You remember how it felt, right? The first time you saw that filter turn your face into a dancing chicken or overlay a flower crown on your best friend's head? Snap made AR feel like magic. But somewhere between the Spectacles launch and the latest redesign, the spark got lost in translation. And while Snap's trying to rekindle that thrill, it's also learned something awkward and quietly important: some ideas just don't belong inside the same building as a messaging app.
That's why Dotmo exists — not because Snap failed at AI, but because it finally admitted success looked nothing like what Snapchat's roadmap needed. The company announced this week that its internal generative AI video team is spinning off into an independent startup called Dotmo, focused on interactive gaming experiences powered by that same video tech. It's a clean break, not because the team wasn't good, but because Snap's mission and AI's future are pulling in different directions.
Here's what I love about this move: Snap didn't bury the lede. They didn't quietly fold the team into Snap's GenAI research unit, claim a win, and hope nobody noticed. No — they acknowledged the mismatch head-on: AI video needs patient capital, technical autonomy, and a product environment that rewards experimentation over quarterly ad impressions. And Snap? That's not it anymore.
The Dotmo Deal Is Simple, But Clever
The structure tells its own story. Bobby Murphy — yes, the co-founder and current CTO of Snap — is stepping in as lead investor with a personal stake. He'll keep his day job at Snap, continuing to run the company's GenAI research, but he'll also hold a meaningful ownership position in Dotmo. That dual role is not just convenient, it's strategic: Murphy has built his credibility on betting on the right tech at the right time. Investing personally signals belief without overcommitting company capital.
Dotmo itself will be staffed by current Snap staff leaving to join the new venture, which means the core team isn't being spun out — it's moving out intact. Snap will license its existing technology to Dotmo so the new company can adapt those models for gaming and interactive entertainment platforms, then retain a large equity stake. If Dotmo takes off, Snap gets upside without carrying the R&D overhead. If it flops, Snap only loses a minority investment, not years of development time.
And yes — cost is absolutely part of the equation. Snap explicitly cited the high price tag of running a generative AI team internally, and anyone who's followed tech over the last five years knows AI expenses have exploded. But cost alone doesn't explain why this wasn't done earlier, or why Snap went down the spinoff path rather than just cutting the budget. The real reason is time: AI development cycles and social media product cycles are fundamentally misaligned. One needs patience, the other demands velocity.
Why Interactive Gaming Makes Sense for Dotmo
Let's be honest — interactive gaming is an odd place to land if you're coming out of Snap's AR team. Snapchat is about connection, quick messages, and fleeting stories. Gaming is about deep engagement, long sessions, and investment in virtual worlds. The overlap isn't obvious at first glance.
But that's exactly why Dotmo gets it right. The generative AI video models Snap was building were designed for real-time, on-device processing — they needed to run fast and look good on phones. That's exactly the stack Dotmo needs for interactive gaming: responsive, visually compelling, and adaptable across different formats.
Think about it: imagine an AR experience where your real-world environment transforms based on the game you're playing, where characters respond to your voice or gestures, where your avatar adapts in real time to your movements. That's not just a cool gimmick — it's what happens when you take Snap's video tech and apply it to gaming logic instead of social filters. Dotmo isn't trying to replace Call of Duty; it's building the kind of playful, camera-driven gameplay that feels like a natural evolution of Snap's legacy.
This isn't uncharted territory, either. Snap has already dabbled in gaming through Snapchat's AR games and minigames. But those were limited by Snap's product priorities — they needed to support engagement, not immersion. Dotmo can go all-in on depth without worrying about how it affects the main app's load times or notification patterns.
The Specs Precedent Was a Warning Shot
Dotmo isn't Snap's first pivot this year. Earlier in 2026, the company spun off Specs into a separate entity focused on smart glasses development. That move didn't go well — the new hardware launched at $2,200 and triggered a stock drop as investors questioned whether anyone would pay that much for a device most people see only as an accessory, not a must-have.
So why try spinoffs again so soon? Because this time, Snap learned its lesson. Specs was hardware: heavy, capital-intensive, and unforgiving when mispriced. Dotmo is software — lighter, faster to iterate, and easier to pivot based on user feedback.
The real difference is in the positioning. Specs was sold as a hardware play, and the market punished it for failing to meet that framing. Dotmo is positioned as an AI company developing gaming experiences, which is a much stronger narrative in today's market. Investors understand AI — they don't entirely trust it, but they know how to value the potential.
Plus, Snap handled Specs differently. The company kept 100% ownership and direct control over the hardware rollout, which left it bearing all the risk. With Dotmo, Snap is taking a minority stake while letting the new company stand on its own. It's a shift from operator to investor — a much cleaner separation for both parties.
Bobby Murphy's Role Is the Key Detail
If you read between the lines of the press release, Bobby Murphy's involvement is the headline. He's not just backing Dotmo in an advisory capacity or lending his name to a fundraising effort — he's stepping in as lead investor with a personal stake while staying on as Snap's CTO.
That's unusual. Most founders who invest in spinouts do it after stepping down from executive roles, or they use investment firms as intermediaries. Murphy is going full direct: he's putting his money where his judgment is, and he's doing it while holding onto his seat at the table.
This tells me three things. First, he genuinely believes Dotmo has a shot — not just as a tech play, but as a business. Second, he's signaling confidence to the rest of Snap's leadership: if Bobby bets on this, maybe it isn't as risky as it looks. And third, he's creating a self-reinforcing loop: the more Dotmo succeeds, Snap benefits from the equity stake and the licensing revenue, which makes Snap a stronger company overall.
It's also worth noting that Murphy co-founded Snap with Evan Spiegel. He knows how the company operates from the inside out, and he's been through enough down cycles to understand when to cut loose and when to double down. His decision to invest personally — not through a fund, not through Snap's treasury — is the strongest signal yet that he sees real value here.
This Is Part of a Bigger Industry Pivot
What's happening with Snap and Dotmo isn't an isolated move — it's part of a quiet but profound shift across the tech industry. Early in the AI boom, companies rushed to build everything in-house. Every major platform had an "AI research lab," often staffed by star researchers hired away from universities. They built models, fine-tuned them, deployed them in products… and then struggled to explain what value they actually delivered.
The euphoria has faded. Companies are now doing the hard work of separating AI projects that align with their core strengths from those that don't. If you're Meta, your strength is social graph and ad infrastructure — so your AI team focuses on recommendation engines and content moderation. If you're Microsoft, your strength is enterprise software — so your AI team focuses on Copilot and developer tools.
Snap doesn't have a clear area of strength for generative AI video. It's not a gaming company like Unity or Roblox, it's not an enterprise platform like Adobe or Autodesk. Dotmo gives Snap a way to stay in the game without pretending to be something it's not.
This pattern is going to spread. Expect more spinouts, more licensing deals, and more creative structuring as companies finally stop trying to be everything to everyone in the AI era. The winners won't be the ones who built the most AI — they'll be the ones who knew exactly what to keep and what to let go.
Why This Isn't a Retreat, It's a Refinement
Some people will see this as Snap retreating from AI. I see it as the company finally getting serious about what it's good at.
Snap built its reputation on playful, inventive filters that made AR feel magical. But as the company scaled, those features became part of a larger product matrix focused on engagement metrics and ad load. The tension was always there — the more successful Snap became, the harder it was to justify spending months refining a single filter when the board wanted quarterly growth.
Dotmo changes that calculus. Now, Snap can license technology it helped develop, earn equity in a company running with the same vision it once had, and still focus its internal teams on what it does best: social connection, ad platform optimization, and hardware that enhances — not distracts from — the core experience.
The beauty of this structure is its asymmetry. Snap doesn't need Dotmo to succeed spectacularly for this move to pay off. Even modest success gives Snap a financial return, technological insights, and a way to retain key talent who might otherwise leave for a company with clearer AI priorities. If Dotmo becomes a standout success, Snap gets to ride the wave without building the boat.
That's not compromise — it's precision. It's knowing when to hold, when to fold, and when to pass the chip to someone else at your own table.
The Real Test Is Still Ahead
None of this works if Dotmo can't find its footing. The gaming AI space is crowded, to say the least — Unity and Unreal Engine have their own AI initiatives, startups like Kaiber and Runway ML are already pushing boundaries, and big tech players are experimenting with their own approaches. Dotmo will need to find a defensible niche quickly, or risk getting swept up in the noise.
The team's origin is a double-edged sword. Coming from Snap gives Dotmo access to proven technology and a built-in credibility boost, but it also means the team has to shake off the perception that they're just Snap's former lab run out to play in the world. They'll need to develop their own identity fast.
I'm hopeful, though — not because I think Dotmo is guaranteed to succeed, but because the model is right. Companies that try to build everything internally usually fail at one thing or another. Dotmo gives Snap the best of both worlds: control where it matters, and distance where it's needed.
What Comes Next for Snap and AI
If Dotmo takes off, Snap's next move will be interesting to watch. Will the company double down on its AI licensing strategy? Will it spin off other teams that don't quite fit — maybe the AR team itself, if it decides to refocus on core social features? Or will this be a one-off experiment that doesn't set a precedent?
I have my guesses, but I won't pretend to know. What I do know is that Snap has made the right call for itself, right now — it stopped pretending and started optimizing. That's rare in this industry, and worth celebrating.
Because here's the truth nobody wants to admit: sometimes the best way to build AI is to stop building it yourself.
Source: TechCrunch
Related Reading:
- Tech Stocks Reassess AI Valuation as Market Recoil Intensifies — As investors question whether AI spending can be justified by current valuations, companies like Snap are finding creative ways to maintain exposure without bearing the full cost.
- Meta Business Agent Goes Global: How Nicola Mendelsohn's 1B+ Conversations Drive Enterprise AI at Scale — While Snap spins off its AI video team, Meta is doubling down on enterprise AI agents built for the messaging platforms it already dominates.
