In a Rapid Response interview recorded live at the Masters of Scale Summit 2026, Replit chief executive Amjad Masad traces his company’s jump from a $3 billion valuation six months ago to $9 billion today, arguing that “AI psychosis” is distracting Silicon Valley from a cybersecurity crisis it keeps ignoring.
Fast Company published the abridged transcript this week, and the contrast at its center is stark: two years ago, the company was flatlining near $2 or $3 million in annual revenue.
The framing matters because Masad is not painting a smooth ascent, Fast Company reported. He describes a layoff, a mostly empty office in Foster City, and years of investors and recruits declining to believe his core thesis — that anyone with an idea should be able to build software. The turnaround, in his telling, only arrived when the underlying technology finally caught up with the promise.
That thesis dates back further than the current AI boom suggests. He says he began promoting universal software creation in 2010, building an open-source precursor to Replit, then joined Codecademy as a founding engineer before starting the company in 2016. The mission then, as he describes it to host Robert Safian, a former Fast Company editor-in-chief, was to blur the line between learning to code and doing code.
“We want to blur this distinction between learning to code and doing code, and build an environment that really makes it so that anyone with an idea can build software,” Masad said, recalling the 2016 founding pitch in the interview.

A layoff, then vibe coding
The low point, per the transcript, came just before the breakthrough. Replit was six or seven years into its life, burning heavily, and still stuck at $2 or $3 million in annual recurring revenue, Masad said. The company had upgraded from a small San Francisco office to a massive one in Foster City, chosen because the space was cheap — and it sat largely unoccupied after a layoff.
“Every day, I would go into the office and sit at my desk and wait for another person to quit,” he told Fast Company’s Rapid Response.
He also described the deeper cost of cutting staff. “The thing about doing a layoff is you also burst a bubble. You burst a dream,” Masad said, adding that every founder needs some version of a “reality distortion” to sell a vision — and that the dream dissipates the moment a founder says the company is running out of money.
The inflection came in 2024. Masad claims Replit invented what would later be called vibe coding that year — “the first fully natural-language-based programming environment,” in his words, at a time when nothing comparable existed.
One early user, he recounts, posted a video crying on YouTube and Twitter after building a personal app that overlaid their photos and memories on a map — an idea they had chased across multiple technologies for years without success. Fast Company reports that stories like this one drove early adoption among people who had never shipped software before.
Replit valuation triples to $9 billion
The valuation math puts the current moment in context. Replit now sits at $9 billion, triple its $3 billion mark from just six months ago, per the interview — richly valued even by the standards of the current AI-funding cycle, and a number that assumes the vibe-coding category keeps expanding rather than consolidating around larger platforms.
It also leaves Masad openly uneasy. He notes he has angel-invested long enough to watch early AI companies ride hypergrowth and hyper-deceleration within a single year, and says a “fair amount of paranoia” is built into how Replit operates. Vindication and terror, he tells Safian, coexist.
Masad says some of his AI engineers joined specifically saying they wanted to automate their own jobs. The broader bind he describes applies to every knowledge worker — the work itself generates the training data that may eventually replace the worker.
What ‘AI psychosis’ means — and the risk Masad says counts
The sharpest elbows in the interview are reserved for the AI-safety discourse. Masad has publicly criticized Dario Amodei specifically at Anthropic for overheated predictions about AI-fed job losses, Fast Company notes, and he frames much of the extinction-and-machine-consciousness conversation as what he calls “AI psychosis.”
“When they go out and talk about how AI is conscious, and they go meet the Pope and do all this random stuff, it’s really distracting, and it’s also, in many ways, destructive and bad for humanity for us to care about the well-being of these machines when we haven’t really figured out the well-being of humans,” Masad said.
His alternative priority is concrete: cybersecurity. He points to major platform breaches in the two or three days before the interview was recorded, says Replit’s team worked overnight on mitigation, and calls the current moment “a very serious moment for the world.” He also points to an earlier warning sign.
The turnaround, in sequence
Map the arc against Masad’s own account and the shape of the risk becomes visible:
- 2010 — Masad begins advocating for universal software creation while building the open-source precursor to Replit.
- 2012 — He immigrates to the United States, describing years of grinding to find his footing before the company found its market.
- 2016 — Replit is founded on the learning-to-code-meets-doing-code thesis; he says few believed it.
- 2024 — After a layoff and years flatlining near $2–3 million ARR, the company ships the first fully natural-language programming environment, later named vibe coding.
- 2026 — Valuation reaches $9 billion, up from $3 billion six months prior; Masad warns the cybersecurity threat is a “disaster brewing.”
The interview leaves the hard question — whether hypergrowth at this pace can be sustained — unanswered, and Masad does not pretend otherwise. He told Fast Company it is “wild” scaling from a few million to a billion dollars in such a short span, and conceded he mistimed the market once before, hiring ahead of a future that “hadn’t arrived as fast as I thought it would.”
His closing frame, on the risk profile: the extinction debate is a distraction, and the real work is defensive.
“The talk about AI extinction is just not relevant right now. We need to focus on cybersecurity,” Masad said.