See the pitch deck AI observability startup Sazabi used to raise $8 million seed round from YC and J2 Ventures

Sazabi, an innovative startup based in San Francisco, has successfully raised $8 million to enhance its AI observability platform. The company specializes in a monitoring, debugging, and incident response solution tailored for agile engineering teams at startups utilizing AI coding tools such as Cursor and Claude Code. Sherwood Callaway, the founder and CEO, previously an engineer at Brex, identified a significant issue with existing observability platforms. He found that many were inundating software engineers with excessive dashboards and irrelevant datasets. Callaway emphasized the importance of log data—a digital black box that meticulously records time-stamped actions and errors within applications. "Whenever there was a problem in production, I would reach for a log stream," he stated, noting that this type of data felt most intuitive. Having emerged from stealth mode in March 2026 as part of Y Combinator's latest cohort, Sazabi's seed funding round was spearheaded by J2 Ventures, Village Global, and Y Combinator, attracting over 60 angel investors from renowned companies including Cursor, OpenAI, and Anthropic. The startup positions itself against established observability tools like Datadog and Grafana, with Callaway humorously remarking that internally, they refer to their mission as "Operation Waterloo," aimed at outperforming Datadog, a French firm. Sazabi offers a pricing model that begins with a free tier, transitioning to a credit-based system linked to token usage and log volume. The newly acquired funds will bolster its eight-member team and facilitate growth in engineering and market operations. Currently, the company has engaged paying customers through a closed alpha phase, with plans for a public launch slated for this summer. "We have a clear and unique product vision, but we don't have the luxury of time," Callaway remarked, emphasizing the urgency to deploy capital efficiently and execute swiftly.