Will Intel Follow Applied Materials into Batteries?

By Michael Kanellos

First Applied Materials (AMAT) jumped into the battery market. Now Intel (INTC) may follow suit.

Former Intel CEO Andy Grove told The Wall Street Journal that he is urging the company to get into the market for batteries for electric cars in some manner. While Grove is no longer a member of Intel's day-to-day management team, he casts a long shadow and his advice is often well-heeded.

Earlier this year, Grove threw himself into the electric car debate, writing editorials, agreeing to lecture at Stanford on electric cars and speaking at conferences.

Whether or not (or even how) Intel might follow his advice remains to be seen. Intel has a long history of trying to diversify its business by branching into other markets, as well as a long history of embarrassing, expensive retreats from new markets. In the last decade, Intel has tried and subsequently scrapped efforts to sell Web-hosting services, name-brand consumer electronics, communications gear, digital cameras and TV chips.

Nonetheless, sometimes it works. In the beginning of the 1990s, Intel had less than 3 percent of the market for server processors. Now it has over 95 percent of that market. Server processors, however, are somewhat similar to desktop chips. Grove also pushed Intel to get into health care technology. Skeptics sneered, and the effort has yet to pay off, but the effort could reel in some dough once the massive overhaul occurs.

The battery business is drastically different than the market for processors. Processors improve in performance approximately 60 percent a year, allowing manufacturers to produce cheaper, faster and smaller PCs and consumer electronics gewgaws on a regular basis. Batteries improve about 6 percent a year – it takes roughly 10 years for batteries to double in performance, say many battery executives.

A battery is also a complete system: Engineers have to wrestle with chemistry problems while others try to devise new casing materials and controllers. A processor is a single component.

The easiest way for Intel to get into the market would likely be to produce controllers, software and chips that can regulate charging or power delivery rather than complete battery packs. These chips are essentially scaled down versions of the processors that Intel already makes. Intel has a wealth of technology for controlling and regulating power in batteries from its notebook division.

This sort of technology could also potentially be ported to the emerging, and potentially massive, market for industrial electricity storage or electronic filling stations. Currently, wind farms and solar plants do not have economical ways to store power. If it's produced at times when consumers don't need it, these power producers often have to dump the power. In Texas, some wind farms are giving away electricity and paying grid operators to take power from them so that they can still quality for wind tax credits.

Several ideas for storing power – compressed air chambers, chemical reactions with zinc, methanol fuel cells, flywheels, flow batteries – have been proposed. Most are in the prototype stage and all will need the sort of fine-grained computer processing chips provide.

Now, the downside. Controllers don't sell for a lot of money and several people are in the market already.

Intel spokesperson Christine Dotts said the company can't say anything at the moment, but noted that Intel's battery work to-date potentially could be used in cars.

"We already have some investments in battery companies via Intel Capital, and have had for some time, for example Cymbet – solid state battery for wireless sensor networks –and Polyfuel – improved fuel cell membrane," Dotts wrote in an email. "We look at a lot of things, and we certainly consider battery technology important. Whether we will do anything more in this area we can't say at this time. It should be noted that battery technology developments for computer uses and for automotive applications are not necessarily mutually exclusive."

Any shift to batteries, of course, would reinforce one of the biggest trends in greentech in 2008: the migration of the computer giants into the field. Applied Materials kicked off the effort in 2006 when it started buying companies that make solar panel manufacturing equipment. This year, Intel and IBM touted solar advances while Sun Microsystems (JAVA) and others increased their efforts to reduce power in data centers.

Applied recently has started to form internal groups and invest in startups in lighting and energy storage.

Comments

[Nose:]

Profit margins on power controller chips is ok, but not anywhere near what Intel gets on their CPUs, and probably similar to what they would get on their chipsets, if they were actually priced separately. Why start a lower-profit margin business, when you have competition locked in a price war with you every other year?

There are also at least a dozen companies, some with decades of experience with power control, already working on next generation power controllers. And many more in China. Intel helped by requiring 135W in a few square inches for the high-end CPUs, and GPUs doubling that. The controller chips for these are in the square mm size range, and hovering around a buck or two. Full processors that are used in microwaves and by many for power management are a fraction of a buck, and do a good job. Now - compete on price? More features? Don't think so. I can just read the headlines, "Intel crafts new 4-bit slave microprocessor" and their stock dropping.

As Sun and Google have experienced, total power consumption, especially for large computer "farms", is the biggest hurt. Why have 4 or more cores, when most of them do nothing most of the time? Turn them off, make it possible for different layers of the system to turn off and save power. Look at most monitors, CRT, LCD... take 5W or less when in power save, versus >100W running. Mainstream CPU's can't do that yet, and it isn't an easy problem to solve. Give it a few years.

Yes, we absolutely need newer ion-based high-density power storage, but I doubt that any silicon-only company can bend their binds completely in that direction without basing a new division on a buyout. Sure they have the money, but my bet is with materials specialist giants that know their phosphates from their photons to cook up some new recipes. Let's hope universities can still get funding for this kind of research.

2008 Dec 17 01:25 PM