After more than a year of hype and promises, companies are starting to settle on best practices for deploying and managing AI agents alongside critical business workflows. Nine members of our Roundtable discussed how they successfully rolled out AI agents for tasks other than software development.
Today: Anthropic's latest Sonnet model could make getting capable AI agents up and running much cheaper, Meta makes a big bet on Nvidia's (other) chips, and the latest funding rounds in enterprise tech.
Today on Product Saturday: Cisco unveils a new networking chip for massive AI clusters, GitHub automates software-development busywork, and the quote of the week.
Today: AWS posts its best growth rate in years as it becomes clear the AI boom is lifting all boats, Microsoft suffers a big Azure outage, and the latest enterprise moves.
Welcome to Runtime! Today: AWS posts its best growth rate in years as it becomes clear the AI boom is lifting all boats, Microsoft suffers a big Azure outage, and the latest enterprise moves.
(Please forward this email to a friend or colleague! And if it was forwarded to you,sign up here to get Runtime each week.)
Success and scale
There's no question that the rise of generative AI in late 2022 caught AWS off guard, given that former CEO Adam Selipsky spent basically the entirety of 2023 urging customers and investors to think of the AI boom as a marathon where you can't judge the eventual result by the standings after the first couple of miles. It's hard to know exactly where we find ourselves in that race in late 2025, but after several years of underperformance AWS seems to have figured something out.
Amazon reported Thursday that AWS grew revenue by 20% during the third quarter of 2025 to reach $33 billion, the fastest growth rate it has reported in several years. Rivals Microsoft and Google Cloud are still growing faster, but it's clear that AWS is also benefitting from the once-in-a-generation race to rebuild business applications around a new technology.
AWS is now on a $132 billion annualized revenue run rate, and is sitting on a backlog of $200 billion in customer deals that have yet to be fulfilled and that doesn't include several deals finalized in October, said Amazon CEO Andy Jassy on a conference call with analysts.
Jassy devoted several minutes of the call to its custom Trainium chips, which are a big part of Anthropic's infrastructure strategy and feature prominently in the enormous Project Rainier data center that just became operational this week in Indiana.
The second-generation Trainium chip "is now a multibillion dollar business that grew 150% quarter over quarter," Jassy said.
The AWS results arrived one day after Microsoft and Google Cloud also reported strong earnings results, which indicates that businesses across the board are spending more on infrastructure technology after a lull in investment during the interest-rate reset period in 2022 and 2023. While examples of companies that have put generative AI apps into production are still relatively rare, enterprise tech spending is accelerating across the board after a few years of uncertainty.
Microsoft's cloud revenue — which includes the commercial portion of Microsoft 365, Azure, and its Dynamics 365 suite — rose 26% to $49.1 billion.
While Microsoft once again resisted the temptation to report revenue from its most important business unit, Azure revenue grew 40% during the quarter coming off a year in which it recorded $75 billion in revenue.
And Google is increasingly finding traction among the large companies that used to only get two bids from AWS and Microsoft: "We have signed more deals over $1 billion through Q3 this year than we did in the previous two years combined," Google CEO Sundar Pichai said on a conference call.
Amid widespread talk of an AI bubble, it's hard to say that demand for enterprise AI is waning after the growth numbers trotted out over the last two days. And the Big Three continue to believe that businesses are poised to spend even more.
All three companies raised their forecast for capital expenditures for the current quarter, and unlike other speculative AI investments they are funding that expansion with actual cash, rather than Nvidia's largesse.
AWS has doubled its computing capacity since 2022 and "we're on track to double again by 2027," Jassy said.
For its part, Microsoft isn't slowing down either, despite allowing OpenAI to see other people: "We will increase our total AI capacity by over 80% this year, and roughly double our total datacenter footprint over the next two years, reflecting the demand signals we see," CEO Satya Nadella said on Wednesday's call with analysts.
It's anyone's guess how consumer demand for generative AI technology will play out over the next couple of years heading into what looks like a very shaky couple of quarters for the U.S. economy, but enterprise demand is accelerating.
Shut the front door
One week after AWS customers slogged through a prolonged outage, Microsoft repaid the favor Wednesday with a widespread, hours-long outage linked to problems with its Azure Front Door content-delivery network. The problems started Wednesday morning Pacific Time and ensnared several Azure services that use Front Door to serve content, including the extremely important Microsoft Entra ID identity management service as well as Microsoft 365, arguably the company's most visible service.
"An inadvertent tenant configuration change within Azure Front Door (AFD) triggered a widespread service disruption affecting both Microsoft services and customer applications dependent on AFD for global content delivery," Microsoft said in a preliminary incident review. The recovery process took several hours because it "required reloading configurations across a large number of nodes and rebalancing traffic gradually to avoid overload conditions as nodes returned to service," it said.
AWS also experienced some minor hiccups around the same time, which led some publications who still think DownDetector is a reliable source of information about infrastructure outages to blame AWS for Microsoft's issues. Still, Runtime would like to apologize to customers of both companies for the real root cause of this week's issues; our report on Tuesday that explained why AI providers are having trouble matching the reliability of big cloud providers.
Enterprise moves
Jason Ambrose is the new CEO of People.ai, after serving as senior vice president of marketing and strategy at the enterprise data management company for the last few months.
Tom Krazit has covered the technology industry for over 20 years, focused on enterprise technology during the rise of cloud computing over the last ten years at Gigaom, Structure and Protocol.
Today: Anthropic's latest Sonnet model could make getting capable AI agents up and running much cheaper, Meta makes a big bet on Nvidia's (other) chips, and the latest funding rounds in enterprise tech.
Today on Product Saturday: Cisco unveils a new networking chip for massive AI clusters, GitHub automates software-development busywork, and the quote of the week.
Today: OpenAI releases a new model that could dent Anthropic's hold over enterprise software teams, Anthropic reloads to fight back, and the latest enterprise moves.
Today: Kubernetes installations that use the Ingress NGINX controller have a month to make new arrangements, Salesforce sends Heroku into early retirement, and the latest funding rounds in enterprise tech.