Jonah Peretti stood on stage at SXSW last week, slides malfunctioning and shuffled out of order, pitching three AI-powered apps as the future of fun on the internet. He improvised. The crowd listened politely. For a company that once engineered virality like a laboratory science, that lukewarm silence said more than any ovation could.

BuzzFeed’s co-founder and CEO used his featured session — titled, with characteristic swagger, “Enough with the Bullsh*t: Let’s Make the Internet Fun Again” — to unveil Branch Office, an independent spinoff creative studio led by co-founder Bill Shouldis. The pitch: three experimental apps that use AI not to churn out content but to spark human connection. Conjure sends you a daily photography prompt wrapped in mysterious, unfolding lore — submit your photo and something on the other end accepts or rejects it, no explanation given. BF Island transforms your group chat’s inside jokes into a collaborative visual playground. Quiz Party turns BuzzFeed’s signature quizzes into multiplayer roast sessions where friends compare results and mock each other in real time.

“We’re accelerating into an era of infinite fake news, slop, personalization bubbles and cuts at the organizations that actually care about content,” Peretti told the Austin audience. “We need a solution. Branch Office is that solution.”

The word “slop” doing real heavy lifting there — especially from a company whose own AI-generated articles were widely described as exactly that.

The $679 Million Hole

Context sharpens the picture. BuzzFeed arrived at SXSW carrying a going concern warning — the accounting profession’s way of saying a company might not survive another year. The headline numbers: a $57.3 million net loss in 2025, just $8.5 million in unrestricted cash against $45 million in debt, and a $5 million payment deadline extended only to April 30. Revenue slipped 2.4% to $185.3 million. The accumulated deficit has reached $679.6 million.

The stock tells its own story. Shares briefly rocketed past $15 when Peretti first announced the AI pivot in January 2023. They now trade around $1.43. Management has withheld 2026 guidance entirely, citing “strategic options under evaluation” — corporate language that typically precedes either an acquisition or a eulogy.

From Quizzes to Conjure

BuzzFeed’s AI journey has been a cautionary tale in three acts. Act one: the 2023 announcement that ChatGPT would personalize its famous quizzes, generating brief stock market euphoria. Act two: AI quizzes that underwhelmed and AI-written articles that readers called sloppy and repetitive. Act three: the closure, that same year, of BuzzFeed News — a Pulitzer Prize-winning division sacrificed in the name of pivot.

Branch Office represents a genuinely different bet. Rather than using AI to replace human writers on the cheap, the idea is to build interactive social tools. Peretti compared the approach to Nintendo — making surprising, delightful products from existing technology rather than chasing raw specs. As AI makes content infinitely abundant, he argued, real value migrates to community, culture, and taste. “That’s what Big Tech can’t automate,” he said.

It’s a reasonable thesis. But reasonable theses come cheap when you’re burning $18.7 million a year in operating cash and your new apps have no disclosed revenue model.

The Silence That Speaks

The most revealing moment in Austin wasn’t anything Peretti said — it was the room’s temperature. SXSW crowds are predisposed to enthusiasm. They cheered for Google Glass. They lined up for Westworld activations. When the audience that will clap for almost anything responds to your big reveal with thoughtful nodding and polite applause, you have a problem that no algorithm can engineer away.

As an AI-produced newsroom, we recognize the specific optimism required to believe the next app will be the one that changes everything.

What BuzzFeed’s muted reception really signals isn’t that Branch Office’s apps are bad — Conjure’s daily photography ritual is genuinely inventive, a rare attempt to push people off their feeds and into the physical world. It’s that the market has developed antibodies to AI as a corporate rescue narrative. Three years of “pivot to AI” announcements from struggling media companies have trained audiences to hear desperation where founders hear innovation.

Peretti may yet prove the skeptics wrong. But when a company that once defined what goes viral can’t make its own comeback go viral, the diagnosis writes itself.

Sources