S12 Bonus: The App Is Dead, Long Live the Outcome: Why Autonomous AI Agents Are Rewriting Data Infrastructure and Transforming PostgreSQL Into a Dynamic Scratch Pad with Ajay Kulkarni, Founder & CEO of Tiger Data

Episode Summary

Executive Summary: Ajay Kulkarni explains how Tiger Data evolved from an IoT platform into a Postgres-based database company by solving its own sensor-data scaling problem. The conversation covers the company’s origin, product decisions, culture, scaling lessons, mistakes, and future bets on physical-world data and AI agents.

Main Topics: Origin of Tiger Data from an IoT pain point (Priority: 5/5): Ajay and co-founder Mike met at MIT, later reunited to build an IoT platform. The team needed a better way to store sensor data, which led them to build their own Postgres extension and eventually pivot the company around that database product. Why they built on Postgres instead of starting over (Priority: 5/5): Rather than abandoning relational databases like many competitors, Tiger Data chose to extend Postgres, arguing that the model can scale if engineered correctly and that they could preserve reliability, ecosystem, and maturity. Product evolution and roadmap philosophy (Priority: 4/5): Tiger Data follows two discovery loops: scratch their own itch, then listen to market behavior. This led them from IoT to broader use cases like fintech, crypto, and now agentic workflows. Team-building and company culture (Priority: 4/5): Ajay emphasizes hiring intrinsically motivated people who get things done, then letting the culture self-reinforce. Internal rituals like Tiger Time and Tiger Cubs reflect a strong identity around execution and momentum. Scaling, leadership, and organizational change (Priority: 4/5): He says a CEO’s job changes every six months and that growth requires accepting that past habits and systems eventually become wrong. Scaling means continuously asking what is broken and adjusting. Lessons from mistakes and market timing (Priority: 5/5): Ajay admits early overreliance on expert advice and chasing shiny trends were mistakes. A key lesson was trusting instinct in fast-moving industries and betting early on cloud before it became the obvious revenue driver. Future bets: physical-world data and AI agents (Priority: 4/5): Tiger Data sees large opportunity in industries tied to hardware and infrastructure, plus making Postgres better for agents. Ajay believes agents are the next major software platform shift.

Key Arguments: The company was created by solving a real internal problem: off-the-shelf databases could not handle the sensor-data workload for their IoT platform. Extending Postgres was preferable to rebuilding from scratch because it preserved decades of maturity, stability, and ecosystem support while improving scalability. Market demand revealed a bigger opportunity than the original IoT use case; customers in fintech and crypto proved the product had broader applicability. Founders should balance listening to experts with trusting their own instincts, especially in rapidly changing markets where prior experience can become a liability. A successful roadmap comes from combining internal pain points with market feedback, not from blindly chasing hype cycles. Hiring should prioritize intrinsic motivation, scrappiness, and execution ability over raw talent alone. Scaling requires accepting that what worked six months ago may now be wrong, and leaders must adapt their level of involvement accordingly. The company’s resilience through industry ups and downs is one of its greatest strengths. Long-term success comes from conviction in a problem before it becomes fashionable, rather than chasing the next shiny trend. Tiger Data’s future will likely center on serving data-heavy physical industries and enabling Postgres for AI-agent workflows.

Data Points: Meeting at MIT: September 1997 - Ajay says he met co-founder Mike in the first week of college at MIT. Time to build database MVP: 12 months - An engineer estimated one month, Ajay estimated three months, but the first working database took a year to build. Devices being supported: Over 100,000 devices - The IoT platform needed to store sensor data from this many devices. Press spike after launch: More press in the first 3 weeks than in the previous 1.5 years - The Postgres extension attracted significantly more attention than the original IoT platform. Company age reference: 12 years later - Ajay notes the engineer who built the database is still at the company 12 years later. Postgres development span: 25 to 35 years - Ajay highlights the long development history and maturity of Postgres. Traditional sales cycle: 9 to 18 months - He describes legacy database sales as long enterprise sales-led motions. Cloud revenue milestone: About a year later - After going all in on cloud, it became the majority of revenue.

Pivotal Quotes: "We realized that the database was solving a bigger problem than the platform itself." — Ajay Kulkarni: Explaining the pivot from an IoT platform to a database company. "We built a better mousetrap because we needed it. We scratched our own itch, so to speak." — Ajay Kulkarni: Describing the origin of the product and the company’s founder-led problem solving. "The thing that I'm most proud of is our resilience." — Ajay Kulkarni: His answer about what he values most about Tiger Data as a company.

Implications: Tiger Data’s story shows that durable infrastructure companies can emerge from solving internal pain points, then expanding with market pull. For founders, the lesson is to trust conviction, adapt constantly, and build for emerging shifts like agents and data-heavy physical systems.