Founder Story: Matan Grinberg of Factory
From String Theory to Startups: An Unlikely Origin
Most founder stories begin in a garage or a dorm room. Matan Grinberg's begins at the frontier of theoretical physics. He graduated from Princeton University in 2020, and while still an undergraduate he did work serious enough to land him a co-authorship with Juan Maldacena — the Institute for Advanced Study physicist whose AdS/CFT correspondence is one of the most cited ideas in modern theoretical physics. Their paper, "Proper time to the black hole singularity from thermal one-point functions," was posted to arXiv in November 2020 and published in the Journal of High Energy Physics in March 2021. It is not the kind of credential you expect on the resume of an enterprise-software CEO.
From Princeton, Grinberg went on to a PhD program in theoretical physics at UC Berkeley, continuing to work on string theory. On paper, he was doing exactly what a brilliant young physicist is supposed to do. But somewhere inside the doctoral program, the story he was telling himself started to crack. Years later he would put it with brutal self-awareness:
"I was only doing theoretical physics because it was hard, not because I actually loved it."
That single realization — that he had confused difficulty with meaning — is the hinge on which his entire life turned.
The Leap From Academia
Leaving a physics PhD is not a small decision. It means abandoning a decade of accumulated expertise, a clear status ladder, and the identity that comes with being one of the smartest people in almost any room. Grinberg didn't have a startup idea when he decided to go. What he had was a growing conviction that his talents were pointed at the wrong target, and the humility to ask for help.
He did something that sounds almost naive in hindsight: he cold-emailed Shaun Maguire, a partner at Sequoia Capital who, crucially, also holds a PhD in physics. Maguire agreed to meet, and the meeting turned into a three-hour walk. By the end of it, Maguire had laid out three concrete paths for a defecting physicist: go join Twitter, go work at one of his portfolio companies, or go start something of his own.
Grinberg's reasoning for choosing the hardest of the three options reflects the same clarity that made him leave physics in the first place. He had come to believe that:
"Code is fundamental to machine intelligence. I was naturally attracted to the role that it plays there."
For a person who spent years hunting for the fundamental structures underneath reality, software was not a downgrade from physics. It was the same instinct pointed at a different substrate — one where the equations could actually ship.
Founding Factory in 72 Hours
The catalyst arrived almost immediately. Around the time of his walk with Maguire, Grinberg attended a hackathon in San Francisco, where he crossed paths with Eno Reyes, a machine learning engineer and fellow Princetonian (Class of '21). Despite a web of mutual friends from their overlapping years on campus, the two had never met. Grinberg describes the encounter with a phrase that has since become part of Factory lore:
"Intellectual love at first sight."
The day after the Maguire walk, Grinberg and Reyes met for coffee and decided to build a company together. Then they did something that tells you almost everything about how the pair operates: over the next 72 hours, they built a working demo of Factory. Factory was founded in April 2023, with Grinberg as CEO and Reyes as CTO.
That velocity was not an accident — it was a philosophy from day one. Grinberg has since distilled it into a maxim he repeats often:
"The only way you can win is by executing faster and being more obsessed."
Building Droids: Agents That Meet Engineers Where They Are
Factory's product is a family of AI agents the company calls Droids — a deliberate nod to Star Wars, and a deliberate rejection of a word Grinberg has come to distrust. In an era when "agent" had become a marketing buzzword, he wanted to signal something more dependable:
"Agent is synonymous with unreliable, stochastic, demoware, vaporware... we want to build systems that aren't just cool examples but valuable today."
Droids are built to automate the parts of the software development lifecycle that engineers are genuinely happy to hand off. Rather than chase glossy demos of "AI writes your whole app," Factory went after the unglamorous, high-friction work that eats real engineering hours:
- Testing, debugging, and refactoring
- Large-scale code migrations
- Code review — which Grinberg memorably compares to a trip to the DMV
- Documentation and codebase Q&A
- Incident response and feature development
The strategic insight underneath this is that Factory positions its Droids with the developer, not against them. Grinberg is explicit that the goal is to "position ourselves aligned with the developer instead of antagonistic with them" — a stance that matters enormously for adoption inside large engineering organizations where tools imposed from above tend to die quietly. Just as important, Factory is deliberately model-independent, abstracting away which underlying foundation model powers a given task. Grinberg is candid about why: "Training foundational models — we're not going to win that battle." By refusing to run a race unsuited to Factory's abilities, he freed the company to win the one that was.
Defining Moments: A Rapid Climb to Unicorn Status
Factory's funding history reads like a compressed time-lapse of the AI-coding boom, and each round marks a defining moment in the company's trajectory:
- Late 2023 — Seed (~$5M): Backed early by Sequoia Capital, the firm whose partner had sent Grinberg on his fateful walk.
- 2024 — Series A (~$15M): Fuel for building out the enterprise product.
- September 25, 2025 — Series B ($50M at ~$300M valuation): Led by NEA and Sequoia, with participation from NVIDIA and J.P. Morgan, plus notable operators including Frank Slootman, Nikesh Arora, and Aaron Levie. Factory formally "unleashed the Droids" to the market.
- April 16, 2026 — Series C ($150M at $1.5B valuation): Led by Khosla Ventures, with Sequoia, Insight Partners, Blackstone, Evantic Capital, 20VC, NEA, and Mantis VC participating. Khosla's Keith Rabois joined the board.
In roughly three years, Factory went from a coffee-shop handshake to a unicorn. The Series C didn't just mint a valuation; it validated a thesis. By that point Factory reported that revenue had doubled month-over-month for six consecutive months, that hundreds of thousands of developers were using Droids daily, and that its enterprise roster included NVIDIA, Adobe, EY, Palo Alto Networks, Adyen, and Morgan Stanley. The company also claimed the #1 position across leading software-development agent benchmarks — a metric that matters to the engineering buyers Factory is selling into.
Innovation Philosophy: Conservation Laws and the Token Budget
What makes Grinberg an unusual technology CEO is that he thinks like a physicist even when he's talking about go-to-market. He has cited Emmy Noether's theorem — the deep result linking symmetries to conservation laws — as a lens for reasoning about what stays invariant while everything around it changes. In a field where the underlying models are re-shuffled every few months, asking "what is actually conserved here?" is a genuinely strategic question, and it explains Factory's bet on being model-agnostic.
That same first-principles instinct shows up in what Grinberg calls the "token budget problem." As enterprises adopt AI agents, leaders are being forced to trade off three resources that used to live in separate spreadsheets: headcount, compute, and tokens. Grinberg argues the winning organizations will be the ones that reason about these as a single budget rather than three silos. He is also notably allergic to AI doom-mongering, pushing back on the fashionable claim that "all jobs are going away" as both inaccurate and, in his words, psychologically harmful. His framing is more grounded: Factory exists to "raise the leverage" of each software engineer — not to erase them, but to give them a force multiplier.
Industry Impact and What Comes Next
Factory sits at the center of one of the most consequential shifts in software: the move from AI autocomplete — helpful suggestions inside a human-driven workflow — toward AI autonomy, where agents own entire tasks end to end. Grinberg's thesis, that autonomy belongs in enterprise software engineering rather than just in flashy consumer demos, has proven prescient as the biggest technology companies race to reorganize their engineering orgs around agents.
His discipline about the enterprise is a differentiator. Plenty of startups can produce an impressive coding demo; far fewer can survive the security reviews, the messy legacy codebases, and the skeptical senior engineers of a Fortune 500. By insisting that Droids "be useful in the enterprise and meet engineers where they are with tasks they're eager to automate away," Factory chose the harder, stickier market. Grinberg's read on his own competitive moat is also unsentimental: "If you're unable to build a product that people are actually using with these incredible models, then chances are fine tuning and training will not save you." In other words, the durable advantage is not a proprietary model — it is real usage, real workflows, and real trust inside real companies.
Closing Thoughts
There is a neat symmetry to Matan Grinberg's story that Emmy Noether might have appreciated. He spent years trying to understand black holes — regions where the ordinary rules break down and information seems to vanish — and then walked away to build systems that impose order on a different kind of chaos: the sprawling, half-documented, migration-averse codebases that run the modern enterprise. The physicist's habits never left him. He still asks what is fundamental, what is conserved, and which battles are worth fighting.
What makes his arc instructive for other founders is not the physics pedigree — it's the honesty that preceded the pivot. Grinberg's willingness to admit that he was chasing difficulty rather than meaning, to cold-email a stranger for help, and to commit to a co-founder and a company within a single week, is a case study in acting decisively on self-knowledge. From a three-hour walk to a $1.5 billion company in roughly three years, the throughline is a founder who treats speed and obsession as first principles. Factory's Droids may end up reshaping how software gets built, but the more durable lesson is simpler: the hardest problem worth solving is usually the one you actually care about — and the courage to switch toward it can compound faster than any model.
Your Story Could Be Next
Every founder in this series started exactly where you are now — with an idea and the will to build. What has changed is the leverage. Frederick gives you a team of AI agents that take you from idea to a real company, helping you design, build, and ship software on demand — no giant team and no endless runway required.
If Matan Grinberg's journey resonated, there is no better time to begin your own. Start building your company with Frederick.
References
- https://paw.princeton.edu/article/young-alumni-shake-business-world-ai
- https://sequoiacap.com/podcast/training-data-factory/
- https://www.generalist.com/p/the-token-budget-problem
- https://factory.ai/news/series-c
- https://www.businesswire.com/news/home/20250925993478/en/Factory-Unleashes-the-Droids-Raises-$50-Million-Series-B-from-NEA-Sequoia-Capital-NVIDIA-and-J.P.-Morgan
- https://arxiv.org/abs/2011.01004
- https://www.fastcompany.com/91279593/factory-ai-code-generation-matan-grinberg-eno-reyes