SignalFire

Today we published research that, as far as I know, has no precedent: the first measured census of AI agents inside live enterprise environments. Everyone in this industry is talking about agents right now. The explosion, the models, the adoption curves, the MCP debates, panel after panel after panel. I've sat through plenty of it, and nearly every conversation runs on the same fuel: anecdotes. The market has been debating a phenomenon it couldn't measure. So we spent the past months measuring it. 90 days of observed agent activity across tens of live enterprise environments, millions of endpoints, every major sector. Real telemetry from the machines where agents actually run. I'm going to allow myself one superlative, because our research team earned it: this is groundbreaking work. For the first time, the market can KNOW instead of guess: - AI agents outnumber engineers 3.6 to 1 in the typical enterprise, and the population doubles every ten weeks - The average agent uses 15 distinct credentials - 78% of enterprises have agents touching production systems, and in 83% we caught agents running with every approval prompt disabled - 73% of agent tool activity bypasses MCP entirely What it means, in one line: a second workforce has arrived, it holds real keys to real systems, and the identity program was built for the other one. I broke down the findings and what I'd do about them on the blog: https://go.clut.ch/f09d1d The full report: https://go.clut.ch/7a99f6

When asked how SignalFire grew from $0 to ~$3 billion AUM in a decade, I say it's these 3 obsessions below about how we apply AI. I’ll be sharing the lessons on Grace Gong’s Venture With Grace. Hit ‘View event’ to join us, but here's a preview: 1. Venture firms' support should be as scalable as the companies they fund, building self-serve software in addition to human teams so we're not stretched thin as the portfolio grows 2. Venture can be a science, not just an art, because it's possible to track the biggest determinant of startup success: talent. If AI lets you see that the best builders are joining a company, you know they're creating something special 3. Venture firms must adapt their fund structures and strategies as fast as AI is transforming the landscape. If you're deploying capital in exactly the same way you did 10 years ago, you're almost assuredly wrong. Reinvention is the only alternative to being replaced! https://lnkd.in/gTWYBNXz

Chris Farmer is the Founder and CEO of SignalFire, one of Silicon Valley's leading venture capital firms built around proprietary data and AI. ~~~~~~~~~~~~~ This episode is brought to you by Nebius — the ultimate cloud for AI innovators. Nebius provides AI infrastructure you can count on, combining reliability and speed with flexibility and engineering support unmatched by hyperscalers. AI leaders like Meta, Shopify, and Higgsfield already partner with Nebius to run their AI workloads. Plus, venture-backed startups can save up to $150,000 on compute costs when they apply for access. Visit nebius.com or nebius.com/startups to learn more ~~~~~~~~~~~~~ Since founding the firm in 2013, he has pioneered a data-driven approach to venture investing, backing category-defining companies across enterprise software, AI, infrastructure, and consumer technology. SignalFire combines machine learning, proprietary talent data, and operational support to help founders build and scale high-growth companies. Before launching SignalFire, Chris was a Venture Partner at General Catalyst, where he established the firm's California office and led early investments in companies including Stripe, Coinbase, Discord, Fivetran, Zapier, Venmo, ClassDojo, and Alation. Earlier, he was a Vice President at Bessemer Venture Partners, a consultant at Bain & Company, and a successful entrepreneur who led the turnaround of SkyBitz into a profitable SaaS business that was later acquired. Chris has also published research on venture capital in Harvard Business Review and other leading journals, bringing decades of experience at the intersection of technology, entrepreneurship, and investing. Topics - How AI is transforming venture capital: building a data-driven investment firm for the next generation of startups - Lessons from backing Stripe, Coinbase, Discord, and other category-defining companies at the earliest stages - What founders need to build AI-native companies that scale from seed stage to market leadership #ArtificialIntelligence #VentureCapital #Startups #AI #Entrepreneurship

Today we announced Credible's $10 million in seed financing, backed by Gradient, SignalFire, and K5 Global, with angels including Godard Abel, Wes McKinney, Alexander Dean, SV Angel, and others. Here's the mission they're backing: make data matter. Data only matters when the people and AI using it understand what it means. I spent 17 years at Google building data analytics infrastructure and watched the complexity pile up -- sprawling pipelines, overlapping tools, meaning littered across systems. That was hard enough for people. For AI, it's untenable: without context, you get confident answers no one trusts. We started Credible to build a simpler, more elegant data stack for the AI era: capture what your data means, and deliver your data and its meaning to every surface where decisions get made -- fast, secure, and efficient. Built on open-source Malloy, because meaning belongs to the organizations that create it. To our investors: some of you built the tools the data world runs on every day. That you see the same gap we do, and trust this team to close it, means more than the capital. Thank you. Thank you to lloyd tabb and Michael Toy, our Malloy collaborators. Credible stands on the shoulders of giants -- and the language you built. Getting to build with you is a privilege. Thank you to our early customers, who bet on us when the product was young and have made it better every week since. Teams like VideoAmp, whose AI now answers to some of the largest advertisers in the world -- with Credible underneath so every answer is grounded and defensible. And thank you to the team in Boulder. Small crew, hard problem, zero shortcuts. Watching what you ship every week is the best part of this job. See article in the comments.

Rain Capital x SVB x SignalFire Cyber Dinner Last night, Rain Capital had the pleasure of co-hosting a cybersecurity dinner alongside SVB and SignalFire, bringing together an incredible group of founders, practitioners, and investors ahead of #BlackHat. From AI-native security to agentic infrastructure to risks around rogue agents, the discussion reflected both the pace of innovation and the importance of building alongside your customers. Thank you to everyone who spent the evening with us. We’re grateful to be part of such a collaborative ecosystem and excited to continue supporting founders building the future of cybersecurity. See you at Black Hat!

Chris Farmer got banned from casinos before he ever raised a fund. He was a card counter. Semi pro blackjack for years, until they stopped letting him play. He told me the whole story on this week's Very First Win. What stuck with me is how he thinks about a bad beat. When you lose big early in a deck, you have to ask: did I just get unlucky, or are all the aces still in the back? One means lean in harder. The other means you're chasing losses. He applies the same question to every struggling startup. Is this a timing problem you can outlast, or is the deck just empty? Chris went on to work at Bessemer Venture Partners and General Catalyst before founding SignalFire, one of the most data driven firms in the business. We got into when to double down vs when to walk, how AI is changing what a winning team looks like, and why entry level jobs are disappearing while engineering hiring goes up. Full episode linked in the first comment.