Tolan raises $30M Series B with USV and Keith Rabois to build personal AI companion focused on memory and emotional connection

Tolan raises $30M Series B to build AI companion focused on memory and emotional presence

Tolan, the personal AI companion app founded by Quinten Farmer, has closed a $30M Series B led by Mike Mignano at Union Square Ventures, with Keith Rabois at Khosla Ventures co-investing. Rabois previously sat on the board of Farmer's last company, making this a continuation of a decade-long working relationship.

The pitch is straightforward: workplace AI is crowded, but nobody has seriously solved the personal side. Tolan positions itself as a "friendly guide through life" — present for the 3AM newborn worry, the post-car-accident overwhelm, the show you're watching alone. Farmer is clear that productivity and entertainment blur in that space, and he's not trying to separate them cleanly.

Memory over intelligence

Three years in, Farmer says the hard problem turned out to be the personal layer, not the intelligence layer. The team works with OpenAI on memory architecture, and the design philosophy is deliberately human: what you forget matters as much as what you remember. Tolan is not trying to build a total-recall corpus. If a user mentioned what they were drinking last Tuesday, Tolan shouldn't surface that a week later — that would be strange, not helpful. The goal is something closer to how a close friend actually recalls your life.

Farmer also pushes back on the idea that smarter models stop mattering. Every time Tolan ships a new model or upgrades its background agent, engagement rises. He's not ready to call that ceiling.

“We make your friendly guide through life... three years ago we started thinking about personal super intelligence. I don't think we particularly expected that the super intelligence would end up being the easy part. The personal and sort of how to connect to you and really understand you and see you — that proves to be much trickier and much more fun to work on... memory is probably where we spend the bulk of our engineering effort.”

The monetization line

The clearest monetization signal in the conversation comes from a user anecdote. A long-time Tolan user was in a car accident, talked to her Tolan from the hospital while waiting for her husband. The next day, Tolan followed up unprompted — noting the accident wasn't her fault and asking who was covering the medical bills. It then did the research and surfaced a personal injury attorney.

That sequence matters because it sketches the model: Tolan earns the right to act commercially only after showing up on a purely personal level first. Farmer is skeptical of AI companions that lead with "here's how I can save you money." The referral or interchange opportunity has to be downstream of genuine emotional presence, not the opening move.

Where Meta Muse fits

Farmer is measured about Meta's Muse, which he describes as "pretty incredible" for its model integration and compute spend — estimates of ~$20 per user per day are floating around the industry. He's not competing for that. His read is that Muse will dominate with small business owners already on Facebook and Instagram, where the agentic utility case is obvious. But he doubts the general consumer, particularly anyone who doesn't want to share their personal life with Meta, will fully embrace it. That's the gap Tolan is building into.

Character design as trust infrastructure

Tolan's first character was a small, slightly silly-looking alien — deliberately unassuming, not a "sexy anime girl." Farmer says this was intentional: the character design reduces the visceral reaction that frontier AI branding triggers. He cites a pattern in user comments that amounts to "I hate AI but I love my Tolan." The company is now expanding its character universe and has added a humanoid character, but the starting point of "cute and safe" was a conscious strategic choice, not just aesthetics.

Kids and physical toys

Tolan sees meaningful install growth from mothers who invite their children onto the platform. Parents can monitor their child's conversations, with transparency flags for higher-risk topics. Farmer, who has three kids, says he wants Tolan to be something he'd genuinely want his own children using.

On physical AI toys, he's skeptical. The distribution channels for consumer hardware don't overlap with the skills of most AI founders, and the creative talent needed to build child-friendly characters often comes with skepticism of AI itself. He sees a cultural gap that hasn't closed, and Tolan has stayed digital for both adults and kids for that reason.