AI Companion Startup Tolan Raises $30M as Rivals Collapse

AI companion startup Tolan has raised $30 million, roughly doubling its total funding and pushing its valuation past its previous $120 million mark.

Union Square Ventures, Khosla Ventures, and Homebrew participated in the round, according to a new report by Upstarts Media. The startup did not disclose its new valuation. PitchBook estimates it at more than $268 million, with the data provider attaching a "medium confidence" label to that figure.

The round matters because Tolan is one of the few AI companion products with demonstrable traction in a market that has devoured better-funded competitors. Character.ai, once the category's leading name, recently filed for bankruptcy after its founders returned to Google; its remaining staff is transitioning to Disney, which bought some of its assets.

The takeaway: investors are still willing to bet on AI companions, but only on the ones that can show real usage and revenue.

What is Tolan, and who uses it?

Tolan is an AI companion app designed to help people manage their daily lives, navigate relationships and emotions, and accomplish their goals. Users interact with the product through a cartoonish character intended to offer support and help them work through their to-do lists.

The startup claims more than 250,000 active users who have held more than 70 million conversations over the past year, according to its team's statements to Upstarts Media. Around 70% of its users are women, and 40% are parents.

Anecdotally, the app is said to appeal in particular to neurodivergent people, who may benefit from additional support, whether emotional or organizational. That claim is not backed by disclosed data, and Tolan has not published research substantiating it.

The company monetizes through tiered memberships, which are required for full access to all features. Tolan told Upstarts Media its annual recurring revenue is now in the "double-digit millions" — a vague formulation that could mean anything from $10 million to $99 million, and the company did not narrow it.

The takeaway: Tolan has real users and paying subscribers, but its headline metrics come from the company itself and lack independent verification.

Why has Tolan survived where other AI companions failed?

Tolan appears to have found a balance between being helpful and being perceived as too weird, which is more than several of its competitors can say. The contrast with Friend, a pendant-based AI companion, is instructive: users dubbed Friend "very cringe," The New York Times called it "hated," and Wired said it "has no soul."

Public perception has been a recurring execution risk in this category. A companion that reads as intrusive or unsettling loses users quickly, regardless of its underlying model quality. Tolan's cartoonish presentation appears calibrated to avoid that failure mode.

Character.ai's collapse adds a second lesson: talent concentration at large labs can hollow out a startup overnight. When its founders returned to Google, the company lost the leadership that had defined its product, and it never recovered its independence.

The takeaway: distribution and tone, not raw model capability, appear to decide which companion products survive.

What competition does Tolan face next?

The harder threat may come from AI agents rather than companion chatbots. Agents help users plan, organize, and accomplish tasks and goals, not just talk their way through them — a functional overlap that goes straight at Tolan's stated purpose of helping users get things done.

Meta has already moved in this direction. The company gave its AI agent Muse a cute avatar for users to talk to, and it plans to release Muse as a hardware product in the future, according to the reporting. A social platform with billions of users shipping an avatar-based agent would compress the space Tolan currently occupies.

Tolan's defense is its specific audience: women and parents, a demographic mix most AI products have failed to retain. Whether that loyalty survives the arrival of free, bundled alternatives from Meta and others is the open question the $30 million is effectively meant to answer.

The takeaway: Tolan's competition is shifting from failed startups to the largest technology companies on earth.

What about the safety risk?

Tolan also has to navigate the risks of offering AI-powered emotional support. Multiple lawsuits have been filed against general-purpose AI app makers alleging that chatbot interactions contributed to users' suicides.

The legal exposure is not hypothetical for the category, and Tolan's user base skews toward people seeking emotional support, including anecdotally neurodivergent users. The company's disclosed safety measures, moderation practices, and clinical safeguards — if any — were not detailed in the report, which is a notable omission for a product in this space.

A startup generating double-digit-millions in ARR from emotionally dependent users faces a risk profile that conventional software metrics do not capture. No amount of funding insulates a companion company from the liability and reputational consequences of a worst-case incident.

The takeaway: the same engagement that drives Tolan's revenue is what makes its legal and ethical exposure acute.

What happens next?

Tolan now has the capital and the runway to defend a niche position while the broader field consolidates around it. Its reported metrics — 250,000 active users, 70 million conversations, double-digit-millions ARR — put it ahead of most direct peers, but all of those numbers are self-reported.

The company's trajectory will likely be decided by two forces it does not control: whether Meta-class competitors bundle similar avatars into products with far larger distribution, and whether litigation over AI-related harms produces regulatory or liability standards that reshape the emotional-support category entirely.

By James Calloway