Airbnb's Chesky: AI Writes 60% of the Code, and 'Founder Mode' Is How You Win

Airbnb CEO Brian Chesky has a message for corporate America as AI reshapes how software gets built: the winners will be the companies that refuse to grow up. In an interview published by Fortune on August 14, 2026, Chesky said AI now writes 60% of Airbnb’s new code — and argued that sustaining “founder mode,” not “manager mode,” is what separates the companies that will thrive in the age of AI from those that will be left behind.

“The winners of AI are going to be the companies that embrace transformation,” Chesky told CNBC last week. “You might call it founder mode, not manager mode.”

The numbers behind the claim

The 60% figure is not new — Chesky first disclosed it publicly at an AI summit in San Francisco in May 2026, and repeated it on Airbnb’s earnings call — but what’s notable is how quickly it has gone from a headline statistic to a baseline operating assumption. Airbnb’s engineers, and even its managers, now routinely work alongside AI coding tools (Chesky has previously acknowledged that managers use Anthropic’s Claude Code). Earlier this year, Chesky told investors that AI had helped the company ship improvements faster and cut the time from concept to launch by as much as 60%.

The results are showing up in the business. On August 6, Airbnb reported a second-quarter earnings beat with $3.6 billion in revenue. During the earnings call, Chesky said Airbnb plans to spend even more on AI tokens this year as it pushes to become an “AI-native company.” He also highlighted the company’s AI-powered customer service tool, which now resolves 45% of user inquiries without any human intervention, and confirmed plans to test AI-powered search.

“AI is accelerating this work across search, sign-up, checkout, and payments,” Chesky said on the call. “By reducing friction across the guest journey, we are converting more traffic into bookings, and that’s become one of the biggest drivers of our growth.”

What “founder mode” actually means here

“Founder mode” entered the Silicon Valley lexicon in 2024, when Y Combinator cofounder Paul Graham published an essay — inspired by a talk Chesky gave at a YC event — that dismantled conventional wisdom about delegating and empowering managers. The argument: great founders stay in the details, and the professional-manager class that replaces them often destroys what made the company special.

Chesky has spent the two years since turning that essay into a working management doctrine. Despite running a company of roughly 8,200 employees with a market value around $110 billion, he still monitors progress on AI initiatives at a granular level — including token usage and the overall output of the AI systems. That level of attention to detail, he argues, is precisely what lets a Fortune 500 giant keep moving like a startup.

“We should start in the details, develop trust, develop muscle memory, and then let go,” Chesky told Fortune’s Ruth Umoh. “Great leadership is presence, not absence.”

He dismisses the alternative — running companies as autonomous structures that operate without the founder’s involvement — as “such a fallacy in modern corporate America.” Companies, in his framing, are organizations in the literal sense: people organized to work together, and that coordination doesn’t happen on its own.

Chesky is in notable company. According to Fortune’s calculations, 26 of the 2026 Fortune 500 companies are founder-led, including Dell Technologies, BlackRock, Meta, and Nvidia.

The next bet: consumer AI and a new lab

Chesky’s bullishness extends well beyond internal engineering efficiency. He is launching his own AI lab — details remain scarce — and he believes consumer AI is on the verge of a breakthrough, even though he’s blunt about its current limits.

“There’s been very little progress on consumer AI,” Chesky said. “I think the current construction of agents has been pretty limited.” His prediction: “You’re going to see a renaissance around consumer AI that is going to begin to change daily life.”

On the product side, Airbnb is reportedly working to transform itself into a one-stop shop for travel and vacationing — a move sources tell Bloomberg could add as much as $1 billion in revenue. Chesky predicts it could take up to 18 months for the new AI software to debut, calling the effort a “very, very tall mountain.” A year from now, he expects Airbnb’s experience to be “very AI-native,” and he assumes the rest of the industry “is all going to have to follow.”

“I think we’re in the beginning of a thousand-year journey on AI,” Chesky said in a recent Yahoo Finance interview. “I think we’ve barely scratched the surface of what’s possible with AI.”

Why it matters

The Airbnb story is becoming the template for how large, non-AI-native companies are absorbing the technology. The pattern is consistent: AI coding tools compress development cycles, which lets small teams ship like startups; AI customer service handles a meaningful share of interactions at lower cost; and the savings get reinvested into more ambitious AI-native products.

Two things make Chesky’s version distinctive. First, the degree of personal involvement — a CEO tracking token usage is not normal, and it signals how seriously he treats AI spend as a first-class operational metric rather than an IT line item. Second, the “founder mode” framing gives other CEOs a vocabulary for justifying deep engagement at a moment when AI is moving too fast for traditional management cadence.

The open question is whether founder intensity scales. Airbnb’s 45% AI-resolved customer service and 60% AI-written code are impressive, but the harder test is the one Chesky set for himself: shipping a genuinely AI-native consumer product within 18 months in a category — travel — where consumers have been slow to trust agents with money and plans. If he pulls it off, “founder mode” will look less like a management philosophy and more like a survival strategy.