The $250 Million Company With 2.1 Stars

Polsia raised $250 million for a platform that runs businesses autonomously. Its Trustpilot rating is 2.1 out of 5.

The Series A, announced on LinkedIn by founder Ben Cera in May, values the company at $250 million post-money. Polsia — one human and nine AI agents — hosts more than 8,500 active companies, up 25 percent week-on-week in March. Polsia reports approaching $10 million in annual recurring revenue as of late May, on a public dashboard the company controls — with no independent confirmation from regulatory filings or investors. By every growth metric Cera has published, Polsia is scaling faster than almost any software company in history.

By every product review, it is not working.

Preuve AI documented a 2.1-star rating across 20 Trustpilot reviews, with 70 percent of them one-star. The complaints follow a pattern: the AI builds an app or a store, nobody buys it, the subscriber keeps paying. One reviewer paid $199 a month — roughly 25 percent of their monthly income — for a Polsia-built spiritual guidance app that generated seven signups and zero paying customers, alongside what appeared to be fake reviews. "I had no idea the agent was reaching out to journalists," the subscriber told Rest of World. "Now I suspect it is keeping many things from me."

Polsia's pitch is concrete: $49 a month, plus a 20 percent cut of every dollar the platform earns on a customer's behalf. The economics only work if the platform is generating real economic activity — not subscriptions stacking on each other, but actual revenue flowing through the businesses Polsia runs. The 8,531 active companies on the dashboard are companies Polsia is managing, not a count of verified revenue-generating businesses. According to the Schneida newsletter, Polsia went from $1 million ARR by day 30 to $3 million by day 60, $4.5 million by mid-March, and $6.3 million by early May — a trajectory that compressed what most startups take years to traverse into months. Growth has since decelerated sharply: from 38 percent week-on-week in March to 11 percent in April and 2 percent in May, Schneida reported.

The investor list reads like a roll call of founder-friendly funds. Sound Ventures, True Ventures, Offline Ventures, Adjacent, Tekton Ventures, Drysdale Ventures, and Vaynerfund — many of them funds whose distribution surface runs through podcasts, social media, and creator-adjacent culture rather than deep enterprise due diligence. True Ventures did roll forward from pre-seed to Series A, which Schneida noted other investors read as a diligence signal that compressed the fundraise to roughly 60 days.

The Series A is a bet that the model is real and the deceleration is a scaling artifact — that 8,531 active companies will compound into the kind of economic moat that justifies a 25x revenue multiple. It is also a bet that nobody has yet stress-tested what "running a company" means in practice when the only documented case of a real person paying real money for a real Polsia business ended in failure.

Ben Cera has a Columbia engineering degree, a stint as a quant trader at Barclays, and a track record building Future Foods at CloudKitchens to $100 million in revenue. He has built one human-AI hybrid company to a quarter-billion-dollar valuation in five months. Whether the product underneath it can survive contact with customers who expected to actually sell something is a question Polsia's investors are now answering with $30 million.

What to watch: the Series B timeline will tell. If ARR growth re-accelerates with genuine platform economics, the valuation holds. If the deceleration settles into a new plateau, the 25x multiple becomes a narrative liability for every VC betting on AI agent infrastructure — and the next fundraise conversation will look very different from this one.