Character.AI Files for Chapter 11 Bankruptcy: Who Made Money on Companions

⚡ The 5-Second Rundown

• On October 12, Character.AI filed for Chapter 11 protection: assets are being sold under court supervision, and no buyer has been found yet.
• At its peak — 22 million MAU in August 2024 and over 90 minutes per user per day. The company still ended up in bankruptcy.
• While the leader sinks, mobile revenue in the AI-companion market grew from 316 million yuan in January 2026 to 656 million in August.

🔍 What Happened

On October 12, Character.AI filed for Chapter 11 bankruptcy in the US. This is not liquidation: the company expects to sell off its assets under court supervision. If no buyer is found, the app may simply be shut down. 36Kr reported this in a piece dated November 12, 2026.

Not long ago, the numbers looked like a textbook case: in March 2023 — $150 million from a16z and a $1 billion valuation less than two years after founding. The mobile app was downloaded 1.7 million times in its first week. In September 2023, according to Similarweb, it had 4.2 million MAU on phones in the US — almost on par with mobile ChatGPT's ~6 million. In August 2024, Sensor Tower counted a record 22 million MAU and first place in the AI-companion category. Users spent on average more than 90 minutes a day and started around 25 conversations — TikTok, with its ~70 minutes, nervously smokes on the sidelines.

Now for the contrast. According to Sensor Tower, mobile revenue for the entire AI-companion category jumped from 316 million yuan in January 2026 to 656 million in August. South Korea's Zeta pulled in 70.9 billion won (about 340 million yuan) in the first half of the year and is already profitable. In October, the CHAI app reported an annual revenue run rate of $120 million — with a team of 10–20 people. Tipsy, PolyBuzz, and Crushie are climbing the global revenue charts.

🧠 Our Take

This looks like a classic engagement trap. Shazeer, Character.AI's founder, framed the bet beautifully: the global entertainment market is $2 trillion a year, and people get attached to fictional characters. The logic works — only TAM doesn't pay the bills. 90 minutes a day is a great line for a pitch deck, but investors need money, not session depth. Our assessment: the company built a mass-market product with free access and hoped to monetize later.

Latecomers went the opposite way. Zeta, CHAI, and others aren't chasing hundreds of millions of users — they sell subscriptions and paid messages right away to a small but paying audience. A 20-person team with a $120 million annual run rate is no longer a startup — it's a money printer. We believe the AI-companion market has split into two models: "audience first, money later" and "money now." The first produced a bankruptcy, the second — profit.

The winners are small teams with a narrow niche and a short path to a paying customer. The losers are those who burned venture money on user acquisition and kept access free for the sake of metrics. For users, this is bad news: the characters they've grown attached to live only as long as someone pays for the servers.

⚠️ What We Don't Know

The article omits the key facts for a bankruptcy story: the debt amount, the list of creditors, the names of potential buyers. The dates should also be double-checked — the piece is dated November 12, 2026, while the filing was made on October 12. There is a single source, with no links to court records. A Chapter 11 filing is a public document and can be verified via PACER; until then, the figures should be treated as a secondhand account.

Prediction: within a quarter, Character.AI's assets will get a public bidder, or the court will convert the case to liquidation. I'm betting on the former: 22 million MAU and the brand are too valuable an asset to simply switch off.

Disclaimer: not officially confirmed, single source — no court documents cited in the text.

Source: 36Kr