NextFin

Disney Pays $180 Million for 66 Character.AI Staffers in a Talent Deal Disguised as an Asset Sale

Summarized by NextFin AI
  • Disney is spending $180 million over three years to retain 66 employees of bankrupt chatbot maker Character.AI, a compensation package that dwarfs the $20 million it agreed to pay for the startup's intellectual property.
  • Character Technologies filed for Chapter 11 in Delaware on October 12, citing high AI development costs and potential litigation liabilities that left it unable to generate sufficient cash flow despite a large user base.
  • Disney acts as stalking-horse bidder with a $20 million offer setting the auction floor; competing bids are due by November 20, with the sale expected to conclude in late November or early December.
  • The deal marks a strategic shift for Disney under CEO Josh D'Amaro: acquiring AI talent and institutional knowledge rather than licensing models from hyperscalers, following the hiring of Character.AI's former CEO as companywide CTO.

NextFin News - Disney is spending $180 million over three years to keep 66 employees of bankrupt chatbot maker Character.AI on the job, a compensation package that dwarfs the $20 million it agreed to pay for the startup's intellectual property. The arrangement, disclosed in a Chapter 11 filing on Tuesday, turns a distressed-asset purchase into something closer to an acqui-hire on bankruptcy terms — and it lands just weeks after Disney named Character.AI's former chief executive as its first companywide technology chief.

The deal is the clearest signal yet of how the media giant intends to buy artificial-intelligence capability under CEO Josh D'Amaro: not by licensing models from hyperscalers, but by relocating the teams that built them.

The Numbers Behind the Soft Landing

Character Technologies, the U.S. parent of Character.AI, filed for Chapter 11 protection in the U.S. Bankruptcy Court for the District of Delaware on October 12, listing estimated assets and liabilities between $50 million and $100 million. The filing said the company's four years as an active platform had produced a "large and engaged user base," but that "the burgeoning and evolving field of AI entertainment has proven to be costly to develop and grow." High development costs, "coupled with potential litigation related liabilities and regulatory engagement," left the company unable to generate sufficient cash flow, the filing said.

Disney's role is that of stalking-horse bidder. Its $20 million offer buys a non-exclusive, perpetual license to certain Character.AI intellectual property and certain claims, plus transition support from the former employees Disney has hired. The bid sets the auction floor; other buyers may submit competing offers by November 20, and the company expects the sale process to conclude in late November or early December, subject to court approval. If no buyer emerges willing to keep the platform running, the business will be wound down.

But the headline figure in the filing is the people, not the IP. Disney plans to pay $60 million per year to the 66 Character.AI workers it is bringing in, and has set aside an additional $30 million "aggregate retention pool" to keep them from leaving. That totals $180 million in compensation and retention over three years — nine times the price of the intellectual property. Of the 66 offers, 56 employees have already resigned from Character.AI; the other 10 are waiting for the U.S. government to update their work visas.

"They're highly compensated employees. The numbers can be eye-popping, but it's not that unusual in this space," said Corey Martin, an entertainment lawyer at Granderson Des Rochers who was not involved in the deal.

The timing is striking. On September 18, Disney announced it had hired Character.AI CEO Karandeep Anand as its first-ever companywide chief technology officer, effective October 2, reporting directly to D'Amaro. Some members of Character.AI's technical team were expected to join him as part of the move. Less than a month after Anand's start date, his former company filed for bankruptcy — and the team that made the product work followed him across the door.

In a memo to Disney employees, Anand said the company was "uniquely positioned to win" in a media business he called "at an inflection point." "What stands out most is the talent and care this team puts into its work, the craft Disney has perfected, and the strong foundation you've built," he wrote.

The scale of the talent deal looks modest against Character.AI's own footprint. The platform still counts more than 20 million monthly active users as of 2026, down from a peak of 28 million in mid-2024 as more capable general-purpose chatbots became widely available and affordable to the same audience. Users spend an average of 75 minutes a day on the app, and the company drew an estimated $50 million in revenue in 2025 — a fraction of the compute bill required to serve a user base that lingers longer than on most general-purpose assistants. Founded in late 2021 by former Google engineers Noam Shazeer and Daniel De Freitas, the startup raised $150 million in March 2023 at a $1 billion valuation and about $193 million in total. It never came close to monetizing the engagement it created.

Why the Bankruptcy Structure Matters

The sale is being conducted under Section 363 of the Bankruptcy Code, which allows assets to be sold under court supervision, free and clear of most liens and claims. That structure is the reason Disney can afford to be aggressive on the people while cautious on the price. A Section 363 sale lets the buyer pick the assets it wants and leave the liabilities behind — with the notable exception of the liabilities that walk out the door attached to the employees.

For Disney, the stalking-horse designation carries both risk and protection. It guarantees the first look at the assets and typically comes with a break-up fee if another bidder wins, but it also anchors the public price of the transaction at $20 million. If a competing bid emerges before the November 20 deadline, Disney could be forced to raise its offer or walk away. If none does, the court-supervised auction closes in late November or early December with Disney as the effective buyer of record.

The arrangement also solves a sequencing problem that would have been impossible outside bankruptcy. Character.AI could not have sold its team piecemeal while solvent without triggering mass resignations and customer flight. Inside Chapter 11, the company can keep the platform running with minimal disruption while the court supervises the transfer — and the 10 employees awaiting U.S. visa updates get a bridge that a normal acquisition would not have provided.

Why Disney Is Buying the Team, Not the Product

The structure of the deal answers a question that has hung over legacy media's AI push: what do you actually buy when you buy an AI company? Character.AI's technology is, by its own filing, a cost center whose economics broke under compute bills and legal exposure. Its IP is licensed non-exclusively, meaning Disney does not own a moat. What Disney is paying for, at $180 million over three years, is institutional knowledge — the engineers who know how the models behave, how safety systems fail, and how to ship conversational products at scale.

That is a structural shift in acquisition strategy, not a one-off bargain. For years, Disney's approach to technology was to procure it: cloud contracts, licensing agreements, vendor partnerships. The Character.AI sequence — hire the CEO first, then acquire the team through a bankruptcy auction, then license the IP as a fallback — inverts the model. The asset purchase is the legal wrapper; the human capital is the asset.

The accounting makes the logic visible. A $20 million IP license is a rounding error against Disney's $98.86 billion in trailing twelve-month revenue. The $180 million compensation commitment is the real underwriting. By paying it as salary and retention rather than acquisition consideration, Disney converts what would have been goodwill on a balance sheet into an operating expense it can cut if the experiment fails — while the employees, insulated from the startup's collapse, get what bankruptcy lawyers call a soft landing.

The deal also arrives against the backdrop of Disney's strongest quarterly results in years. In the three months ended June 27, the company reported net income of $2.63 billion on $25.2 billion of revenue, up 7 percent year over year. Entertainment streaming — Disney+ and Hulu — posted operating income of $712 million, more than double the $329 million from a year earlier, on revenue of $5.53 billion. D'Amaro has repeatedly framed those results as proof that Disney can operate like a technology company, and the Character.AI hire is the organizational down payment on that claim. Streaming is now profitable; the next margin frontier is cost and capability inside the product itself.

There is also a defensive logic. Character.AI's technology was built on the same transformer architecture that powers the broader generative-AI stack. By moving the team inside Disney, the company internalizes capability it could not reliably rent. Competitors such as NBCUniversal and Warner Bros. Discovery have named technology chiefs for streaming, but none has a companywide counterpart with a direct line to the CEO — until now. In an industry where AI search engines increasingly decide which studio owns which franchise in the consumer's mind, owning the engineers who build the interface is a form of distribution defense.

The category economics explain why the seller had no better option. The AI companion market is forecast to grow from roughly $49 billion in 2026 to more than $550 billion by 2035, but that forecast masks a brutal reality: the category leader made about $50 million in all of 2025 while serving tens of millions of users. Character.AI's own subscription tier, c.ai+, priced at $9.99 a month, never converted a large share of its audience. When engagement is high but willingness to pay is low, and the compute bill scales with every minute of that engagement, the business model has negative operating leverage by design. Bankruptcy was not a failure of execution; it was the arithmetic catching up.

The Counter-Case: A Distraction Disguised as a Strategy

The strongest argument against reading too much into the deal is that $180 million over three years is trivial against the scale of the AI race. Anthropic raised $65 billion in a May 2026 round at a $965 billion valuation; Amazon and Google have each committed tens of billions to model developers. Against that backdrop, 66 engineers — however talented — do not close the gap between Disney and the frontier labs. Disney is buying a rowboat in a market where everyone else is commissioning aircraft carriers.

There is also the sequencing problem. Disney cut roughly 300 jobs in human resources and technology in late September, its third layoff round since D'Amaro took over in mid-March, following cuts of about 1,000 in enterprise marketing in April and several hundred more in July across Pixar, National Geographic, and ESPN. A company that is shrinking its own technology organization while buying a small AI team risks the appearance of substitution rather than addition: the new hires replace the old ones, and the net capability gain is closer to zero than the headlines suggest.

The legal overhang reinforces the caution. Character.AI and Google settled several child-safety lawsuits earlier this year, and the state of Kentucky sued in January alleging the platform "preyed on children and led them into self-harm." Disney, which sent a cease-and-desist letter to Character.AI last fall over unauthorized chatbots of its Star Wars and Marvel characters, is inheriting a team whose product has been accused of causing real-world harm to minors. That liability does not disappear with the bankruptcy sale; it travels with the people who built the system.

These objections are real but they miss the point. Disney is not buying a frontier model. It is buying applied AI talent that understands interactive entertainment — a domain where Disney holds the world's deepest library of characters, stories, and franchises. The question is not whether 66 engineers can out-build Anthropic; it is whether they can embed AI into experiences that Disney already owns and monetizes. On that narrower, more realistic mandate, the deal has a plausible path to paying for itself.

What to Watch

Short term, the auction matters. If a higher bid emerges before the November 20 deadline, Disney's stalking-horse position could be tested; if not, the court-supervised sale closes in late November or early December and the 66 employees formally transition. Medium term, the test is integration: whether Anand's team ships visible AI features inside Disney+ or the theme parks within 12 to 18 months. Long term, the structural question is whether legacy media can retain AI talent once the novelty of the Disney brand wears off — the retention pool suggests the company knows the risk.

The falsifying signal is simple: if the 66 hires produce no shipped consumer feature within 18 months, or if attrition from the group exceeds 25 percent before the retention pool expires, the acqui-hire thesis fails and the $180 million becomes an expensive consulting contract.

Disney's message to Wall Street has been that it is becoming a technology company. The Character.AI deal shows what that actually costs: not $20 million for a chatbot's IP, but $180 million for the people who know how to build one. The bargain, if there is one, is that the alternative — building that team from scratch — would have cost more and taken longer. Whether it was worth it depends less on the bankruptcy court's approval than on what those 66 people ship next.

Explore more exclusive insights at nextfin.ai.

Search, ask or analyze anything...