Gradient Ventures

Gradient does not walk into a room and announce itself. It shows up in the cap table before the noise hits, when the product is still half-formed and the conviction has to do the heavy lifting. Born in 2017 inside Google, Gradient started as a quiet experiment in backing AI before it was fashionable to say it out loud. Anna Patterson, an engineer with real weight in search and machine learning, helped set the tone early: if the model is not the product, you are already late. That DNA stuck, even as Gradient stepped out into its own identity, raising capital beyond Google and sharpening into a seed fund that treats AI like oxygen, not decoration.

Today, General Partners Darian Shirazi and Zach Bratun-Glennon operate with a simple bias: get in early, get in technical, and stay close when things get weird. Around them, Andrew Brackin, Clayton Petty, Denise Teng, Eylul Kayin, and Vig Sachidananda bring range across infrastructure and application layers, while Operating Partner Kyle Duffy works the ground game where most firms disappear after wiring funds. This is not a tourist group taking pictures of AI. This is a crew that understands what breaks between model and market, and sticks around long enough to help fix it.

Gradient invests where the signal is still faint. Pre-seed and seed dominate the playbook, with Fund V, a $220M vehicle, doubling down on the idea that the earliest check often carries the most leverage. The focus stays tight: AI-native companies only. Not AI sprinkled on top of SaaS like parsley on a bad steak, but systems where the model, the data, and the workflow are inseparable. Infrastructure, developer tooling, applied AI across operations and productivity, all fair game as long as the product could not exist without the underlying intelligence.

The portfolio tells that story without needing theatrics. A multi-fund track record, a decade inside the AI curve, and a spread of companies that have gone on to raise serious follow-on capital or carve out real market position. Gradient tends to show up before consensus forms, when categories still feel like rumors and the founders are closer to whiteboards than press releases. That timing is not luck. It is pattern recognition built from staying in one lane while everyone else keeps switching highways.

What separates Gradient is not access. Plenty of firms have capital. It is discipline. AI-only. Seed-first. Technically literate enough to call nonsense early and double down when the architecture actually matters. The Google origin gave it early gravity, but the spinout proved something more important: the model works outside the mothership. Independent, focused, and still wired into the conversations that shape where AI goes next.

For founders, the message is clear without being loud. If you are building something where AI is the core engine, not the marketing line, Gradient is already thinking in your direction. For operators, engineers, and researchers, the opportunity sits one layer deeper. These are companies being built from first principles, and they are hiring across roles, stages, and geographies right now.

If you want in, start with the portfolio. The Gradient jobs hub is live with roles across engineering, product, and go-to-market. That is where the real action is, inside companies shaping how AI actually gets used, not just talked about.

Follow this firm. Study their founders. Track their plays.