Stephen Huang spent decades developing chips in Silicon Valley, but the launch of ChatGPT convinced him that he was ready to bring to market the AI chip company he had long considered starting.
By then, Huang was working on GPUs from MediaTek and GPUs from Apple. Face ID technology and Amazon AI chip team.
When ChatGPT arrived in late 2022, Huang believed the industry had reached a tipping point. “We felt the market had arrived,” he said.
So, now 55 years old, Hwang decided to start over.
In 2024, he founded Tranxform AI, a Taiwan-based AI chip startup that develops power-efficient processors designed to run AI models outside of large data centers.
The company currently has about 40 employees and is preparing its first chip, which Huang expects to have ready by next year.
The CEO is one of a growing number of entrepreneurs chasing the opportunities presented by the AI boom. But unlike many founders who build companies around large-scale language models, he spent decades designing chips before launching his own startup.
However, Huang never considered age a disadvantage.
“Morris Chan started TSMC in his 50s,” he said, referring to Morris Chan, the founder of Taiwan Semiconductor Manufacturing Company, who founded the world’s largest contract chip maker at the age of 55.
In fact, Huang believes age may be to blame, arguing that hardware startups, unlike software startups, often value experience.
Semiconductor design is a long-term game, he says. Building systems-on-a-chip (integrated processors that power devices ranging from smartphones to AI systems) requires balancing countless tradeoffs across hardware and software, skills that can take decades to develop.
“Building a good SoC takes experience,” Huang says. “Otherwise, you won’t know how to balance different tasks.”
Huang’s startup adventure was anything but obvious.
Before founding Tranxform, he had a stable career and good income in the United States. Starting a company meant taking a big risk and spending most of his time in Taiwan, a decision his family initially struggled to accept.
As Tranxform grew and reached important milestones, his family’s attitude changed. “Today they are proud of what we accomplished,” he said.
The timing was also helpful. Mr. Huang’s two sons were adults when he founded the company. One works in the technology industry and the other is a recent college graduate.
“Their independence makes it easier for them to dedicate the time and energy needed to build Tranxform,” he said.
yellow said he The demand for dedicated AI hardware will continue to grow as companies seek faster and more energy-efficient ways to run increasingly complex models, so we thought it was a risk worth taking.
Huang’s optimism comes as investors pour new money into AI hardware.
Venture funding for AI and machine learning chip startups rose more than 70% year over year to $16.2 billion in 2025, according to PitchBook. But the number of deals fell from 266 to 232 in the same period, as investors increasingly wrote larger checks to a smaller group of breakout companies.
As of June 22, the company has raised $9.9 billion and made 87 deals this year, according to PitchBook.
going home
While AI drew Huang towards entrepreneurship, the increasing competition for talent in Silicon Valley made him reconsider where to build his company.
Years of experience in Silicon Valley taught him how difficult it is for startups to compete for engineering talent with tech giants like Google, Apple, and Nvidia.
“We kept training people, but they were pulled out,” he said.
Huang founded his company in Hsinchu, Taiwan’s chip hub. We believed we could build a more stable engineering team there.
One of those new hires is university classmate Weixin Li, who joined Tranxform last year as chief technology officer after leaving U.S. semiconductor giant Qualcomm.
One of Huang’s (right) key new hires was his college classmate Way-Shing Lee (left), who is now Tranxform’s CTO. Steven Huang/Tranxform
Now, startup life has replaced Huang’s previously stable career with fundraising, recruiting, customer meetings, and constant technical problem solving.
“Starting a company is very difficult,” Huang said. “You have to find a business partner. You have to sell your story. You have to find funding.”
Huang said Tranxform is preparing for its next round of funding, but did not provide details.
The company is still in the early stages of licensing and has little revenue, he said.
Still, Huang believes the biggest opportunity in AI lies ahead.
The AI industry is “probably still in its infancy,” he said.