←
Greenfield Partners107 posts
Greenfield Partners avatar

Greenfield Partners

@GreenfieldVC

Greenfield partners with disruptive tech businesses globally at the Early Growth (B/C) stage, and then applies its Greenfield Growth Momentum (G2M) playbooks.

◷ Joined July 2025▧ 30 Photos and videos
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Nov 9

More than $10B in commitments over three quarters, from just 14 of the 1,000+ customers our portfolio company @VAST_Data serves. Each of those 14 has committed over $100M to VAST-powered infrastructure this year, four of them over $1B each. Nine of the top ten frontier AI labs now run on VAST. Kudos to the VAST Data team. As the industry rebuilds its stack for agents and, increasingly, robots, the builders operating at the largest scale are choosing the VAST AI OS as their foundation. That kind of trust is earned through years of relentless execution.

Post attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Nov 2

3. The growing competitiveness of open-weight models is a major tailwind for founders. Open-weight models are closing the gap, letting enterprises fine-tune their own and control their data. For founders, a competitive open-weight ecosystem is existential.

Post attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Nov 2

2. Inference could be Israel’s biggest opportunity in AI. Frontier training takes capital few can match, unlike the capital-light innovation where Israel excels. Inference gives Israeli companies a path to the world stage.

Post attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Nov 2

1. We shouldn’t take today’s enormous compute requirements as a given. Training a frontier model takes ~10²⁶ operations, and inference adds 5× that, much of it discarded. The opportunity is using compute more efficiently, not assuming we need more.

Post attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Nov 2

What comes next for AI infrastructure, and where will the next wave of innovation be built? At our annual Future of AI Infra fireside, @AmnonShashua and Shachar Fienblit of @VAST_Data took on that question. Three ideas that stood out 🧵

Post attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Oct 26

It was our pleasure to host Prof. @amnonshashua and Shachar Fienblit for breakfast and a fireside conversation at our annual Future of AI Infra event. Amnon (Co-Founder & CEO of @doubleai ) sat down with Shachar (Co-Founder & Chief Research & Development Officer of @VAST_Data), moderated by Managing Partner Shay Grinfeld. Together, they dug into what comes next for AI infrastructure and where the next wave of innovation will be built. Thank you to Amnon, Shachar and every founder and AI researcher who joined us that morning. We'll share the key takeaways from their conversation later this week. Building something exciting in AI Infra? Our door is open, and we'd love to hear from you

Post attachmentPost attachmentPost attachment
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Oct 21

What does it take to build effective AI tools across a company? We were joined by @sebgoddijn , former Internal AI Lead at Ramp, to discuss how he built Ramp Glass so quickly and what he believes are the fundamentals for building useful AI tools. 1. Spend the time defining context - it’s the foundation of everything. 2. You don’t need to build something complex to create meaningful value – focus on memory, skills and connectors. 3. Buy infrastructure that helps you build faster 4. Be clear on ROI before you scale – start with something small and trackable Full webinar in the comments.

1:21
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Oct 20

Congrats as well to co-founders Shai Shalev-Shwartz, Yoav Levine, Or Sharir, Noam Wies and Gal Beniamini, and the whole @doubleai team on taking Maxwell from research to launch. The full story in @WSJ: wsj.com/cio-journal/nvidia-b…

Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Oct 20

How it works: Maxwell watches code as it runs and rewrites the machine instructions for the exact chip and data. No source code changes. doubleAI showcases: - 38× faster data pipelines - 4.25× faster AI inference kernels - 1.87× faster 4K video generation

1:12
Greenfield Partners avatar
Greenfield Partners@GreenfieldVC·Oct 20

What if code could think while it runs? Our portfolio company @doubleai has launched it's first product, Maxwell, designed to tune software to the exact chip and data it meets, with every optimization formally verified. Congrats to @AmnonShashua and the team 👏

Post attachment