AWS hits $100bn annual run rate as AI push accelerates

‘We're at $100bn-plus annualised revenue run rate, yet 85 per cent or more of the global IT spend remains on premises. And this is before we even calculate GenAI,’ says Amazon CEO Andy Jassy

Amazon Web Services has hit the $100bn annual revenue run rate milestone to become the biggest driver of growth for Amazon as a whole.

That milestone is being driven by a shift in how businesses use the cloud, and by a huge push to adopt Amazon's AI and genAI capabilities, Amazon CEO Andy Jassy said Tuesday.

Jassy used his prepared remarks at the company's first fiscal quarter 2024 financial analyst conference call to tell analysts that the company is very bullish on AWS.

"We're at $100bn-plus annualised revenue run rate, yet 85 per cent or more of the global IT spend remains on premises," he said. "And this is before we even calculate genAI, most of which will be created over the next 10 to 20 years from scratch and on the cloud. There's a very large opportunity in front of us."

AWS is seeing several trends that will impact its future, Jassy said.

First, he said, companies have largely completed the bulk of their cost optimisations, and are turning their attention to newer initiatives, including moving from on-premises infrastructure to the cloud to save money and innovate at a more rapid rate to drive more developer productivity.

"Companies are pursuing this relatively low-hanging fruit and modernising their infrastructure," he said. "Our AWS customers are also quite excited about leveraging genAI to change the customer experiences and businesses. We see considerable momentum on the AI front, where we've accumulated a multi-billion-dollar revenue run rate already."

AWS has a three-stack genAI approach, Jassy said.

At the bottom layer, where developers and companies build models themselves, AWS not only has the broadest selection of Nvidia compute instances, but is experienced growing demand for its custom silicon, Trainium and Inferentia, with their favorable price performance benefits relative to available alternatives, Jassy said. Larger quantities of its latest generation Trainium2 are coming in the second half of 2020 or early 2025, he said.

AWS' SageMaker managed end-to-end service has also become a game changer for developers preparing their data for AI, managing experiments, training models faster, lowering inference latency, and improving developer productivity, Jassy said.

"This changes how challenging it is to build your own models, and we see an increasing number of model builders standardising on SageMaker," he said.

In the middle layer the stack, aimed at for developers and companies who prefer not to build models from scratch, but instead prefer to leverage existing LLMs (large language models), customise them with their own data, and deploy secure. high-quality, low-latency, cost-effective production genAI apps, the company built Amazon Bedrock, Jassy said.

"[Bedrock] not only has the broadest selection of LLMs available to customers, but also unusually compelling model evaluation, retrieval augmented generation or RAG to expand models' knowledge base, guardrails to safeguard what questions applications will answer, agents to complete multi-step tasks, and fine tuning to keep teaching and refining models," he said. "Bedrock already has tens of thousands of customers."

Last week, Bedrock launched new capabilities including Custom Model Import, which lets customers use SageMaker to build their own models using Bedrock to make it easier to build high-quality production-grade genAI apps, Jassy said.

"Bedrock Custom Model Import makes it simple to import models from SageMaker or elsewhere into Bedrock before deploying their application. … As more companies find they're employing a mix of custom-built models along with leveraging existing LLMs, the prospect of these two linchpin services SageMaker and Bedrock working well together is quite appealing," he said.

At the top of the stack, where genAI applications are being built, Amazon unveiled the general availability of Amazon Q, which Jassy called the most capable generative AI-powered assistant for software development and leveraging companies' internal data.

"On the software development side, Q doesn't just generate code," he said. "It also tests code, debugs coding conflicts, and transforms code from one form to another. Today, developers can save months using Q to move from older versions of Java to newer more secure and capable ones. In the near future, Q will help developers transform their Dotnet code as well, helping them move from Windows to Linux."

Q also has agents that can autonomously perform a range of tasks from implementing features, documenting, and refactoring code to performing software upgrades, Jassy said.

"Developers can simply ask Amazon Q to implement an application feature, such as asking it to create an ‘Add to Favorites' feature in a social sharing app," he said. "An agent will analyse their existing application code and generate a step-by-step implementation plan, including code changes across multiple files and suggested new functions. Developers can collaborate with the agent to review and iterate on the plan. And then the agent implements it."

Most companies have large troves of internally relevant data in wikis, Salesforce, storage repositories like Amazon S3, and SaaS apps that are hard to access, making answering straightforward questions about company policies, products, business results, code, people, and so on hard and frustrating, Jassy said.

"Q makes this much simpler," he said. "You can point Q at all of your enterprise data repositories, and it will search all this data, summarize logically, analyse trends, and engage in dialogue with customers about this data."

AWS also introduced Q Apps, which lets employees describe in natural language what apps they want to build on top of this internal data and then have the desired app generated automatically, Jassy said.

"This is going to make it so much easier for internal teams to build useful apps from their own data," he said.

Jassy said it is important to not overlook the security and operational performance elements of such genAI services.

"It's less sexy, but critically important," he said. "Most companies care deeply about the privacy of the data in their AI applications and the reliability of their training and production apps. If you've been paying attention to what's been happening the last year or so, you can see there are big differences between providers on these dimensions."

All the changes going on at AWS add to the number of companies moving their AI focus to AWS, Jassy said.

"We expect the combination of AWS' re-accelerating growth and high demand for genAI to meaningfully increase year-over-year capital expenditures in 2024, which given the way the AWS business model works is a positive sign of the future growth," he said

Amazon by the numbers

For its first fiscal quarter 2024, which ended March 31, Amazon reported total revenue of $143.31bn, up 13 per cent compared to first fiscal quarter 2023 revenue of $127.4bn.

That included North America segment sales of $86.34bn, up 12 per cent over last year; International segment sales of $31.94bn, up 11 per cent; and AWS segment sales of $25.04bn, up 17 per cent.

The reported revenue was also divided into net product sales of $60.92bn, up from $57bn, and net services sales of $82.4bn, up from $70.3bn.

For the quarter, Amazon reported GAAP net income of $10.43bn or 98 cents per share, up significantly from last year's $3.17bn or 31 cents per share. The company also reported operating income of $15.31bn, up 221 per cent over last year.

Looking ahead, Amazon expects second fiscal quarter 2024 revenue to be between $144bn and $149bn, or up between seven per cent and 11 per cent compared with second fiscal quarter 2023.

Amazon stock was up a little more than one per cent in after-hours trading Tuesday, trading at $177.16.