Today on Product Saturday: rivals Databricks and Snowflake roll out new tools that promise to help companies get their agents over the finish line and into production, and the quote of the week.
Today: Microsoft outlines its plan for becoming a frontier AI model company, OpenAI's CFO walks back a suggestion that the U.S. government should pay its infrastructure bills, and the latest moves in enterprise tech.
Today: AI companies are making huge spending commitments that require an enormous leap of faith in their revenue-generating abilities, two cybersecurity professionals are indicted in a series of ransomware attacks, and the latest funding rounds in enterprise tech.
Today on Product Saturday: rivals Databricks and Snowflake roll out new tools that promise to help companies get their agents over the finish line and into production, and the quote of the week.
Welcome to Runtime! Today on Product Saturday: rivals Databricks and Snowflake roll out new tools that promise to help companies get their agents over the finish line and into production, and the quote of the week.
(Please forward this email to a friend or colleague! And if it was forwarded to you, sign up here to get Runtime each week.)
Talk to data: As we close out the year still waiting for agentic AI to make inroads in the enterprise, the tools that might help make a difference are starting to emerge. This week Snowflake launched its Snowflake Intelligence service, first introduced a year ago as a central resource that allows employees to access and analyze corporate data without needing the developers to build them a custom tool.
"Snowflake Intelligence is an enterprise intelligence agent that provides every employee with the ability to answer complex questions in natural language and puts insights at their fingertips," the company said in a press release. Snowflake also introduced a method for testing the output of tools like Snowflake Intelligence called Agent GPA (Goal, Plan, Action) "that catches up to 95% of errors when tested on standard datasets, achieving near-human levels of error detection," it claimed.
Brick by brick: Data companies like Snowflake and Databricks play a very interesting role in the enterprise push toward generative AI apps given that unless they have access to properly managed data, those apps just won't work. This week Databricks added new features to its Agent Bricks service that the company said will improve the accuracy, governance, and interoperability of agents built on the platform.
Databricks released new code for the MLflow project that "traces every agent interaction, scores outputs with custom judges, and provides an app for collecting subject-matter-expert feedback—making MLflow the standard for evaluating agents across any environment," it said in a blog post. Agent Bricks also now comes with a marketplace for third-party servers that support MCP and a document-parsing tool that "not only does it identify or translate text, but it also chunks that document for usage in a vector database," Databricks' Craig Wiley told SiliconAngle.
Find an edge: Before the generative AI boom sucked all the oxygen out of the room, edge computing was a hot topic in enterprise circles, based around the idea that putting compute closer to the end user — as opposed to in a rural data center — can be extremely powerful when building some types of applications. But as with most things in this world, making it all work is harder than it sounds, and this week Cisco introduced a new product that could help.
Cisco Unified Edge is "the industry’s first edge computing platform that unifies compute, networking, storage, and security in one modular system," the company said in a blog post. And as the technology behind all those components gets better, customers can swap in new hardware, Cisco's Jeremy Foster told Fierce Network: "It’s not a one-and-done, it’s an entire system with a long roadmap."
Trust the process: Celonis has been adapting its process-mining technology into what it calls "process intelligence" over the last several years, and this week it introduced new AI-oriented features into its flagship product at its Celosphere event. The Celonis Process Intelligence Platform now works with data lakes from Databricks and Microsoft, and customers can use new tools to build "digital twins" of their operations.
A "digital twin" is a term that usually describes a digital representation of an actual thing, like a building or an airplane, but in this case Celonis believes customers want to "connect desktop actions (keystrokes, mouse clicks, and screen scrolls) to business processes with enhanced Task Mining capabilities and AI-driven Task Discovery," which would allow them to build agents on top of that data, it said in a press release. "Celosphere 2025 isn’t about process mining anymore, it’s about decision mining: finding, understanding, executing where AI should act," according to Constellation Research's Mike Ni.
Ray of blue: Anyscale has raised $259 million to build services around Ray, the open-source project developed by the co-founders of the company to help enterprises run AI workloads across distributed systems. This week it teamed up with Microsoft to ship a new version of that service directly with Microsoft Azure as "a fully-managed, first-party service," which means Azure customers will be able to access it through their core accounts.
"Azure customers can set up and manage Anyscale directly from the Azure Portal, run their Ray-powered AI workloads on Azure Kubernetes Service (AKS), and take advantage of unified billing through Azure," the companies said in a press release. Anyscale's service can be used on all the major clouds, but on AWS it is only available through its marketplace, which generates a separate bill.
GitHub Copilot will always be the seminal AI coding assistant, but the launch of ChatGPT a year later clearly changed the market and probably contributed as much tension to the OpenAI-Microsoft relationship as anything else. According to a new survey from Jetbrains, 41% of software developers are using one of the ChatGPT tools as an AI coding assistant, while 30% are using GitHub Copilot and just 8% are using Anthropic's Claude tools.
"At Microsoft AI, we believe humans matter more than AI." — Mustafa Suleyman, CEO of Microsoft AI, clearing up any confusion as to Microsoft's position on the sanctity of life.
OpenAI suggested that the Trump administration should extend the 35% Chips Act tax credit to companies building AI infrastructure, according to Bloomberg, which isn't the same as asking the federal government to backstop its spending commitments but does show the company is looking to ease that pressure however it can.
Thanks for reading — see you Tuesday!
Today: Microsoft outlines its plan for becoming a frontier AI model company, OpenAI's CFO walks back a suggestion that the U.S. government should pay its infrastructure bills, and the latest moves in enterprise tech.
Today: AI companies are making huge spending commitments that require an enormous leap of faith in their revenue-generating abilities, two cybersecurity professionals are indicted in a series of ransomware attacks, and the latest funding rounds in enterprise tech.
Today on Product Saturday: GitHub's bid to become the center of agentic coding, Cursor and Cognition roll out new coding-specific AI models, 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.
