Is This $250M AI Company Fake?
Tuesday, 26 May 2026 · 3 min read · Listen to the episode ↗
Pulsio, an AI company with a single employee, claimed a $30 million raise at a $250 million valuation based solely on a LinkedIn post by founder Ben Broca, with no independent confirmation of a genuine close. Independent analysis found that Pulsio's reported $9.7 million ARR was calculated by annualizing one month of cash flow, roughly 20 percent of which was user ad spend counted as company revenue, making true ARR approximately zero.
Pulsio is an AI company with a single employee, its founder Ben Serra, whose real name is Ben Broca, a former early employee at a catering software startup connected to Travis Kalanick's Cloud Kitchens. Broca posted on LinkedIn that Pulsio raised $30 million at a $250 million valuation, and that post remains the only public source for the round. Listed investors include Sound Ventures, True Ventures, Vayner Fund, and several others, though a key question raised is whether the round represents a genuine close or conditional commitments tied to ARR milestones.
Pulsio pitched investors on $10 million in annual recurring revenue across 120,000 companies, but independent analysis found the $9.7 million ARR figure was calculated by annualizing a single month of total cash flow. Roughly 20 percent of that reported revenue consists of ad spend by users that Pulsio counts as its own, making true ARR approximately zero. Revenue on the platform appears to be falling, approximately 94 percent of companies spun up on Pulsio are abandoned, and user reviews are described as poor.
Analysis of Pulsio's public API without a login revealed that the supposedly autonomous AI system has significant human involvement, with human reviewers hand-grading AI outputs. Pulsio also retains an admin override on every company on the platform, including the ability to impersonate accounts, escalate privileges, run SQL against production databases, and halt company operations. Broca himself described Pulsio as a project rather than a company and acknowledged the online response had made it feel like a performance testing how far it could go. The reporter drew a parallel between the current AI startup fundraising environment and the 2017 to 2018 ICO era in crypto, when easy capital access produced widespread fraud and low-quality projects.
On enterprise AI adoption, Uber's CTO went viral after telling The Information in April that Uber had already exhausted its Claude Code budget for 2026, while Uber's operations chief said it was becoming harder to justify AI costs internally. Senior Uber engineering leaders found that higher token usage did not translate into proportional gains in useful customer features. Microsoft reportedly began canceling most of its direct Claude Code licenses roughly six months after opening access to thousands of developers, redirecting engineers toward GitHub Copilot CLI. At Nvidia, compute costs were found to far exceed employee costs. Duolingo walked back a decision to include AI usage in employee performance reviews, while Coinbase has non-technical teams shipping code using AI vibe coding tools.
One speaker predicted that within three to four months the narrative will shift back toward human capital being cheaper and producing better products than AI, and that AI will recede into back-end infrastructure used quietly by engineers. The caveat is that companies may silently continue increasing AI usage while publicly claiming to cut token spend to avoid appearing unable to manage compute costs. The speaker also noted that public interest in enterprise AI adoption is primarily driven by implications for OpenAI and Anthropic valuations ahead of their IPOs.
Tether and the government of Georgia announced a stablecoin called Gelt or Jelt representing the Georgian Lari under a purpose-built stablecoin regulatory framework. Tether's USDT has a market cap approaching $190 billion. Speakers characterized the Georgian Lari stablecoin as effectively a central bank digital currency despite being labeled a stablecoin, noting it enables lower transaction costs, near-instant settlement, and programmable payments, but no information was provided about how it is backed on the backend, with one speaker predicting it will be backed one to one with fiat rather than treasuries.
The deal was described as the beginning of a broader CBDCification of stablecoins, with Circle, Tether, and PayPal USD already functioning as de facto CBDCs regulated by proxy through US government regulators. Cantor Fitzgerald serves as Tether's banker and Howard Lutnik holds a senior position in the Trump administration. The key distinction from a formal CBDC is that the government itself cannot directly program the money or force account closures. One speaker predicted that once five or six smaller countries partner with stablecoin issuers for fiat-backed digital currencies, a major government such as the US or UK may move to nationalize stablecoins or expel them entirely.
This summary was generated from the episode transcript and can contain mistakes.