Most AI startups in 2025 are chasing scale the same way: raise hundreds of millions, hire aggressively, and hope the product catches up. Gamma went the opposite route and now has something most of its peers don’t: revenue, users, and profits.
The San Francisco-based startup builds an AI-powered tool for making slide presentations. That may not sound flashy in an era of AI agents and humanoid robots. But the numbers speak louder than hype: over 17 million users, 250 million gammas created, and $50 million in annual recurring revenue. Gamma is profitable, growing, and operating with a team of just 30 people.
“I want to convey to the market that small teams can do amazing things,” says CEO and co-founder Grant Lee.
Founded in 2020 by Lee and two colleagues from Optimizely, Gamma had a slow burn start. The original idea was to fix the pain of building decks, something Lee knew firsthand from years spent in consulting and finance. But by late 2022, the team had only 60,000 users and about a year of cash left.
“We were at a moment of truth,” Lee says.
Three Months That Changed Everything
Instead of raising more money, Gamma rebuilt the product. Over three months, the team rewired its infrastructure, redesigned the interface, and overhauled onboarding. Most critically, it leaned hard into generative AI, embedding models from OpenAI, Anthropic, Perplexity, and Google into its backend.
“AI can feel overwhelming to users,” Lee explains. “Our job is to run the tests, do the heavy lifting, and curate the best AI has to offer so our users can create great content without having to think about which model to pick.”
The bet worked. Gamma’s product can now generate a complete, branded presentation from a block of text, removing the need for design skills or template hunting.
“Our goal was, how does AI become a design partner that’s almost like an expert sitting next to you, modifying the content based on your direction,” says Lee. “They’re raising the ceiling, and our approach has been the opposite: to lower the floor.”
This framing “lowering the floor” is core to how Gamma differentiates itself from design-first tools like Canva or Figma, or bloated incumbents like PowerPoint. It abstracts complexity and turns generative models into invisible assistants.
Behind the scenes, Gamma dynamically swaps between over 20 large language models, using Perplexity for outlining, OpenAI for images, and others for formatting or structure. “AI can feel overwhelming to users,” Lee explains. “Our job is to run the tests, do the heavy lifting, and curate the best AI has to offer so our users can create great content without having to think about which model to pick.”
The result: a product that users don’t need to learn, just prompt.
A Product That Pays for Itself
For many, Gamma replaces an entire design function. Stacie Sussman, CRO at RevUp Advisory, says she would’ve needed a team to create client decks Gamma now produces in minutes. She reached out to Lee directly with feedback and he added her to a Slack group of over 1,000 power users. Lee still responds personally when he can.
“It’s the thousand small choices they’ve made,” says consultant Jordan Crawford, who’s earned nearly $100,000 from Gamma-powered courses.
That obsession with iteration runs deep. Much of the early team came from Optimizely, a company built on A/B testing. Gamma applies similar rigor, not just in design decisions, but in which models to use and when. Leaderboards track model performance; new workflows get tested live.
Accel partner Vas Natarajan, who led Gamma’s $12 million Series A last year, puts it plainly: “It’s just maniacal product execution. They are ruthless about collapsing as much complexity as possible.”
Escaping the PowerPoint Trap
Gamma’s timing is fortuitous. Despite decades of dominance, PowerPoint and Google Slides still lock users into fixed 16:9 frames designed for projection screens. Gamma’s “card”-based format breaks from that completely, fluid layouts that adapt to vertical screens, mobile formats, or interactive elements.
“In traditional PowerPoint land, you’re stuck to 16 by 9,” says Lee. “In Gamma, you can have fluid cards of any height.”
That flexibility has opened the door to new formats like vertical decks for TikTok, modular pages that resemble websites, even AI-generated voice avatars to deliver presentations asynchronously. “People hate having to present, and even if they do a Loom recording, they hate having to record themselves, because if they mess up, they have to do it all over again,” Lee says. “So what if we can help them create a presentation that could present itself in your voice?”
That opens up international use cases too. Users could personalize the same presentation for multiple audiences in different languages, with customized messaging—all without re-recording a thing.
Gamma is also building embedded interactive features, including polling, quizzes, and clickable elements that allow users to get feedback or engagement mid-presentation—even if it’s not delivered live.
“You can insert interactive elements in Gamma,” says Lee. “The idea is for you to not only engage the audience but also gauge their level of understanding or even get input live.”
Profitable, Global, and (Still) Not Fundraising
Despite its traction, Gamma remains selective about growth. The company has raised just $23 million total, an anomaly in a market where $100 million rounds are now table stakes. “We have zero cash burn,” says Lee. “We raise the money not really for the funding itself, but to establish a strong relationship with Accel.”
While competitors like Tome raised double that in a single round before pivoting away from slides, Gamma stayed the course and outlasted them. The company has no plans to fundraise again and isn’t seeking an acquisition. If anything, it may look to acquire smaller teams or double down on product.
“Someday [the big players] are going to get significantly better, and we’ll obviously have to address that,” Lee says. “In the meantime, we just want to continue building.”
Lee also wants to avoid the equity dilution that left him with little after Optimizely’s acquisition. “Employees at a lot of these AI startups will end up left with almost nothing in the bag, and that’s unfortunate,” he says. “If Gamma does great and there’s a great outcome, I want every employee to really participate in that.”
Gamma’s strongest growth today is coming from outside the U.S.. Brazil, Germany, Japan, and South Korea are all expanding rapidly. “Many of those markets feel underserved,” says Lee.
But more than global reach or technical architecture, Gamma’s real advantage may be its restraint. While much of the AI world races toward agents and “general intelligence,” If Gamma seems unusually disciplined, that’s by design. Its founders spent months dual-tracking two products before settling on presentations. One idea, a virtual office for remote teams called The Lobby, was scrapped. The other, Gamma, became the focus after plenty of experimentation.
“Being patient and experimenting early on takes time,” Lee says. “But it unlocks a ton of advantages. It means we can take big swings.”