It's so common that it's almost a cliché: To start a high-tech company, you need to team a scientist with a business person.

Rajit Manohar, Cornell associate professor of electrical and computer engineering, has been working for over a decade on a way to increase the speed of computer chips. He wasn't planning on starting a company, but he happened to describe his idea to John Lofton Holt, a former electrical engineer who has spent the past decade as a business consultant on high-tech matters. And Achronix Semiconductor was born. Holt is chief operating officer; Manohar is chief technology officer.

Holt is the husband of Alyssa Apsel, the Clare Booth Luce Assistant Professor of Electrical and Computer Engineering, so he spends a lot of time in Ithaca, and it happens that he lives next door to Manohar. "Rajit is a frequent front-porch dweller here," Holt says, and it was on the front porch that Manohar explained his invention.

Manohar had submitted his invention to the Cornell Center for Technology, Enterprise and Commercialization (CCTEC), which is responsible for moving research results into the marketplace. CCTEC had applied for a patent, but according to Ernest Davis, the CCTEC technology manager assigned to manage the invention, it seemed, initially, that more licensing opportunities existed with existing manufacturers, rather than through startups.

Manohar was thinking along the same lines. Or perhaps, he thought, starting a company that would just help other companies design products with the new technology. "I knew that John had done a lot of consulting and had a very good handle on the business side of things, so I asked him for advice, and if he would be interested in getting involved," he recalls.

Holt was running a Washington, D.C., consulting company that does about $3.5 million in business. He decided that Manohar's idea was worth more -- a lot more. After doing some research, he decided the market potential would be measured not in millions, but billions.

What Achronix will market is a very fast "field programmable gate array" (FPGA) chip. Never heard of it? These chips are all around you -- and if Achronix has its way, there will soon be a lot more.

Practically everything electrical has a chip or three or four in it: telephones, cars, video games, microwave ovens, cameras, even some toasters -- some $77 billion a year worth of chips, $24 billion of which are "custom logic" chips, with programming specific to the device. About 75 percent of those are "application-specific" chips, with the programming hard-wired into them. The rest are FPGAs, currently representing about a $4 billion market.

A "gate array" is just a lot of transistors that act as configurable switches, turning on or off to represent the ones and zeros of computation. On a programmable gate array you can tell those switches how to set themselves to perform certain tasks, and you end up with a chip programmed for the device in which it works, just like the application-specific chip. Except that a manufacturer can use the same FPGA chip in a lot of different devices, change the programming when a new model comes out, and even change the programming after the device is in the customer's hands via an upgrade sent over the Internet or a telecommunications network.

The catch is that application-specific, hardwired chips have typically been much faster than FPGAs. Until now.

Manohar's innovation is a way to speed up an FPGA chip by, ironically, removing the clock. Most chips generate a "clock signal," a beat that every component follows in lockstep, like a rowing crew following the coxswain's call. Manohar devised a way to let each component on the chip run at its own maximum speed, notifying the next in line when it is finished processing and can pass along its result. Engineers call this asynchronous (i.e., not synchronized) processing. FPGA chips using Manohar's technology can run as fast as application-specific chips while offering the flexibility of programmability, and adding power efficiency as a bonus. "Our competition is not FPGA vendors," Holt says, explaining that Achronix expects to sell to customers who previously used application-specific chips. "We expect to at least double the FPGA share of the custom-logic market," he says.

Holt and Manohar had their first front-porch discussion about forming Achronix in April 2004. By the end of September 2004 they had brought in as partners two Cornell Ph.D. alumni, Virantha Ekanayake and Clinton Kelly IV, both experts in asynchronous design. The four together contributed over $1.2 million of their own capital.

CCTEC's Davis is managing the patent application and has executed an agreement for Cornell to license the patent to Achronix. Under Cornell policy, inventions by faculty are the intellectual property of the university, with licensing revenue shared among the inventor, the inventor's department and the university. So, technically, Manohar's company had to secure a license to market its co-founder's invention. Fortunately, Holt says, speaking from a lot of business experience, "Cornell is a very pleasant organization to work with."

As soon as a prototype chip was designed and tested, customers started calling, Holt says. "It's a very mature market, and people are always looking for incremental improvements -- but what we offer is not incremental, it is disruptive," he explains. The company expects to ship its first orders by the end of next year, he says. Achronix is offering two product lines: a very fast FPGA chip and another device designed to operate in "extreme environments."

It's still very much a start-up. Holt is looking for at least another $9 million in capital but doesn't expect that to be a problem. In between potential customers, he says, he's spending literally hundreds of hours on calls from interested investors, including not only traditional venture capital firms but also investment banks and larger institutional investors.

Manohar, meanwhile, is on sabbatical at the Massachusetts Institute of Technology, popping in and out at Achronix to work on the engineering side of the business and expressing some amazement at how far it has come. After his sabbatical ends and Achronix is well on its way, he says, he will probably spend one day a week, at most, on the company. "It's too much fun to be a college professor," he says.