Ravi Mhatre co-founded Lightspeed and has built a reputation as a "long-cycle" partner who backs infrastructure companies like Nutanix and Rubrik from early stages through scale, often serving on boards for over a decade. He combines his operating experience in enterprise technology with a "conviction wins, depth compounds" philosophy, going very deep on a small number of founders and markets rather than chasing short-term trends.

• A strong conviction that safety and trust are prerequisites for scaling artificial intelligence, rather than obstacles to it. He frequently argues that the most successful AI companies will be those that prioritize reliability and security from day one, citing Anthropic as a primary example of a company that earned a massive valuation by focusing on building systems the world can trust.
• Focuses on the infrastructure and hardware layers necessary to unclog AI bottlenecks. Beyond just software, he highlights investments in companies tackling fundamental limitations like the speed of chip design, noting that better models and data are useless without the physical compute capacity and specialized processors to run them efficiently.
• Values technical pedigree and founder provenance, particularly teams emerging from major tech research labs. He often points to the specific backgrounds of founders, such as the original creators of PyTorch leaving Meta to build new inference platforms, as a strong signal that a startup has the capability to execute on complex infrastructure visions.
• Interest in enterprise-grade AI applications that solve data fragmentation and security. He promotes tools that help businesses actually utilize their internal knowledge, such as platforms that use context graphs to organize corporate data, viewing this as critical infrastructure for making generative models effective in a professional environment.
• Investments in physical AI and scientific discovery. He expresses excitement about foundation models moving beyond text and into robotics and the physical world, referencing companies building brains for robots or integrated systems for scientific research that combine software models with physical lab operations.