debt during the quarter and can now borrow up to $80 billion against AI hardware purchases2
. This aggressive AI infrastructure expansion caused Amazon's free cash flow to fall to minus $42.4 billion over 12 months, compared to a positive $14.8 billion a year earlier2
.Bedrock AI services has become the fastest-growing multibillion-dollar service in AWS history, with token volume growing nearly 17-fold year over year
3
. Spending by Fortune 500 customers on Bedrock increased nearly sevenfold during the same period3
. Jassy predicts Bedrock could eventually match the scale of EC2, AWS's flagship compute service, as enterprises adopt multiple AI models to avoid vendor lock-in3
. The CEO expects six to eight AI models to handle at least 95% of workloads over the next two to three years, with different models excelling in reasoning, coding, visual tasks, latency, and cost3
. Tens of thousands of customers now use Amazon Bedrock AgentCore, with users deploying twice as many agents during Q3 as in the prior three quarters combined3
.Related Stories
AI workloads are driving increased consumption of core AWS services, creating what Jassy described as a flywheel effect for AI-driven cloud consumption
1
3
. When customers select AWS for AI workloads, they position inferencing capabilities near existing data and applications, which then drives demand for core compute and storage services used in reinforcement learning, post-training, and tool usage1
. Core compute and storage services grew at their fastest rates in several years1
. Operating margins for AI services are already well ahead of what AWS achieved for core services at similar stages of development1
. Lambda, AWS's serverless offering, surpassed a $1 billion annualized revenue run rate3
.
Source: The Register
Amazon plans to start selling its Trainium AI chips outside AWS late in 2027, expanding beyond its current internal use
2
. The company recently released TorchNeuron for Trainium, enabling developers to run PyTorch code on the chips2
. Jassy emphasized structural cost advantages from owning chip technology and long-leasing data centers rather than using neoclouds, which he described as about 3,500 basis points worse in operating margin3
. He noted that approximately 85% of global IT spend remains on-premises, with a 10- to 20-year cycle expected to flip that equation to the cloud1
3
. Even traditionally immovable workloads like mainframes and storage area networks are now migrating to cloud1
. AWS maintains a backlog of $578 billion and more than 140,000 partners worldwide3
.Summarized by
Navi
29 Apr 2026•Business and Economy

29 Jul 2026•Business and Economy

01 Nov 2024•Technology
