On a mild Friday afternoon, I sat across from a 5.5-inch-tall AI hologram, fidgeting as its creators prompted it to critique my fashion sense. Fortunately, I passed. “Your style is elegant and soft,” said Sumi, a female avatar with black-and-green hair. “That cream halter top looks really lovely on you, and your wavy hair gives a gentle, classic vibe.” I breathed a sigh of relief and sheepishly thanked Sumi for the compliment.
Sumi—and her male counterpart Nico—are virtual personalities that users can toggle between on Razer AVA, a desktop AI companion powered by xAI’s Grok. Each has a scripted backstory: Sumi is an illustrator from Tokyo who has a calico cat and enjoys “cozy gaming,” while Nico is a freelance engineer and skater living in downtown Los Angeles. Sumi is kind and warm, while Nico is more curt and to-the-point.
They’re some of the newest faces of Razer, a cult gaming hardware brand that’s now trying to become an AI company.
“AI is going to bring a new renaissance to all of gaming,” Min-Liang Tan, Razer’s co-founder and CEO, tells Fortune at the firm’s Singapore office. “Today, gaming is an industry worth hundreds of billions, and we believe that AI will make it a trillion-dollar one.”
The industry generated roughly $188.8 billion in revenue last year, serving over 3.5 billion players worldwide. Tan—who has spent two decades telling anyone who would listen that everyone would eventually become a gamer—is convinced AI is the force that finally gets gaming there.
‘A gamer first’
Tan’s route into the industry was sideways. A lawyer by training, he grew up an avid player of fantasy role-playing games like Ultima, the series created by game designer Richard Garriott in the 1980s.
“I would say I was a gamer first, rather than being a lawyer first,” Tan says. “But like every good Asian child, my parents only gave me the option of being either a doctor or lawyer.”
In 1999, while still practicing law, Tan teamed up with Robert Krakoff, an American businessman he had met while gaming online, to build a gaming mouse. They named it the Boomslang, after a fast-striking African snake, because “we wanted to make a product which eats up other mice,” as Tan tells it. “Everyone else was calling their computer mice the MX2000 or FX500.”
“We went on the servers and we got kicked out for hacking,” he adds.
Tan and Krakoff established Razer in 2005. Now, the firm employs over 1,500 people with dual headquarters in Irvine, Calif. and Singapore’s Queenstown district. About half its business is in the U.S., 30% in Europe, and less than 1% in Singapore, by Tan’s account. “We’re probably more American than Singaporean,” he says.
Within its first five years, the company had built “a cult following”—people tattooing the three-headed snake logo on themselves and building “shrines” to the brand. “We believed everyone was going to be a gamer in the future. Fast-forward 20 years, and I dare say pretty much everyone’s a gamer right now,” Tan says.
He adds that the gaming industry—and the demographic of gamers—has shifted significantly.
“It was a predominantly male demographic aged between 12 and 25, with most of us living in our parents’ basement; it was a little bit of a subculture back then,” Tan says. Now, it’s a 50-50 split between men and women.
AI ambitions
In 2022, Tan and private equity firm CVC Capital Partners took Razer private, ending a five-year-run on Hong Kong’s stock market. (In a nod to its gamer roots, Razer had the ticker 1337, or “leet” in gamer slang). Tan says the company was already starting to work on AI—or what he calls “cognitive gaming”—and going private let him think about the new technology without worrying about the “quarterly commitments of a publicly listed company.”
It’s given him the freedom to think big about Razer’s future. “We are probably going to be one of the largest AI hardware companies on the planet,” he predicts.
On Sept. 28, the firm unveiled a wearable headset named Barracuda AI. The AI-equipped headset comes with built-in computer vision (via integrated cameras), smart audio processing and compatibility across various AI models like ChatGPT and Google’s Gemini.
Previous attempts to release smart glasses have run up against privacy concerns, particularly on non-consensual recording. Tan says the company is taking a “privacy-first” approach to its wearables. “It’s definitely not down the route of where Meta is going,” he says.
Sumi and Nico, Razer’s AI avatars, are another part of the company’s AI lineup. Razer has given its avatars “personality and emotion engines,” Tan says. Razer AVA has yet to be publicly released, but is expected to ship by the “second half of 2026”, according to its website.
Finally, Razer is betting on AI in game development. Last Wednesday (Oct. 7), Google and gaming tools developer Unity launched an AI game creation platform called Playground, allowing users to create video games with a simple prompt.
“Now, everyone can create games with just their imagination, even without technical skills,” Tan says, adding that AI will also fundamentally change how games are experienced and played. “You’re going to see much larger games where there will be virtual worlds run by AI, and that can mimic or transcend the size of our world at this point in time.”
Yet gaming storefronts are already flooded with a sea of so-called shovelware, or low-quality buggy products made with minimal effort. Generative AI, by making it easier to make video games, could make the problem even worse.
One in five games released on Steam last year disclosed the use of generative AI—an eightfold increase in a year, according to a study of the storefront’s catalog. In a single week this June, 120 of the roughly 300 games released on Steam carried an AI disclosure, according to a survey from gaming publication PC Gamer. A separate study of 53,000 titles found that AI-made games account for up to 90% of the growth in monthly releases—and that almost none of them make money.
Last year, Tan and his team launched the Razer QA Co-AI, an AI-powered game testing tool designed to catch performance issues and review AI-generated bug drafts. “We don’t like AI slop,” he says. “When you get a huge glut of content created, which is not very good, we can then help them make it better.”
Tan is unmoved by the debate over whether AI is moving too fast. “AI is going to be inevitable, and the genie is already out of the bottle,” he concludes. “What we can do is to see how we can effectively harness it.”
Disclaimer : This story is auto aggregated by a computer programme and has not been created or edited by DOWNTHENEWS. Publisher: fortune.com
