INTC’s new “3-D” tri-gate microprocessor technology

INTC’s May 4th announcement:  tri-gate

During its most recent quarterly earnings conference call, INTC alluded to a “revolutionary” announcement it would be making on May 4th about its next-generation microprocessor-making technology.  We already knew that the company was going to shrink the distance between elements of the patterns it creates on silicon from 32 nanometers (1 billion nanometers = 1 meter) to 22 nm this year.  But it hinted that there was more to come.

The “more” at a press conference on May 4th.  The company intends to manufacture its 22nm chips using a new technology that it has been developing for almost a decade–a 3-D process it calls tri-gate.

what it is

Up until now, INTC–like everyone else–has been placing chip circuits on a flat silicon substrate.  For 22nm, INTC intends to make the substrate three-dimensional by raising a series of fins up from the substrate, on and around which it will wrap the chip circuits.  If the conventional chip design looks like strands of spaghetti laid side by side, the new one looks more like Rice Chex.

what it does

The the new 3-D “tri-gate” chips cost only 2%-3% more to make than conventional ones.  Just by being 22nm (vs. competitors’ 32 mm offerings), they’re smaller, faster and use less power.  That speed and power advantage is about 15% (I’m eyeballing an INTC chart).  However, the new process allows the chips to be tuned, so that they do much, much more:

–they will use less than half the power of competitors’ chips in high speeds, or

–they will run 37% faster in low-power applications, like smartphones and tablets.

Assuming the adoption of the new process goes according to plan and works as INTC expects, tri-gate will put the company several years ahead of other chip makers.  And it will underscore dramatically the value of INTC having ownership of its own fabrication plants.

when it arrives

Volume production is slated for the second half of 2011.  INTC seems to be putting particular emphasis on shifting its Atom chip–targeted at markets ARMH presently dominates–to the new process.

the big question

The big question for tri-gate, as with any new technology, is whether it will work as expected.  INTC has close to ten years’ experience with the process–and does not seem to me to be the kind of company to announce a development without great confidence in it.  The rest of the semiconductor manufacturing industry is planning to adopt some version of the technology, just not right away.  Still, until we actually see INTC churning out large numbers of tri-gate chips, we won’t know for sure.  We also don’t know whether, even if the chips perform as advertised, that INTC will be able to displace ARMH in new smartphones or tablets.

investment implications

INTC shares are about flat since the announcement.  They’re trading at under 10x earnings and yielding over 3%.  To me, no good news, and a lot of bad, is already factored into the stock price.  On the other hand, merely applying a market multiple to the stock would imply a price of over $30 (see my analysis of INTC’s 1Q11 earnings).  So the risk-reward relationship seems very favorable to me. What I also find interesting is that besides the case made for a value investor, there’s one to be made to a growth investor as well.  It’s the prospect of much better earnings growth than the market expects, for longer than the market thinks.

About ARMH: it’s trading at 90+ times earnings, on the idea that it is the virtual monopoly provider for microprocessors to smartphone and tablet manufacturers.  INTC success doesn’t necessarily translate into ARMH failure; a lot depends on how fast the market is growing.  But while the company may do well, to my mind the ultra-high multiple represents a significant risk in the stock.

Leave a Reply

%d bloggers like this: