One man’s trash is another man’s treasure; and to that point, one man’s clean room full of old semiconductor manufacturing equipment is another’s vision for the future of custom-chipmaking.
So when a company called SVTC had to sell the assets of what used to be the Sematech manufacturing gear in Austin, Texas, Dave Anderson, the CEO and president of Novati snapped it up. Anderson envisioned taking this small fab, where he had worked for SVTC Technologies for years, and turning it into a place where people could get unique chips made in small production runs. So in October 2012 he founded Novati, a custom fab shop for a new type of chip manufacturing, and the company is launching the new facility Monday.
While this is a pretty esoteric service, it’s poised to become an important one as biology and computer sciences converge and as more application arrise that use new materials as opposed to simply trying to cram more transistors on a chip. The company has developed a new type of image sensor for someone trying to build a better camera, it has integrated some DNA testing devices onto a chip and it has improved night vision goggles for the military by putting a different type of material on top of the traditional silicon used by chips. In doing these things, Novati has become a data point in the the upheaval hitting the chip world.
Greater gigahertz and more storage isn’t the only goal
I’ve covered outlines of this upheaval, as the players change on the client side and in the data center, but there are also technological and business shifts at play. Moore’s Law, the idea that the number of transistors on a chip will double every 18 months, has been under attack for half a decade at least. Pushing this goal is expensive (for example, Intel spent 10 years developing a new type of transistor) and it’s also no longer solving the problem that people want their chips to solve.
In the computing sector, the goal was always chips that could store more data in the same space, or chips that could perform more calculations or chips that simply did everything faster. This was true across a variety of high end chips, whether they were for the telecommunications industry or you just wanted to create iPads with more storage capacity. But now chips don’t always need to be faster — sometimes that just makes the chips more power hungry while with others speed just offers something the end product doesn’t need.
Instead, companies in the toy industry for example, are trying to make cheaper yet more modular chips so they can make cheap toys with connected or interactive features. In the life sciences business, they might not care about the latest process technology (the art of cramming more transistors on a chip) because they’re interested in putting an analog sensors on the chip. The military and telecommunications industry might want more performance, but don’t need a huge number of chips, so they have turned to new materials such as gallium arsenide to boost performance for some of their applications. Those new materials are more expensive, but because these companies can’t guarantee large production runs, they are locked out of the major fabs producing the fastest silicon-based processors.
Novati thinks this is a market opportunity
Unlike the large foundries and Intel that are chasing Moore’s Law in their production facilities, Anderson is stepping away from that race. With Novati, which is backed by Tezzaron as a majority stakeholder, Anderson aims to produce runs of 200-500 wafers for clients that want to take advantage of computing but don’t need to chase Moore’s Law to the new frontiers of physics.
The team at Novati includes the process engineers that were most recently at SVTC Technologies and before that at the Advanced Technology Development Facility, which was a subsidiary of the SEMATECH consortium.