You can observe a lot by watching

Welcome to Runtime! Today: Why Snowflake's acquisition of Observe underscores the importance of observability to enterprise AI, Microsoft attempts to fend off the AI coding upstarts, and the latest enterprise moves.

Please forward this email to a friend or colleague! If it was forwarded to you, sign up here to get Runtime each week, and if you value independent enterprise tech journalism, click the button below and become a Runtime supporter today.

Open and shut

Operating a business on the internet at scale has never been more complex than it is today, even though that statement has been true at basically any given point over the last 20 years. Simply keeping track of everything running across a modern enterprise is a herculean task, and if enterprise agents take off in 2026, the amount of activity that IT organizations will have to catalog will explode.

Observability tools rose to meet this challenge during the transition from on-premises data centers to cloud computing, and they're evolving once again as enterprise AI moves from concept to production. Snowflake's decision to acquire Observe, announced Thursday, is a bet that moving those tools closer to the data they're recording will be a cheaper and easier way to make sure enterprise agents are performing as designed.

• Terms of the deal were undisclosed, but Observe had raised $463 million according to Crunchbase and The Information reported that Snowflake likely paid around $1 billion for the company.
• Observe's software was built around Snowflake's data platform, and it allows customers to keep an eye on the performance of their apps and infrastructure in hopes of mitigating small problems before they turn into big ones.
• That software also comes with an "AI SRE" that uses its own AI agents to plan and execute a response to new errors just like a human site-reliability engineer would during an incident.

Snowflake became a $76 billion enterprise software company thanks to its business-friendly data analysis tools, which arrived more than a decade ago as many companies were moving their operations into the cloud and generating reams of data to parse. But a few years ago it was forced to acknowledge that customers really wanted to store and manage their data in open formats like Iceberg, which is similar to the transition that the observability market is currently going through with the maturation of OpenTelemetry.

• Companies loved the insights obtained from early observability tools from companies like Datadog, but balked at the cost of storing and accessing that data in proprietary formats.
• OpenTelemetry, which is managed by the Cloud Native Computing Foundation, was designed to provide a standardized way to collect and read data across the three pillars of observability: metrics, logs, and traces.
• Adopting the project allows businesses to store that data however and wherever they like, which is much cheaper than previous options allowed and also lets that data be accessed by a variety of analytics engines.
• Snowflake plans to blend Observe's technology into its core data platform to let customers apply the same analytics and AI tools they use for corporate data to their telemetry data.

There are signs that homegrown enterprise AI agents are finally starting to get traction, but they won't really take off until companies feel comfortable they can manage the complexity. Observability will be an important piece of that puzzle, especially as the data produced by agent activity threatens to make older enterprise transitions look quaint.

• "As AI reshapes how applications are built, the bottleneck has shifted from writing code to operating and troubleshooting complex systems in production," Observe CEO Jeremy Burton said in the press release announcing the deal.
• Coders on the bleeding edge of AI development (like Sourcegraph's Steve Yegge) are already looking for ways to manage running multiple agents at the same time, and similar workflows could soon find their way into other types of enterprise software.
• That won't be possible without cheap data analysis tools and trustworthy feedback loops, which Snowflake thinks it can provide with the Observe deal.
• And it's yet another sign that the days of the standalone "best of breed" enterprise software applications are probably numbered as AI collapses distinct categories of software into broader packages.

Time to git going

Between Visual Studio Code and GitHub, Microsoft has enjoyed a decade's worth of developer mindshare that presumably funneled a few bucks into Azure and other enterprise products and services. But honest-to-god platform shifts have no sympathy for their ancestors, and the generative AI revolution Microsoft spent so much time and money promoting over the last several years appears to be hitting the company in a delicate place.

Business Insider reported Thursday that Microsoft is moving personnel into the now quasi-independent GitHub organization after deciding that AI-powered coding upstarts like Cursor are having an impact on its standing among developers. In August Microsoft announced that GitHub would become a more integrated part of its CoreAI division, but "Microsoft and GitHub have remained somewhat separate" in the intervening months, according to the report.

The stakes are enormous; the company that produces the best daily driver for software development stands to reap enormous rewards from that effort. Microsoft's success building compelling and open software development tools burnished the company's image among a generation of developers that remember its quixotic battles against open-source software, but the vibes may have shifted.

Enterprise moves

Hossein Nowbar is the new president and chief legal officer of ServiceNow, following nearly 30 years in legal leadership roles at Microsoft.

Leonard Speiser is the new chief operating officer at Lambda, joining the neocloud after several years in investment roles.

Scott Manchester is the new chief product and technology officer at Nerdio, joining the Microsoft partner after 25 years at Microsoft.

Alison Wagonfeld is the new chief marketing officer at Nvidia, after nearly 10 years in marketing leadership roles at Google Cloud.

Kady Srinivasan is the new chief marketing officer at Freshworks, joining the IT management company after marketing leadership roles at You.com and Lightspeed Commerce.

Adnan Chaudhry is the new chief revenue officer at Ping Identity, following several years in sales and operational leadership roles at Slack and Salesforce.

The Runtime roundup

Crowdstrike acquired SGNL for $740 million as it expands into the identity security market.

A maximum security flaw in the widely used n8n open-source workflow automation platform could allow attackers to compromise n8n servers without having to obtain legitimate login credentials.

Thanks for reading — see you Saturday!