Some companies chase uptime. Resolve AI chases the moments when production goes quiet, engineers hold their breath, and the truth lives somewhere between logs, metrics, and the Slack thread nobody...
Some companies chase uptime. Resolve AI chases the moments when production goes quiet, engineers hold their breath, and the truth lives somewhere between logs, metrics, and the Slack thread nobody bookmarked. On February 4, 2026, Resolve AI, Inc. came out swinging with $125M Series A at a $1B valuation, led by Lightspeed Venture Partners with Sebastian Duesterhoeft in the mix, joined again by Greylock Partners, Unusual Ventures, Artisanal Ventures, and A*. San Francisco energy, global ambition, no wasted motion.
Spiros Xanthos, CEO, and Mayank Agarwal, CTO, did not wake up one day and decide to teach machines how to babysit prod. They met 20 years ago at the University of Illinois Urbana Champaign, built together since 2012, helped create OpenTelemetry, and shipped companies into Splunk and VMware. That kind of shared scar tissue matters when software breaks at scale. It is the difference between guessing and resolving. Name checks out.
Resolve AI lives where production actually hurts. Multi agent systems working across code, infrastructure, and telemetry. Alerts triaged without panic. Incidents investigated without 20 people piling into a bridge. Context delivered while engineers are still writing code, not after customers notice. Coinbase cut investigation time by 72%. Zscaler saw fewer engineers per incident and faster answers. DoorDash, Salesforce, MongoDB, and MSCI are not shopping for vibes. They are buying sleep.
Lightspeed Venture Partners saw the gap forming as AI cranks out code faster than humans can keep it alive. Greylock Partners stayed close from seed. Unusual Ventures and Artisanal Ventures leaned in above pro rata. The signal is not the size of the round. It is the agreement that production is now the product.
AI for prod is not a slogan here. It is a system that sits in the mess, learns the tribal knowledge nobody documents, and adapts to each stack and business rhythm. Foundation models meet custom agents trained on real telemetry, real infrastructure, real consequences. Available where teams already work, from observability tools to cloud and code, the platform resolves before humans even argue about root cause.
Resolve AI is not promising a future. It is standing inside the systems companies already depend on, learning the undocumented knowledge nobody writes down, and resolving what used to wake people up at 3 a.m. As software velocity keeps accelerating, the quiet confidence of systems that heal themselves starts to feel less like luxury and more like table stakes. That shift changes how teams hire, ship, and trust their own systems.