The engineer who spent forty years chasing a cleaner signal

In the RF world there is a rule of thumb: the higher the frequency, the harder everything gets. Wires start behaving like antennas, a millimeter of packaging tolerance eats your performance, and getting real power out of a chip without smearing the signal becomes a small act of engineering violence. Seyed A. Tabatabaei has been living inside that difficulty since 1986. Most people who start that early move into management, or out of the field entirely. He kept designing chips - and then, at an age when many engineers are winding down, he started a company to build them at scale.

That company is mmTron, an 18-person fabless semiconductor outfit tucked into an office park on Twin Dolphin Drive in Redwood City. It does one deceptively narrow thing: it makes MMICs - monolithic microwave integrated circuits - that push clean, linear power through millimeter-wave frequencies. Those chips end up inside satellite links, 5G and 6G radios, radar systems, test instruments, and defense hardware. You will never see one. If they work, you will never think about them at all.

The long apprenticeshipA career measured in decades, not quarters

Tabatabaei's technical foundation was poured at the University of Maryland, College Park, where he did his PhD and post-doctoral research alongside work at the Laboratory for Physical Sciences. His first industry job was in the millimeter-wave R&D division of M/A-COM - the kind of place where you learn the hard physics before the market catches up to it.

In 1998 he joined Hewlett Packard, which would later become Keysight Technologies, and there his education took a turn. He moved to the manufacturing and operations side of the microelectronics organization and led a cross-functional team. It is one thing to design a chip that works on a bench. It is another to make ten thousand of them come out identical. That second lesson would matter later.

In 2006 he went to Endwave Corporation with a specific mandate: build a semiconductor MMIC division from the ground up. He did. When Renesas acquired Endwave in 2011, he left and started Teramics LLC, a design-services firm. For roughly nine years, Teramics quietly designed more than 46 products for top-tier customers - a working proof that a small team could out-design much larger ones at the hardest frequencies.

The pivotWhy he stopped selling design and started selling chips

Teramics was a good business. It was profitable, it had the best customers, and it never had to carry inventory. But it had a ceiling that bothered its founder. When you sell design services, your best ideas leave with the client. The differentiated technology - the models, the architectures, the tricks learned over decades - reaches only as far as one contract at a time.

The trigger came in 2020, in the middle of a pandemic - an odd moment to start a capital-intensive hardware company. A few angel investors approached him. He and co-founder Mona Molaasgari, who runs operations, quality and supply chain as President and COO, decided to form mmTron Inc. and ramp up product development. The plan they had carried for nine years finally had funding behind it.

Molaasgari's role is not decorative. At millimeter-wave frequencies, the package around a chip can degrade its performance as much as a design flaw. A stray fraction of a millimeter in a laminate or ceramic package changes how the signal behaves. Her focus on dimensional precision and supply-chain rigor is a direct answer to a physics problem, not just a business function.

The thesisClean power is the whole game

Ask most people why 6G or satellite broadband is hard, and they will talk about spectrum or standards. Tabatabaei talks about power. Bandwidth demand keeps climbing, but at millimeter-wave frequencies power is scarce and expensive. You need higher output power, better linearity, and higher efficiency all at once - and the three tend to fight each other. mmTron exists to win that three-way fight.

The company mixes and matches four semiconductor technologies - GaN, GaAs, InP and SiGe - choosing the right one for each chip rather than forcing everything onto a single process. That flexibility, plus proprietary linearity-based device models and custom epitaxial materials, is how a tiny team claims the highest linear power at mmWave frequencies in the world. Their frequency coverage runs primarily from 15 to 60 GHz, with one part, the TMC774, stretching all the way from DC to 160 GHz for instrumentation and 6G work.

The product list reads like a spec sheet because it is one. The TMC211 puts out 20 watts of linear power at high temperature in Ka-band. The TMC212 was tuned for the 5G n260 band. The TMC200 is billed as the highest-power low-phase-noise GaAs broadband amplifier available. None of it is glamorous. All of it is the sort of thing systems engineers lose sleep over.

The founder premiumWhat customers actually buy

There is a reason a hardware startup lives or dies on its founder in a way a consumer app rarely does. A chip is a long, expensive commitment. A customer designing mmTron's part into a satellite payload is betting years and millions on it working. Tabatabaei is blunt about what that means.

That is also the honest description of mmTron's moat. There are very few engineers in the world who genuinely specialize in mmWave and MMIC design. Tabatabaei is one, and he has hired several more. You cannot download that expertise or raise it in a single round. It accumulates the slow way, over careers.

The outlookBetting the demand curve never bends down

mmTron serves four segments: mmWave SATCOM, 5G, instrumentation, and aerospace/defense. Tabatabaei's bet is simple and, so far, correct - that appetite for data will not slow. AI and VR pile new load onto wired and wireless networks, satellite constellations multiply, radar and data centers need the same clean high-frequency signal processing. Every one of those trends needs more of exactly what mmTron sells.

The 2026 Series A - $45 million led by Maverick Silicon, with Tipping Point, J&M and others - is meant to accelerate development and production and push mmTron toward system-level solutions rather than one-off parts. For a company that started three years earlier with angel money during a pandemic, it is a meaningful vote of confidence in a founder who has been at this longer than some of his investors have been alive.

What is striking is how little the core question has changed across his career. M/A-COM, HP, Endwave, Teramics, mmTron - five stops, one problem. Get more clean power out of a chip at frequencies most engineers would rather avoid. He has spent forty years getting better at answering it, and by his own account he is enjoying it more than ever.

It is an unusual thing for a veteran to say. Most careers flatten into routine. His seems to have done the opposite - the deeper he went, the more interesting the frequencies got. mmTron is, in a sense, the payoff of a very patient plan: build the expertise for four decades, then finally point it at products of your own.