The Closest Thing Yet to Sam Altman's One-Person Unicorn
Polsia just raised $30M at a $250M valuation. One human. Zero employees. And the AI ran the fundraise.
In September 2023, Sam Altman told Alexis Ohanian he expected the first one-person billion-dollar company to arrive within a few years, something he called “unimaginable without AI.” He started a betting pool with tech CEO friends over when it would land. Two and a half years later, Ben Cera’s Polsia just closed a $30M Series A at a $250M post-money valuation with one human and zero employees. As far as we can find, that is the highest valuation any truly solo company has ever crossed, putting Polsia closer to Altman’s prediction than anyone else.
Ben’s announcement opened with the sentence most founders would have buried in the footer: “Polsia handled the data room, briefed investors, ran the back-and-forth on diligence. I joined the final calls. That was the job.” The product is an orchestration layer that lets AI agents run an entire company, coding, marketing, support, outreach, without employees. The “job” was Ben proving it works by using Polsia to run his own fundraise, in public, on a live dashboard anyone could watch. The proof was the product.
The Founder and the Company
Ben Cera has done this twice before. Once badly. Once at scale.
Engineering degree from Columbia. First job out of school was as a quantitative trader at Barclays, which he found boring enough that he taught himself to code on the side. He shipped a string of consumer apps, then joined CloudKitchens, Travis Kalanick’s post-Uber venture, where he built Future Foods. That business scaled to $100 million in revenue. By the end of it, Ben had 400 people reporting to him across multiple geographies. Travis became his mentor; the operating philosophy he attributes to that period is “push through the pain.”
Then he tried again, this time with a co-founder, building a gifting company. It lasted two years. Every decision required alignment. Momentum died. He came out the other side with one conviction: next time, he would build alone.
That was the prologue. The story starts in 2024, when Ben spent twelve to sixteen hours a day vibe-coding with whichever model was current. He was building what he initially called Company OS, an operating system that could run a company. The pieces existed, but the agents were not yet good enough to chain them together reliably. When Anthropic’s Opus 4.5 shipped in December 2025, Ben called it his “game over moment.” The model was finally strong enough to be the CEO of a workflow.
Polsia launched on December 15, 2025. Users describe a business idea, and Polsia provisions infrastructure (a Render web server, a Neon database, a GitHub account, a Stripe account, a Meta Ads account, an email address), generates a website, runs ad campaigns, handles support, and emails the user a structured report every morning summarizing what it did overnight.
Five months later, the platform is approaching $10 million ARR. There are more than 7,600 companies running on it. Daily active users sit at 3,627. Month-two retention is 85 percent. Ten percent of signups convert to paid. Ben is still the only human at the company.
Tony Conrad at True Ventures led the pre-seed in March and wrote the public conviction piece. Sound Ventures, Vaynerfund, Tekton Ventures, Adjacent, Offline Ventures, and Drysdale Ventures joined for the Series A.
The Fundraise Was the Demo
The cleanest demo Ben could have given any of those investors was the one they all watched in real time. A public URL, polsia.com/live, where the AI agent doing his investor outreach was visible to anyone who wanted to look.
The mechanic was straightforward. Polsia’s agent identified investors who matched the profile, drafted personalized outreach, sent it, parsed the replies, scheduled the calls, prepared the data room, and answered diligence questions. Ben joined the conversations where signing the term sheet required a human. Every other step ran without him.
He described the design intent in his announcement post: “The round itself is the clearest example of what I built. Polsia handled the data room, briefed investors, ran the back-and-forth on diligence. I joined the final calls. That was the job. That is not a marketing line. It is the product.”
This is the rare case where a founder’s marketing claim and operational truth collapse into the same artifact. Investors who wrote the check did not have to take Ben’s word that Polsia could run a company. They had personal evidence: it ran his.
That is what dogfooding looks like when the product itself is autonomous. Not a screenshot in a deck. Not a video walkthrough. The investor sitting across from the founder reads an email the founder did not write, scheduled by an agent the founder did not instruct, summarizing diligence the founder did not personally handle. Then they sign the term sheet.
Know which funds are actively deploying right now? Fundmomentum.vc tracks freshly raised VC funds (930+) and personal investor signals across the globe - so you know exactly who has dry powder and is ready to write checks. If you’re raising, stop guessing which investors are the right fit.
The Raise
The Series A was $30 million on a $250 million post-money. Sound Ventures led. True Ventures rolled forward from the March pre-seed. Offline Ventures, Adjacent, Tekton Ventures, Drysdale Ventures, and Vaynerfund joined.
The cap table is not assembled the way most B2B AI Series As are. Sound Ventures is Ashton Kutcher and Guy Oseary’s firm. Vaynerfund is Gary Vaynerchuk’s. These are not the funds that show up to underwrite an enterprise SaaS ARR multiple with deep cohort analysis. They are funds whose distribution surface is podcasts, X, LinkedIn, and creator-economy adjacent culture.
This is the cap table for a product whose primary acquisition channel is build-in-public storytelling on the platforms those investors already amplify. The architecture is deliberate, not opportunistic. If your product’s growth depends on the founder going viral with revenue milestones and one-person-company numbers, the optimal investor list is the one whose audiences will retweet you. Ben picked exactly that list.
True Ventures sits in a different role on the cap table. Tony Conrad led the pre-seed personally in March and published a 500-word conviction essay the same week. “While venture capital typically rewards pattern recognition,” Conrad wrote, “we’ve always believed the truly transformative bets are found in pattern breaking.” His sign-off line was unusually direct: “The solo Founder was a myth until Ben made it a reality. We’re proud to be the first to believe in Ben and Polsia.”
That essay carried a specific operational weight. When Tony writes a “first to believe” post with the founder’s name in the headline, the implicit promise is that True will continue conviction through the Series A. True did. Their roll-forward into the new round was a cross-round diligence signal that other Series A participants priced into the round mechanically, not analytically. The pre-seed lead’s continued participation is the single most credible piece of due diligence the next round can offer; when it is absent, every new investor reads its absence. With True, that question never even surfaced.
The Series A closed roughly sixty days after the pre-seed. That cadence is fast even by 2026 AI standards. The compression was made possible by two things: the public revenue ramp ($1M ARR by day thirty, $3M by day sixty, $4.5M by mid-March, $6.3M by May 8, approaching $10M by close), and the live fundraise dashboard, which let new investors observe the product in operation against the function they were evaluating.
Ben put almost no money into marketing the round itself. The product marketed it
Raising (soon)? I’ve helped 50+ startups to accelerate their fundraising campaigns and match them with the right investors. If you want hands-on support - from pitch refinement to warm intros to term sheet negotiation - let’s talk.
What Happens When You Run Ten Thousand Companies On One Platform
The growth has not been frictionless. Operating that many autonomous companies through a single platform produces friction that does not exist when humans are in every loop.
Some users have flagged tasks marked complete that did not actually execute. Customer service responses have been uneven. The platform is expensive for users on the low-income end of the addressable market; Rest of World profiled a Chinese factory safety trainer paying roughly a quarter of his monthly salary for a subscription and getting modest results. Several respected operators on X have publicly questioned whether some of the cosmetic outputs, landing pages, AI-generated reviews, cold email volume, are valuable or just noise.
These are real, and they are the kind of growing pains any platform handling thousands of autonomous workflows at agent-speed is going to encounter. Polsia has shipped product fixes for several of the complaints publicly through its own AI-authored product newsletter, the one signed “Hi. It’s Polsia.” The most recent release, called God Mode, lets users buy uninterrupted agent runtime by the hour or week, which materially reduces the “task marked complete but didn’t run” class of complaint.
The honest read is that Polsia is operating at the frontier of what autonomous AI can ship into production today. Some of it works exceptionally well. Some of it breaks. The trajectory across the last five months suggests the broken parts are getting fixed faster than the platform is taking on new ones.
The Playbook, 5 Lessons
1. Run the product on the fundraise. Whatever function your product automates, use that function on your own raise. If you build sales software, run your sales on it. If you build trust infrastructure, prove it on yourself. If you build agents, let them do the outreach. The most credible demo a founder will ever ship is the one where the investor watches the product perform the job they are evaluating you on. Polsia’s live dashboard at polsia.com/live did more underwriting work than any slide deck. Make your own product the load-bearing evidence in your own round.
2. Pick investors whose distribution matches your acquisition channel. Sound Ventures and Vaynerfund are not enterprise SaaS diligence funds. They are amplification funds. Ben picked them because Polsia’s customer acquisition runs on build-in-public storytelling on X and LinkedIn, exactly the surfaces those funds amplify. If your growth depends on going viral with operational milestones, the optimal cap table is the one whose audiences will retweet you. If your growth depends on enterprise procurement cycles, your optimal cap table is different. Map your investor list against your real distribution channel, not the brand value of the logo.
3. The pre-seed lead’s roll-forward is the single most credible diligence the next round can offer. True Ventures led the March pre-seed and continued into the Series A. That fact alone removed a class of questions every other Series A investor would otherwise have raised. When your prior lead does not participate, every new investor reads the absence. Engineer the relationship with your seed lead from day one as if their follow-on participation is the most important asset of your next round. It is.
4. Build alone if a prior co-founder dynamic taught you the cost of alignment overhead. Ben spent two years on a gifting company that died because every decision needed alignment. The conviction to build alone next time was the price of that lesson. The transferable framing is not “you should build alone” (many companies need multiple founders). It is “be honest about which co-founder dynamic is actually adding velocity versus subtracting it.” If your last venture died from alignment debt, your next venture should be structured to eliminate it.
5. Know which foundational model unlock is yours, and ship the week it lands. Ben spent all of 2024 building agent infrastructure that was not yet strong enough to be useful in production. When Opus 4.5 shipped in December 2025, the underlying capability he had been waiting on arrived. Polsia launched ten days later. The lesson for founders building on top of foundation models: identify which specific capability would make your product viable, watch for it, and have the deployable product ready the week it lands. The window between “the model can finally do this” and “someone else ships first” is short and shrinking.
What’s Next
The $30 million is targeted at deeper telco and platform integrations, broader Meta and Sora ad coverage, and continued investment in the agent reliability stack. Ben’s stated long-term goal is to take Polsia to 100 percent autonomy, where the platform builds, fixes, and ships itself without him as a decision node. The aspirational milestone he has named publicly is “just spin up a Polsia” becoming as natural a phrase as “just make a Shopify store.” He wants to give the 99 percent who do not have a network or a seed round a way into the AI economy. That ambition is the next thing the round has to fund.
Want the deeper playbook, the live-fundraise-dashboard build spec, the cap-table-by-acquisition-channel matrix, the model-launch readiness checklist, and the full operational template for running an autonomous fundraise? The detailed playbook is available for paid subscribers.