He inherited a name from 1985 and pointed it at a problem from 2030. The bet: an open-architecture chip company can outrun the incumbents into the Physical AI era.
The MIPS in MIPS once stood for an instruction set you learned in undergrad computer architecture. The processor that ran the Nintendo 64 and a thousand routers. Then it disappeared into the kind of corporate twilight where a brand has more recognition than revenue. Sameer Wasson took the CEO chair in September 2023 and started rewiring it - new architecture, new market, new bet.
The new bet is RISC-V. The open instruction set that has spent the last decade convincing the industry it could be a real alternative to Arm. Wasson did not invent that conviction. He showed up at the moment it stopped being a hobby. In his words, the past year felt like ten years of maturation compressed into one. MIPS, the company that once published its own ISA into obscurity, is now an active member of the RISC-V economy. It also happens to be owned by GlobalFoundries, the foundry spun out of AMD, which gives the strategy an end-to-end shape that did not exist three years ago.
The target market is what Wasson and a growing chorus call Physical AI. The phrase is doing real work. It means the compute that sits inside the car, inside the robot, inside the inspection arm on the factory floor. The compute that has to make a decision before the next sensor frame arrives. The compute that does not get to phone home to a hyperscaler. It is a market where latency is measured in microseconds and a watt of power is a budget line.
In January 2026, MIPS shipped the proof point: the S8200, a software-first RISC-V neural processing unit pitched at the autonomous edge. Wasson framed the moment in the press release with the language of a man who has watched too many architecture wars: scalable, efficient, open, standards-based. The subtext, less polite, was the line he had been delivering on stage since CES 2024: quit designing around Arm constraints.
That sort of talk is easier when you have spent eighteen years inside the largest broadline analog and embedded vendor on earth. Wasson came to MIPS from Texas Instruments, where he ran the Processors business unit. Before that, he stood up TI's mmWave CMOS radar franchise - one of the few examples in modern silicon of a high-volume product line built from a research curiosity into a billion-chip industrial standard. He was the operator who took a piece of physics and turned it into a price list.
That instinct - the operator's instinct - is what MIPS hired. The company has had architects, evangelists, and turnaround specialists in the CEO chair across its various lives. It had not, in a long while, had a Processors BU manager.
We need to do more processing without blowing up the power consumption. AI has not yet been designed with power budget as a first constraint. - Sameer Wasson, CEO, MIPS
It is rare, in modern silicon, to find an executive who stayed at one employer long enough to see a market be born, scale, and then need to be replaced. Wasson did it twice at Texas Instruments without changing his badge.
He started in TI's Communications Infrastructure Processors business - the side of the company that quietly powered base stations and routing gear. Useful, profitable, and absolutely not the part of TI anyone wrote magazine covers about. He learned how a high-margin semiconductor business actually runs: roadmaps, design wins, customer enablement, the long, untweetable arc from sample to socket.
Then radar. TI's mmWave CMOS radar program was a research bet on the idea that 77 GHz signals could come off a standard CMOS process - no exotic gallium nitride, no specialty fab. Wasson took the resulting silicon and pointed it at automotive ADAS, then industrial sensing. The mmWave business is now a fixture in any modern car that decides whether to brake for you. He built that.
His last role at TI - VP and BU Manager for Processors - was a different kind of work: a re-founding. Wasson re-established Texas Instruments as a mainstream MPU and MCU supplier into automotive and industrial, and put the company on the embedded-AI, software-defined-vehicle, and electrification maps. He left in 2023 to take the MIPS job. Eighteen years, three product eras, one operator playbook.
"We're saying to quit designing around Arm constraints. Have the freedom to innovate. Go innovate. Silicon people now have the tools."
CES 2024"AI has not yet been designed with power budget as a first constraint."
On the next chip generation"As AI moves beyond the datacentre into physical environments, platforms require real-time data orchestration capabilities far beyond what traditional processors can deliver."
S8200 launch, 2026"From a RISC-V perspective, the past year has felt like ten years of maturation compressed into one."
On industry tempo"MIPS is bringing Physical AI to the autonomous edge platforms that need modern AI capabilities in a scalable and highly efficient way that is also open and standards based."
On the product mission"Driving intelligence into action."
The MIPS framing under WassonWasson keeps drawing a line between datacenter AI and the AI that has to act in the world. The car braking. The arm welding. The drone reading a power line. That AI has different rules: deterministic latency, sealed power budgets, no fallback to the cloud. It is a chip problem, not a model problem.
The complaint that AI was not designed with a power budget is more than a soundbite. It is the design brief for everything MIPS is putting into the autonomous edge - architectures sized for the watt budget of a vehicle, not the megawatt budget of a hyperscaler.
His framing of RISC-V is less ideological than practical: the tools are good now, the customer wants the freedom, and the alternative is a tax. Silicon teams have permission to extend the instruction set instead of routing around someone else's.
One detail does most of the work in understanding Sameer Wasson: he spent eighteen years at one company. In a market where executives change logos every twenty-four months, that is its own kind of credential. It means he saw a product cycle finish. It means he had to live with the second-order consequences of his own roadmap decisions. It also means the move to MIPS was not a routine job change. He left the broadest semiconductor portfolio on earth to run a pure-play IP company. That requires a thesis.
The thesis - again - is Physical AI. The autonomous platforms that need a brain at the edge of the network, in the chassis of the car, on the body of the robot. MIPS gets to be the open, standards-based answer. GlobalFoundries gets to fab it. The customer gets to extend it. The story is internally consistent enough that the engineer in Wasson can sell it without losing the operator in Wasson.
What does he want next? The S8200 to find sockets. RISC-V to win the autonomous edge before the field tilts back to incumbents. MIPS at fifty to be more interesting than MIPS at forty.
If the bet works, the company that gave a generation its first lesson in pipelining will become the company that gives the next generation its first lesson in Physical AI. That is a tidy arc - even tidier because the CEO will be the one who, before he ever sat in this chair, taught a different company how to build a market from physics.