AWS is now a $100 billion revenue run rate company, executives confirmed late Tuesday as Amazon reported its Q1 earnings.
Here are five key takeaways from its earnings call.
1: AWS Growth
AWS growth of 17% is not the same as the 48% growth of mid 2018 or the 37% growth of early 2022. But it’s growth from a much larger base…
AWS is now a $100 billion run rate company, Amazon CEO Andy Jassy confirmed, with the cloud provider generating $25 billion in Q1 revenues.
Jassy said: “Companies have largely completed the lion's share of their cost optimization and turned their attention to newer initiatives.”
Forrester’s principal analyst Lee Sustar added: “Noteworthy was a near doubling of AWS operating revenue in the quarter to $9.4 billion…
“...AWS has managed to invest to stay in the AI game without dragging down the operating income that helps drive Amazon overall.”
2: AI generating billions already
AWS’s managed AI service Bedrock already has “tens of thousands of customers,” Jassy claimed, including Adidas, New York Stock Exchange, Pfizer, and Toyota. That means “on the AI front… we've accumulated a multibillion-dollar revenue run rate already” he told analysts.
That’s a striking success in just 18 months since ChatGPT launched and lit a fire under generative AI. It also requires sustained capital investment…
3: CapEx is set to soar
Investment in the GPUs and networking gear as well as broader data center expansion needed to make AI run fast is going to be expensive.
Amazon’s capital expenditure (CapEx) in Q1 was $14 billion. It expects that to be the lowest quarterly figure in 2024, putting CapEx in the $60 billion+ region (executives did not break out an annual CapEx budget.)
As CFO Brian Olsavsky put it: “We are seeing strong demand signals from our customers and longer deals and larger commitments, many with generative AI components” – 2023’s CapEx was $48.4 billion he added.
4: Q and Custom Model Import
When it comes to AI, AWS’s Copilot-equivalent, “Q” is now generally available, the company added, saying it “tests code, debugs coding conflicts, and transforms code from one form to another…”
With less fanfare AWS also recently announced a “sneaky big launch” of Custom Model Import. That means that organisations who fine-tuned or “domain adapted” open models on-premises can, if they decide to do so, import these customised models to run on AWS infrastructure using Bedrock’s invoke model API, creating what the company described as a “a unified experience across base, custom, and imported models.”
This is, for now, however, only a preview available in US East (Virginia) and can’t be used with Amazon Guardails or Cloudformation. See docs.
5: Dividends or share buybacks? No chance
With all this growth is Amazon splashing out on dividends or buying back stock like Alphabet and Meta have done this year? No chance.
Amazon’s philosophy of reinvesting in the business stands, Olsavsky said: “Our first priority is to support the growth opportunities and long-term investments within our businesses… we still have many opportunities for that capital to use that would generate meaningful returns,” he added.
All in all, Jassy concluded, “we're seeing strong AWS demand in both generative AI and our non-generative AI workloads, with customers signing up for longer deals, making bigger commitments.
He added: “When you think about how many generative AI applications will be out there over time, most will end up being in production when you see the significant run rates. You spend much more in inference than you do in training because you train only periodically, but you're spinning out predictions and inferences all the time.. we see both training and inference being really big drivers on top of AWS,” Amazon’s CEO said.
“We remain very bullish on AWS.”
