The standard Silicon Valley origin story requires a garage. Steve Bisset could not supply one, because his garage was already occupied by a home brewery, and the beer, by his own account, was excellent. So the workbench went into the bedroom, and Bisset slept on a mattress on the floor beside it, illuminated, as he later put it, by the green glow of an oscilloscope. That bedroom eventually became Megatest, a company that listed on NASDAQ in 1993 and sold to Teradyne two years later for $245 million. Founding myths are usually retrofitted with better lighting. This one comes with a fermentation schedule.
Bisset grew up in Canberra, the son of a civil servant, in what he has called a very normal, boring suburban life in the bureaucratic capital city. He has been admirably unsentimental about it. "I could have had an easy life if I just stayed there," he told the Computer History Museum in 2014. "But no."
By his final year of school he and a friend were running two businesses: a light show and a surfboard export operation. College held no appeal. What he wanted was a summer in America, which required a visa, and the cheapest route to a visa was a university application. He asked a careers adviser which ones to try. She asked what subjects he liked. Maths, science, he said, more or less at random. She suggested Caltech and MIT, institutions he had heard of in the way one has heard of the Pyramids.
He took the offer. He also spent the summer of 1969 driving a van full of surfboard samples from Maine to Key Largo, signing up distributors along the way. Caltech beat MIT on a scholarship he could afford and a geographical misunderstanding: he believed Pasadena was near the beach and had, he admits, "visions of Annette Funicello and all that." Anyone who has driven the 210 freeway in August will appreciate the scale of the disappointment.
The kid who found the 8080
What Pasadena did supply was Carver Mead, whose courses Bisset credits with the whole subsequent trajectory. "I got the smell of it," he said of Silicon Valley. "I could feel the swirl, the vortex of all of the innovation and the energy that was flowing into it and I kind of went, I want to go there and be part of that." Asked whether there had been a master plan, he was brief: "No master plan."
He went to Hewlett-Packard's IC design group for six months and found it incredibly boring. Then Intel, hired by Ralph Ungermann into Federico Faggin's microprocessor group alongside Masatoshi Shima. His first assignment on his first day has the tidy improbability of something invented afterwards, except that it is on tape. The team had a fresh mask run of 8080 chips and no confirmation that any of them worked.
Over eighteen months he contributed to the definition of seven chips in the 8080 peripheral family: interrupt controllers, floppy disk controllers, the connective tissue of early microcomputing. He never completed a chip design. He kept being pulled off projects to start new ones, because Intel had no shortage of people who could design a chip and rather fewer who could sit with a customer and tease out what the chip ought to do.
The period had its compensations. New engineers were put in rooms with Gordon Moore, Robert Noyce and Andy Grove, who asked them what they thought. "The aura of them in the room was like giants," Bisset said. He is careful to note he did not leave in a huff. "I was having a thrilling, thrilling time. I just had this other compulsion."
In his Intel job interview, Ungermann had asked the standard question about five-year plans. Bisset answered that he intended to be running a startup as soon as humanly possible. He was hired regardless, which says something flattering about Intel in the mid-1970s. Eighteen months later he made good on the threat.
Thirty thousand dollars and a factor-of-two error
The co-founder was Howard Marshall, a fellow Caltech alumnus he had somehow never met at Caltech but ended up sharing a house with. They spent roughly two years of nights and weekends working through about six ideas - digital automotive tune-up meters, medical devices - before settling on a tester for the microprocessors that Intel and its rivals were about to produce in volume. The incumbent machines were the size of refrigerators and cost about $300,000.
The capital was $10,000 from vested Intel options plus $20,000 borrowed from friends. They had budgeted $20,000 total to design, source, build and sell the first machine, and were, Bisset says, off by a factor of two. The friends declined to take equity. Their reasoning was a small masterpiece of pessimism: you two are employable, so when this fails you will get jobs and pay us back. They charged seven per cent.
Eleven months after he quit Intel, they delivered the first product. A third man, an Intel engineer named Roman Reiser, had wandered in one day and asked to work there. Told there was no money, he said that was fine. Bisset has said flatly that without him the company might not have cleared the threshold. The office was a room roughly twenty feet square on Walsh Avenue in Santa Clara.
The CEO title arrived by the least ceremonious route in corporate history. Marshall pointed out that customers were confused by a company with no defined organisation and suggested they needed a chief executive. Bisset's recollection: "And so we actually tossed a coin. And that's how I became CEO of Megatest, not just founder." He held the job for about fifteen years, which suggests the coin had judgement.
Twenty to one
By 1980 Megatest was doing $10 million a year at better than forty per cent pre-tax margins. Between 1980 and 1985 it raised roughly $42 million in venture capital, an experience Bisset describes with a wince. He would explain the strategy; the investors would nod. "And I don't think they were listening. They just saw that we were making money and growing." His summary is worth engraving somewhere: "Easy money with inexperienced management can lead you into some traps."
Then 1985 happened. In the fourth quarter of 1984 the company booked $10 million of its legacy product. In the second quarter of 1985 it booked half a million. A twenty-to-one collapse in two quarters.
Megatest quarterly bookings, legacy product
Relative scale — figures as described by Bisset, CHM oral history
Headcount went from about 450 to about 200. He did most of the laying off himself, and has never dressed it up. "It was really heartbreaking. Those are my friends. I kind of thought it was the end of the world." For eight consecutive quarters in 1986 and 1987, the single sale that decided whether the quarter was profitable or not was booked, sold and shipped in the final week - in an industry with three to six month lead times. Competitors helpfully spread word that Megatest was in Chapter 11. Every sales call opened with the financial statements.
He is unusually specific about which mistakes were his. The new tester-per-pin architecture was kept secret from customers on the theory that it was too clever to share. Test engineers were made to program in Unix. The test head was too big. "What we think is right, if it's not what the customer thinks is right, it doesn't matter what we think is right. And if we didn't ask, they would have told us." He attributes it to "a bit of hubris that you get by being young and successful and having people throw money at you."
Three takers out of six hundred
The moment he chooses as the best of his career is not the IPO. Deep in the bad years, with the stock's fair market value at a penny and the investors having written the whole thing off, the company offered to buy shares back from employees at $1.50. Six hundred-odd people received the offer, many of them already laid off.
Some of those holders eventually sold at around $100. The refusal also left the company with more than 500 shareholders, making it an SEC-reporting entity years early and, Bisset notes with an accountant's satisfaction, making the 1993 IPO remarkably smooth. Recovery came the unglamorous way, through customer-funded redesigns with IBM and Intel. Revenue was back around $50 million by 1990. NASDAQ, ticker MEGT, 1993. Teradyne, $245 million, 1995.
Ask him what the legacy was and he does not talk about testers. Megatest, he says, was "not only a venture, it was a social experiment and a management experiment." Company cafeteria. No suits. Consensus decisions. Hiring by panel, in which a candidate met the boss, every peer and every future report, and the group debriefed together afterwards - a method he found lethally effective at revealing whether a manager was, in his words, productive or political, since political candidates adjust their manner according to who outranks them. All of this reads as unremarkable now. "These were radical ideas when we were playing with them. A lot of people told us we were crazy." Then the team read In Search of Excellence and thought: that's us.
He also never managed to explain any of it at home. "I never could explain to my mother. An iPhone, yeah, that's pretty cool, but test equipment for a production line, why is that cool?"
The market nobody was talking about
In 2008 he decided to go into energy, on the grounds that it was the biggest and most worthwhile thing available. His reasoning is not the standard cleantech pitch. He cites a World Bank correlation between cheap dependable electricity and standard of living, calls power "a fundamental driver to civilization," and then makes an argument that is essentially constitutional: distributing the ownership and control of electricity distributes economic power and political power along with it, which he says "makes the world safer for democracy."
His method for picking the opportunity was pure inversion. Look at a panel of venture capitalists, note what excites them, and cross all of it off. They are ahead of you, they are smart, they are funded, the odds of catching up are close to zero. Then look for the silence. "Nobody was talking about steam engines and solar - wow, this is interesting."
He met Robert Mierisch, an Australian engineer who had been reviving reciprocating steam engine designs. Bisset is candid that most people's first reaction to Mierisch was that he was a wacky engineer with a ridiculous old-fashioned idea. The two turned out to have been born in the same town. Terrajoule was founded in 2009 in Redwood City.
The thesis is deliberately unfashionable. Concentrated solar heat, energy stored as pressurised saturated water at very high round-trip efficiency, and conversion through a piston steam engine. No new physics. "We're not big entrepreneurs who're going to invent some fundamental breakthrough in science. We're too impatient for that." The engines were refined over three centuries and then abandoned commercially around 1955. "If we started from scratch, it would have taken us 100 years to refine it."
Recruiting posed a novel problem. "We can't go and hire 10 people who have years of experience designing and manufacturing steam engines - they're dead. And so we dig out the literature." Mierisch went to the Smithsonian to study a Second World War naval unaflow engine. For manufacturing they went to Michigan, on the grounds that "Detroit is the Silicon Valley of piston engines."
An early working system ran on an almond farm near Oakdale, California, pointed at the irrigation pump market - some 300,000 pumps across the western United States, owned by what Bisset approvingly calls "rational customers who understand the difference between capex and opex." In November 2013 Terrajoule raised $11.5 million led by New Enterprise Associates with Air Liquide participating, on a pitch that did not lean on government subsidy.
The company has since restarted as Terrajoule Energy Inc., based in Palo Alto, with a patented modular storage product it claims will hold full capacity for 25-plus years without degradation, cycle limits, rare materials or a gigafactory. Partnerships run to Coimbatore and Vancouver. In May 2026 he turned up in New Delhi at a TERI consultation on concentrated solar power with thermal storage, still making the same case.
Two companies, forty years apart, in industries with nothing in common. The through-line is a preference for problems other people have stopped looking at, an aversion to inventing anything that already exists, and a piece of advice he offers without much ceremony:
Asked once whether he was a glutton for punishment, given that capital equipment leaves you, in his phrase, "at the end of a whip," Bisset gave the shortest possible answer. "Lack of adaptive behaviour patterns, yes."