|
MAY 2004
Can NVE’s ‘Ideal Memory’
Transform the Computer Chip?
Even
after more than three years as president and CEO of NVE corporation,
Daniel A. Baker still can’t shake the feeling that he has reached
career Nirvana. His job, he says, is “the opportunity of a
lifetime.”
Baker leads a company of 70 employees—most
have technical backgrounds, 10 with Ph.D.s—that is riding a technology
wave that could revolutionize a good chunk of the electronics industry.
With 15 issued patents to support the claim, Eden Prairie-based
NVE appears to be at the cutting edge of a new form of electronic
memory that has broad applications in the computer world. It’s
called magnetic random access memory, or MRAM. With MRAM, data are
encoded in electronic “spin”—a property of an electron
based on the direction of its rotation (“up” or “down”)
about an axis—rather than its charge. (The industry term for this
technology is “spintronics.”) The upshot for consumers
and businesses: instant-on computers and smaller, more reliable
cell phones.
“It’s just such a fabulous
technology and opportunity here,” Baker says. “When I
arrived [in 2001], I couldn’t believe that there was a company
like this, with so low a profile and such great potential.”
Last year, investors also got wind
of that potential. While there aren’t any equities-industry
analysts following it, NVE’s stock (Nasdaq: NVEC) was one of
the market’s hottest properties in 2003. Going from $8.60 with
hardly any trading at the beginning of the year to $51.30 by December,
with an average daily volume in the six figures. The following month,
it climbed into the $60s.
But while NVE’s technology offers
great promise, it also faces big hurdles—and growing investor debate.
Established technologies that are already baked into product designs,
manufacturing processes, and pricing can be hard to dislodge. What’s
more, NVE is a mouse among such elephants as IBM, Hewlett-Packard,
Fujitsu, Toshiba, and German firm Infineon. Even though the company
has what Baker describes as “watershed” intellectual property,
several of its competitors have impressive MRAM IP portfolios of
their own, not to mention more resources to put them to work. What’s
more, one investor who is “shorting” the stock—meaning
he is betting that the shares will trade lower from recent levels—argues
that NVE’s intellectual property isn’t what it’s
cracked up to be.
Meanwhile, NVE’s share price
has declined since January—as of March 31, NVEC stood at $47.68.
The decline in price largely reflects the fact that one of NVE’s
license partners, San Jose-based giant Cypress Semiconductor, has
been slower than expected to get samples of its spintronics chip
into production. Baker’s task is to make sure that this doesn’t
nip NVE’s potentially glorious future in the bud.
In the Spin
MRAM has been called the ideal memory, because it has the potential
of combining onto a single chip the compact nature of dynamic random
access memory (DRAM), the speed of static random access memory (SRAM),
and the nonvolatility of flash memory. Does that sound like so much
gobbledy-gook? Think of it this way: MRAM has the potential of being
to computer technology what the DVD was to record players.
With DRAM or SRAM, every time you
shut down your computer all of the data stored on your machine’s
circuits goes bye-bye, unless you’ve saved it on your hard
drive. The “computerese” for that phenomenon is “volatility.”
Each time you turn on your computer, all the data stored on your
hard drive must be “rewritten” back onto your DRAM and
SRAM before you can use your programs. That’s why it takes
so long to get your computer up and running, and why it’s so
awful when it crashes.
An MRAM memory chip is different.
Each of its cells contains a very thin insulating layer sandwiched
between two magnetic films. Depending on which way their electrons
are spinning, the magnets can create magnetic fields that point
in the same direction or in opposite directions. These directions
determine whether a cell stores a “0” or “1”
of computer data. A small amount of electricity switches the polarity
of each cell, depending on the data being stored. Because MRAM encodes
those zeroes and ones this way, it doesn’t lose the data when
the computer is turned off. That could mean better data integrity,
smaller equipment footprints, and more reliable technology.
Immediate applications include an
entire new generation of cell phones that can be smaller, lighter
and more sophisticated than current technology, which must employ
all three existing types of memory. With MRAM, all that could be
made much simpler. The hot new term in the mobile industry is a
cell phone on a chip. It's a concept that envisions Dick Tracy wristwatch-style
cell phones complete with voice recognition; cell phones in MP3
players, which allow users to download and e-mail music over wireless
networks; and smaller PDAs and cell phones coupled with global positioning
satellite technology that track everything from your dog’s
whereabouts to the precise location of product shipments.
But the mother lode for MRAM is in
computing. Some developers believe that there will come a day when
they can cram at least as much storage onto an MRAM chip as they
now have on DRAM (the chief form of computer memory), without DRAM’s
disadvantages.
In the past two years, the electronics
industry trade press has embraced MRAM technology. If—and that’s
the operative word here—this type of memory replaces existing memory
architecture in electronic devices, current estimates of the size
of the market for MRAM technology exceed $40 billion a year. Memory
Strategies International, a Texas-based memory-industry research
firm, estimates that the cell phone memory segment could make up
almost $9 billion of that total.
Fifteen years after its founding,
NVE may finally be ready to cash in on that promise.
Let’s Get Smaller
NVE was started in 1989 as Non-Volatile Electronics by James A. Daughton, who worked on early applications of NVE’s current
technology as vice president of Honeywell’s Solid State Development
Center. Daughton has been pursuing the dream of using magnetism
to encode data for more than 20 years. He conceptualized the idea
while at Honeywell, working with Art Pohm, a now-retired professor
of electrical engineering at Iowa State University.
“Back then, Honeywell was focusing
its efforts on developing circuitry for the aerospace and defense
industries,” Daughton recalls. “We were working in 1-micron
increments, which is huge by today’s standards.” A human
hair, as a point of reference, is about 80 microns in diameter.
Since the early days of Daughton’s research, however, high
technology has shrunk exponentially. Now nanometers—a unit of measure
equal to 1 billionth of a meter—or, in some cases, angstroms,
are the relevant units of measurement. One micron is 1,000 nanometers,
or 10,000 angstroms.
Daughton left Honeywell and started
his company with the idea of piggybacking on Honeywell’s efforts.
“My initial intention was to let Honeywell develop [MRAM],”
he says. “I left with the rights to commercial development
of the technology, while Honeywell retained the right to military
and avionics [applications]. So I was just going to ride along.”
Honeywell continued to work on MRAM
technology along with NVE, but it then began to focus on “radiation-hardened”
MRAM to withstand space environments. It was a direction NVE didn’t
wish to take, and Daughton and Pohm (who remains an advisor to the
company) pushed forward with their own MRAM research. They won several
awards from the Small Business Administration under its Small Business
Innovation Research program to fund the development of couplers
(which transmit data) and sensors (for high-precision data-acquisition),
as well as spintronics technology.
The result: NVE’s patents—many
of fundamental importance to the MRAM industry—started to pile
up. Meanwhile, some industry giants were leaping into the fray,
with companies such as Motorola and IBM also accumulating dozens
of MRAM-related patents.
Baker entered the story in 2001. Taking
over as CEO from Daughton (now chief technology officer), Baker
has a solid résumé that includes a Ph.D. in engineering,
a University of Minnesota MBA, and a stint as president and CEO
of Printware, an Eagan-based maker of computerized pre-press equipment
for the printing industry. He also was the director of research
at Minntech, a Plymouth-based manufacturer of medical devices, sterilants,
and water purification products.
“We knew we needed to do two
things,” Baker recalls. “Get to market with some of our
core technology to show we knew how to run a business, and establish
strategic relationships with companies that had the financial muscle
to accomplish things that we didn't have the capital to pursue.”
So far, Baker has delivered. In its
most recent quarter ended December 31, 2003, NVE reported $577,000
in earnings on $3.1 million in revenues, which were up 33 percent
over last year. Earnings were up 162 percent year over year. The
company’s revenue-producing businesses include magnetic-based
sensors and data couplers. NVE's tiny sensors, which incorporate
spintronics, are used to position pneumatic Cylinders and robotic
arms, to sense the speed and position of bearings and electric motor
shafts, and in overcurrent and short circuit detection, to name
just a few applications. NVE’s 21,362-square-foot facility
is capable of producing up to 40 million such devices per year,
going to market through a worldwide network of sales representatives
and distributors.
“We believe that we have some
watershed patents, and that our intellectual property is much broader
than the applications being developed by our partners,” Baker
says. His optimism isn’t without foundation. A key factor behind
MRAMs promise is that it allows engineers to cram much, much more
circuitry into a very, very small space. That could allow NVE to
be part of a potentially huge industry—really small technology.
The Next Big Thing?
Some in the investment community have termed nanotechnology
“the new Internet.” (Indeed, one of the hot new books
in the technology space is The Next Big Thing is Really Small,
by Minneapolis nanotech consultants Deb Newberry and Jack Uldrich.
Uldrich was deputy director of the Minnesota Office of Strategic
and Long-Range Planning under former Governor Jesse Ventura.) The
“nano” in nanotechnology is short for nanometer. What’s
more, the technology involved here often is measured in atom. A
conductive layer in one of NVE’s sensors is only 12 atoms thick.
NVE is “on everybody’s nanotech
list,” says Clint Morrison, an analyst at Piper Jaffray in
Minneapolis who has covered the local technology industry for more
than 20 years. “It’s really tough to invest in this area.
You’re often dealing with companies like IBM and Motorola.
And here you have a pure play in that space.”
As for MRAM, “the technology
is real,” asserts Morrison (who owns no NVE stock). Indeed,
he says that MRAM “could be the holy grail of the next generation
of memory. The question now is, can it be commercialized?”
There’s another question: Can
NVE be the one to commercialize it?
A year after its founding, the company
began notching its belt with a who’s who of strategic partners,
licensees, media attention, and more government backing. These include
contracts from the U.S. Army, the U.S. Air Force Research Laboratory,
and the Defense Sciences Office of the Defense Advanced Research
Projects Agency, as well as grants from the National Science Foundation.
Meanwhile, spintronics technology has been featured in an array
of esoteric engineering and science-related trade publications,
and was the cover story in the June 2002 issue of Scientific
American.
While this sort of attention is nice,
it’s the company’s license agreements with the likes of
Motorola, Cypress, Agilent Technologies, and Union Semiconductor
Technology that could take this story to the bank.
In late 2000, NVE cleared away any
lingering licensing complications with Honeywell in exchange for
a payment to NVE of more than $1 million. In the fall of 2001, NVE
granted nonexclusive rights to some of its coupler technology to
Agilent Technologies for upwards of $1 million in fees and advance
payments in the first year of the agreement, along with future payments
based on sales of the Agilent products covered by the agreement.
At this point, NVE’s big-money
bet is with Illinois-based Motorola, which owns about 5 percent
of NVE. This relationship dates to 1995, when Motorola licensed
the company’s MRAM technology for the development of new cell
phone technology. Last October, Motorola announced that it had produced
its first samples of MRAM chips—well ahead of a joint venture between
IBM and Infineon. Motorola has targeted pilot production for late
2004. If Motorola uses NVE’s technology, the Minnesota firm
would earn a 1 percent royalty on the sales of the chip.
But the big payday for the company
will come with licenses from other manufacturers that step through
the door after the technology has been proven. That is, if
that payday comes.
Great Expectations
For all the minuteness of NVE’s products (and its relatively
small revenues), investors have been willing to pay big dollars
for its shares. Even though NVE’s stock price has slipped from
a 52-week high of $69.69 down to the mid-$40s, its shares are trading
at more than 18 times sales—and 126 times projected full-year
earnings for the fiscal year that ended in March.
It was inevitable that such valuations
would attract short-sellers—investors who borrow stock from other
investors in hopes that the stock will drop in price, allowing them
to buy it in the open market, return the borrowed stock at its original
price, and pocket the difference. In mid-March, Manuel Asensio,
a New York-based analyst, launched an attack on the company’s
valuation, arguing that a combination of high insider sales and
what he contends is low-value intellectual property suggests that
the stock is wildly overvalued.
Indeed, insiders have been
selling—with Baker and Daughton selling or announcing the sale
of stock worth more than $14 million. Baker defends these moves,
which took place primarily in December 2003 and January of this
year, noting that he still has more than 70,000 in vested options
that he could sell, but hasn’t.
“I have retained significant
[equity] interest and exposure to the company’s stock price,”
Baker says. As for NVE’s technology, “We believe we have
valuable intellectual property,” he says. “We have constantly
improved and added to our intellectual property portfolio. The standard
isn’t based just on the intellectual property Motorola licensed
in 1995. We’ve continued to add to that portfolio. We believe
we have one of the best IP portfolios in the industry.”
One of the big factors driving NVE’s skyrocketing
stock price was the deal it signed in April 2002 with Cypress. In
exchange for NVE’s intellectual property and 17 percent ownership
of the company, Cypress pumped $6.2 million in capital into NVE. That
investment bootstrapped NVE’s stock from the over-the-counter
bulletin boards to the much more widely traded NASDAQ, greatly increasing
the company’s liquidity and broadening the regulatory eligibility
for investors. But the really big news was Cypress’ plans to
produce an MRAM computer memory chip by March 2004 and then manufacture
chips at the Silicon Valley firm’s Bloomington facility.
But in January, word got out that
Cypress was encountering delays in getting chip production on line.
Cypress had said only that it hoped to have samples by the end of
March. Meanwhile, bulls and bears have whiplashed NVE’s stock,
pushing it down to the high $30s one day, only to have it bounce
back, often as much as $4 to $5 per share, the next.
In a story that seems to evolve by the minute, Cypress indicated on March
19 that it had produced the chip—not, perhaps, in a full production
form, but at least on a proof-of-concept basis. No specific date
was given for when actual production would begin. Nonetheless, the
news caused the stock price to jump again.
Cypress aside, Baker emphasizes that
NVE’s Motorola relationship is still another powerful component
of the NVE story, and that the Cypress project is focused on only
one type of computer memory. He suggests, in other words, that there
will be other opportunities for his firm’s MRAM technology.
“A lot of the valuation [of our
stock] is based on the quality of our intellectual property and
the tremendous potential of MRAM to revolutionize an industry,”
Baker says. “We believe we have the intellectual property and
technology that makes us the current leader in the practical use
of MRAM. Those kinds of things come along very rarely.”
Tony Carideo (tony@carideogroup.com) is president of The Carideo
Group, a Minneapolis-based integrated corporate communications company.
His firm does not represent NVE.
|
