As token costs keep climbing 📈, where else can we find efficiency gains? Partner @SpencerFarrar dug into how @sailresearchco is answering this question with more efficient inference and long-horizon agents. We're proud to back their Series A.
00:00
Theory Ventures
247 posts
Theory Ventures
@Theoryvc
Theory Ventures is an early-stage venture capital firm based in SF. We invest $1-25m in software companies that leverage technology discontinuities.
Joined August 2022
- Theory Ventures reposted
Replying to @lancedb and @aiDotEngineer - The most important companies will soon run on agents that work for hours or days at a time. Today's inference stack, built for a human waiting on a chatbot, can't serve them. That's why we're excited to announce our investment in @sailresearchco, enabling inference as a
- Theory Ventures reposted
Samir Menon @blintzbase and I are thrilled to announce Sail @sailresearchco ! We build infrastructure for long-horizon agents: inference served at unbeatable prices-per-token for open models, plus sandboxes designed to run for days, weeks, or longer. We've raised $80M, w/ our
- The Anthropic Mythos/Fable saga made one thing clear: AI agents are cyber experts at both offense and defense, and as they get smarter, the stakes get higher. The same models that can find and fix vulnerabilities can also be used to exploit them. So what does that mean for the
- Theory Ventures reposted
Chat about the next-gen of AI infra over some 🏓, how else would you want to spend your night after a long day at @aiDotEngineer World's Fair? 😄 Join @lancedb, @Theoryvc, and @ollama for a night of food, drinks, and competition! Tournament?? 👀 - When @ethanyish set out to build @AlliumLabs, good data didn't exist for digital assets. Now, Allium serves as the trust and transparency layer that the industry was long missing. Congratulations again to the team on their $40M Series B today and all the momentum!
00:00 - Every onchain product a bank ships depends on a trustworthy record. That record runs on @AlliumLabs. When we led their Series A, we bet they’d become the trusted data layer for onchain finance, and they have: revenue up 10x and 150+ enterprise customers. From tracking
- In town for AIE? Excited about the boundary between local and cloud AI infrastructure? Come for some back-and-forth with LanceDB, Ollama, and Theory Ventures at Spin on June 30th! 🏓 Food, drinks, and fierce competition: 📅 Tuesday, June 30 ⏰ 6:00–9:00 PM 📍 SPIN San Francisco
- Did you miss our GTM survey on Friday? After interviewing 111 founders, we found that AI expectations have grown but outcomes haven’t, there are fewer AI-driven layoffs than expected, performance is directly correlated to AI-deployment, and buyer-side AI changes objections.
- 111 founders. 25 questions on GTM. We discovered: AI expectations have grown but outcomes haven’t, the "AI = layoffs" narrative is mostly wrong (only 7% of founders frame AI as cost-cutting, companies cutting SDRs are actually underperforming), performance is directly correlated
- Is RL just fine-tuning under a new name? @tuomars( @nvidia ), @zedlewski( @togethercompute ), and @charles_irl ( @modal ) joined our Head of AI @BEBischof onstage at @AICouncilConf to answer this question. We also dug into how smart teams are compressing inference costs, who
