Retirement, for Mark Granahan, apparently required reading material. In 2014, after decades in semiconductors and the sale of his first startup to Texas Instruments, he stepped away. He relaxed. He thought. And, because some people are constitutionally unable to leave a technical problem alone, he read IEEE publications and research journals. One paper described an experiment in photovoltaic cells. Buried inside was a charge-balancing idea that seemed capable of doing something useful in a different neighborhood: the power transistor.
The object of his attention was hardly a celebrity chip. A power MOSFET is a switch. It blocks voltage, conducts current, then switches again, sitting inside chargers, cars, robots, telecom equipment and data centers. The ideal version would waste no energy as heat. The real version always does. Multiply each tiny loss by billions of switching events and millions of devices, however, and the humble component develops a rather expensive shadow.
Granahan knew the terrain. A materials scientist by training, he had moved through engineering, product management, marketing, acquisitions and operating leadership at Texas Instruments and Lucent Microelectronics. He had also founded Ciclon Semiconductor in Bethlehem, Pennsylvania. Ciclon began with three employees in 2004, raised $24 million, built high-performance power devices and was acquired by Texas Instruments in February 2009. It was a compact education in what it takes to get physics out of a lab and into a buyer's bill of materials.
A technical paper interrupts retirement
The photovoltaic research that caught Granahan's eye involved high-k dielectrics, materials that can hold back electrical charge while transmitting its field, and atomic layer deposition, a method for laying down films only nanometers thick. Neither ingredient was science fiction. Both were established enough to be interesting. His question was whether they could be rearranged to change the geometry of a high-voltage silicon switch.
So he purchased time at Penn State's Nanofabrication Laboratory. He tried combinations of dielectrics and process recipes. The work eventually showed that the concept could function, though making it outside a university lab would demand specialized equipment and a cooperative foundry. The romance of invention met its natural companion: process engineering.
“There wasn't this aha moment. But there was this learning process that I had to go through.”Mark Granahan, on SuperQ's beginning
That sentence is a useful antidote to the founder mythology of lightning strikes and napkin sketches. Granahan's idea arrived as a sequence: notice a paper, carry the mechanism into another field, test it, learn what breaks, and find the next starting point. In 2017, he cofounded iDEAL Semiconductor with device physicist David Jauregui and executive chairman Michael Burns. They set up in the Lehigh Valley, near Lehigh University and the Ben Franklin TechVentures ecosystem that had helped Ciclon years earlier.
Put more of the silicon to work
High-voltage power transistors live with an old compromise. A device must block substantial voltage when it is off, but it should offer little resistance when it is on. Conventional superjunction designs balance charge by alternating differently doped regions inside the silicon. The arrangement handles voltage, yet it leaves roughly half the device's cross-section available for the current-carrying n-region. Half the expensive little square is doing the conductive work.
SuperQ changes the arrangement. Its narrow, deep trench is lined with extremely thin dielectric films and shaped asymmetrically. In iDEAL's published description, that architecture can expand the conducting region toward 95 percent of the die. More area for current means less resistance and less energy lost as heat. The clever part is not simply the result; it is the choice of ingredients. SuperQ is designed around conventional CMOS-compatible steps rather than requiring a wholesale departure from silicon manufacturing.
Illustrative cross-sectional comparison based on iDEAL's description of n-conduction area. It explains the design principle, not a product benchmark.
This is Granahan's bet on mature technology: maturity is not the same as completion. Silicon has an enormous base of factories, reliability knowledge, packaging practice and customer confidence. A new architecture that fits those assets can attempt to improve performance without asking the supply chain to learn an entirely new language. That is less theatrical than replacing silicon. It may also be easier to buy.
A startup looks for a factory
The architecture was only one problem. The founders also needed a U.S. facility with the right tools and the patience to help develop a new process. In 2017, they could not find one that matched the requirement. They worked with a small California manufacturer and even invested in equipment there. For a software company, an imperfect development environment is annoying. For a chip company, it can consume years.
There was capital available from Chinese or China-backed investors, Granahan has said, but it often came with manufacturing or intellectual-property conditions. His caution was earned. He said Ciclon's intellectual property had escaped an overseas fab and surfaced in competing Chinese devices. At iDEAL, domestic production became both a supply-chain aim and a form of memory.
The experience widened his focus from one transistor to an industrial system. A country might possess gifted designers and large chip companies yet lack the shared facilities that carry an unfamiliar process from experiment to pilot production. Granahan and Burns went to Washington. They met congressional staff, members from both parties and, in their account, virtually anyone willing to hear the argument. Burns estimated as many as 75 meetings.
“Asia had a pocket veto on semiconductor development in the United States.”Mark Granahan, on the manufacturing bottleneck
The CHIPS and Science Act eventually directed attention and money toward domestic capacity. Granahan's view is more specific than a celebration of factories. Advanced fabrication needs packaging and testing. Production needs trained engineers. Startups need access to pilot tools and venture capital willing to wait through physical development. Remove one leg and the table becomes a rather elaborate way to spill coffee.
He has remained blunt about the talent question. In a 2025 interview, he argued that the United States needs a culture capable of supporting technical innovation, including advanced degrees and long-horizon company building. He also noted a small encouraging signal: greater visibility for chips appeared to be drawing more students toward internships and semiconductor roles at iDEAL.
The long middle reaches a loading dock
In July 2025, SuperQ moved into full production with iDEAL's first 150-volt MOSFET; 200-volt devices began sampling. The company named Polar Semiconductor as its manufacturing partner that September. Polar operates a 200-millimeter facility in Minnesota with experience in high-voltage and power devices. A month later, iDEAL announced automotive qualification for SuperQ and a 200-volt product intended for demanding power systems.
By March 2026, DigiKey had added the devices to its global distribution platform. The sequence is prosaic and therefore important: production, foundry relationship, qualification, distribution. An invention becomes a business through nouns such as inventory, package, sample and purchase order. Seven years of material science ends, for a customer, with an add-to-cart button.
Begins his semiconductor career at Texas Instruments after studying materials science at Penn State.
Starts Ciclon Semiconductor with three employees in Bethlehem.
Texas Instruments acquires Ciclon and its power-device work.
Retires from TI, then resumes technical reading and laboratory experiments.
Cofounds iDEAL Semiconductor with David Jauregui and Michael Burns.
SuperQ MOSFETs enter production and Polar becomes the manufacturing partner.
DigiKey begins global distribution of iDEAL's production devices.
Experience becomes architecture
Granahan's career has an appealing circularity. The materials-science graduate became an operator. The operator became a founder. The founder sold a company, returned to the corporation, retired, and became a laboratory-minded founder again. Each loop carried something forward: acquisition experience from Lucent, power-device knowledge from Ciclon and TI, sensitivity to intellectual-property risk, relationships in the Lehigh Valley, and a view of manufacturing as part of the product rather than a task that happens elsewhere.
There is also a quiet civic streak. iDEAL works near the remains of Bethlehem Steel, a monument to American manufacturing with all the subtlety of a blast furnace. Outside its workplace sits a life-size Benjamin Franklin. Granahan has admitted to a quiet moment or two beside the old electricity enthusiast. It is a charming image, but the company was not built by communing with bronze. It was built through process recipes, fundraising, foundry searches, policy meetings and years when the product could not yet be ordered.
What can another founder take from this? First, read beyond the boundary of the problem; a solar-cell paper may contain a transistor company. Second, choose an architecture that respects the installed world, because factories and customer habits are part of the design brief. Third, treat scars as requirements. And finally, do not confuse the absence of a lightning bolt with the absence of an idea. Learning can be the invention process.
Granahan's stated ambition is larger than one package or voltage class. He wants lower power loss across the growing machinery of electrification, and a U.S. ecosystem capable of turning technical insights into manufactured products. The test will not be whether silicon sounds fresh. It will be whether engineers can buy the devices, put them on boards and measure less wasted energy. Retirement gave him time to think. Production gave the thought somewhere to go.