Kleiner Perkins

Inference is becoming the most important infrastructure market in AI. As models reason longer and agents take on more work, the economics will increasingly come down to how many tokens a system can produce per dollar and per watt. Etched is building for t

For the last decade, the satellite industry chased affordability, and the consensus was that smaller satellites were the only way to achieve it. The problem is that every mission in space comes down to three things, power, aperture, and mass; smaller sate

For years, the satellite industry assumed smaller meant cheaper. K2 Space took the opposite bet. At K2's headquarters in Torrance, Lucas Oliveira sat down with co-founders Karan and Neel Kunjur to discuss why they chose to build bigger, how they developed

The next breakthroughs in semiconductors, batteries, carbon capture, and clean water depend on materials that have yet to be discovered. CuspAI is building an AI-powered search engine for materials. Its platform works backward from the properties a materi

Most big leaps in technology come down to a material. Polymers, steel, silicon. The next set - cheaper carbon capture, better batteries, faster chips, cleaner water - is stuck waiting on materials nobody has discovered yet. And the old way of finding them

The U.S. needs to reindustrialize, but we lost the ability to build. The Empire State Building took 13 months to stand up. Today, an equivalent skyscraper takes six or seven years to build. Construction is the rare industry that has gotten worse over time

Glean is the Work AI platform that connects a company's knowledge, systems, and context, so its assistant and agents can actually get work done instead of just returning links. What started as a better way to search across all your workplace apps has grow

Sail Research is building the inference platform for long-horizon, asynchronous agents. In this conversation, founder Neil Movva and KP partner Aditya Naganath discuss why agentic workloads are changing inference, how cost and scale become the bottleneck,

Inference is already a $100B+ market and one of the fastest growing areas of AI spend. As AI moves from demos into production, the amount of compute required to run intelligence at scale is growing quickly, with agentic workloads driving much of that dema

AI has promised, for years now, to solve our hardest problems. We think the most direct path to that promise runs through AI itself - specifically, AI that can accelerate its own research. And for the first time, there are real signs of models doing meani