News & Analysis
Top 25 predictions for semis in 2010
Mark LaPedus
1/3/2010 10:04 PM EST
Happy New Year! 2010 is just beginning to unfold in the electronics industry and there is already uncertainty in the air.
Looking into my crystal ball, I have released my own chip forecasts--and other predictions--for 2010. These are my top 25 predictions, which are strictly my opinions.
1. IC forecast: Funny money demand?
It's time for my annual chip forecast. Drum roll, please! Most forecasters see a big double-digit growth year in 2010. Unfortunately, I don't. My IC forecast for 2010: Plus 5 percent.
I see a small recovery, but I'm not as bullish as some. First, the worldwide economy still feels weak to me. It's better than last year. The economy is resilient in some nations and weak in others. I'm afraid the governments in some nations (i.e. China, Japan, U.S.) have propped up their respective economies with ''crazed money'' or what I call funny money. Stimulus packages, bail outs, cash for product programs (i.e. appliances, auto clunkers) have certainly jumpstarted some sectors, but will they have a lasting effect?
In some respects, I believe the bail outs and ''cash for whatever'' programs sends the wrong message to consumers. Take the ''cash for clunkers'' programs in the U.S. or similar programs in China. Instead of buying a car when it makes sense, the government is providing false incentives--or bribes. Consumers may end up getting conditioned to buy -- if and only if the government bribes us to do so.
I'm also worried about lackluster consumer demand, the housing woes and unemployment, especially in the U.S. Too bad. For once, there are some cool products in the market: iPhones and iPhone wannabes, cheap netbooks, e-book readers, snazzy LCD/LED TVs, etc.
2. Yawn. Another fab tool downturn
Same old story. Another bad year for equipment makers. My semiconductor equipment forecast for 2010: Minus 10 percent.
I can barely remember a good year in semiconductor equipment. The business was booming when I started in high-tech journalism in the mid-1980s. In those days, chip makers believed that ''real men have fabs.'' Fabs were part of the equation in chip production. Equipment was also a valuable piece of the puzzle.
Those were the glory days in equipment. But those days are long gone. I also remember the fab tool boom during the ''dot.com'' explosion. I also remember the fab tool bust after the ''dot.com'' explosion and 2001 downturn.
Since then, it's been a struggle for fab tool vendors. 2009 was bad. Towards the end of last year, the signs did indeed improve. Intel, Hynix, Nanya, Samsung, TSMC and a few others bought gear.
I'm afraid the party is over. In 2010, there are no new fabs on the drawing board, according to SEMI. Repeat: No new fabs are expected to be built in 2010. That's just simply unbelievable. We will continue to see some technology and capacity buys, but without a new fab, the equipment industry may see another lousy and down year.
3. Solar hype--again?
2010 forecast for solar: Minus 5 percent. The solar industry suffered its first-ever downturn in 2009. The recession put a damper on credit, which impacted the procurement of solar cells for businesses and homes. I'm told credit is still tight, especially in the U.S. Solar still depends on attractive subsidies. That still concerns me.
But here's the real problem: In my opinion, the solar industry has failed to deliver on its promises. Efficiencies remain low. I see little evidence that grid parity is on the horizon. We were sold a bill of goods by some solar vendors. That doesn't mean solar is a flop. There are still opportunities for new and breakthrough solutions. I am still an optimist on solar, but let's cut down on the hype.
4. DRAM blues
Look out for a rebound in demand, but here comes another round of consolidation. My surprise prediction is that South Korea's Hynix Semiconductor Inc. will become U.S.-based company in 2010. Recently, Hynix' shareholders put their stake on the block. The only bidder was a shaky Korean firm, which has since withdrawn its bid.
Hynix' shareholders have put its stake up for sale again. I don't see any bidders coming from Korea. So, Hynix could end up like MagnaChip Semiconductor Ltd., the logic spin-off of Hynix. Last year, MagnaChip fell into bankruptcy. Last year, a U.S.-based investment firm, Avenue Capital Management II L.P., acquired a majority stake in MagnaChip. All told, MagnaChip is a U.S. company now. I see a similar scenario with Hynix, that is, the company being acquired by a U.S. investment firm. Perhaps Avenue Capital could also buy Hynix.
In Taiwan, the island's DRAM vendors are still hanging on. That's a miracle in itself. But in 2010, I predict Powerchip Semiconductor Inc. will fold. Look for Elpida Memory Inc. to buy them. Speaking of Elpida, the Japanese DRAM house is in danger of becoming the next Qimonda. In other words, it could go under--if we see another bad year in DRAMs. Elpida has skipped a node to play catch-up with its rivals. Its diversification strategy, including moves into the foundry and LCD driver market, is still in question.
Back to Taiwan. Taiwan's ProMOS Technologies Inc. will continue to limp along. ProMOS is already looking for niche DRAM markets. Nanya Technology Corp. appears to be a contender, thanks to help from its technology partner, Micron Technology Inc.
5. Micron's fate
Speaking of Micron, this is a make or break year for CEO Steve Appleton. Micron recently reported its first profit in three years. Its performance has not been stellar to say the least. And one quarter does not make it a success.
So if Micron falls back into the red, look for Micron to merge with Numonyx BV. At one time, rumors swirled that Micron and Numonyx were set to merge. The chatter has quieted. Perhaps a deal was once put on the table. Then, Micron resisted.
In 2010, I see Intel putting pressure on Micron to merge with Numonyx -- if Micron goes back into the red. A backroom hostile merger? Intel has been known to engineer such moves. Micron and Numonyx have plenty in common with Intel. Micron has a joint NAND venture with Intel. Numonyx was formed from the former flash groups from Intel and STMicroelectronics Inc.
Comments
LouCovey
1/4/2010 1:07 PM EST
Mark, have to say I agree with most everything but I'm beginning to think that the companies that still have fab capacity will be rethinking their plans to abandon the business. Right now, capacity is incredibly tight and TSMC just hasn't kept up with the potential for growth, which even you say is coming in the next 12 months. No one has invested in new technology, much less new equipment and anyone with capacity will have a chance for increasing business. The industry, as a whole, has had it's head in the sand for almost a decade and when they emerge, everyone will be scrambling for new answers. VCs who walked away from semi investment will be pounding their heads on the wall at the lost opportunity. i think the next 12 months will see a chaotic stampede to catch up with worldwide demand. If course they will over correct and it will tank in 2011.
Sign in to Reply
chipchap42
1/4/2010 3:28 PM EST
Mark,
great article and congratulations on sticking your neck out in such a concrete fashion. Unusually specific and focused predictions!
On the CEO front, I think you missed one segment: FPGAs. Xilinx's Gavrielov is reported to be "enjoying" a popularity rating of 10%, with rumors circling about old management being brought back. Altera's Daane has had 10 years with almost no progress in any of the metrics that matter - top-line growth, market share or profitability. I can't remember who's running Lattice, but apparently neither can most of its employees. So what about that segment?
Sign in to Reply
Mark LaPedus
1/5/2010 2:07 AM EST
Hello,
Thanks for the feedback. Here's my responses:
1. Lou, Thanks. Good points. Like you, I'm troubled by the fab capacity trends. The foundries will soon control the (semi) universe. They will dictate the terms of chip production. In terms of VCs, agreed. I am not a big fan of VCs right now. Investing in semis seems to be a four-letter word.
Do you agree or disagree?
2. chipchap42. Thanks. The world is flat--and so is the FPGA industry. FPGAs are a flat business going nowhere. Altera and Xilinx will plod along. Ho hum. Lattice will go under one day. Xilinx will buy Silicon Blue. The FPGA startups? They may sample, possibly in the year 2050, if we're lucky!
Tell me more about the Gavrielov rumors?
3. Roger. Thanks. I'll check out the book. I like what Microchip is doing. I'm worried about their 32-bit efforts. Atmel is giving them fits in 8-bit. Microchip is in better shape than Renesas-NEC, Atmel, etc. Any thoughts out there?
4. Regarding my predictions, someone sent me an e-mail about Rambus. Do they even matter? If so, let me know. In fact, I'm still wondering when the IP houses will make money. Anybody know?
5. Lastly, the semiconductor industry will post a revenue decline for just the sixth time in the last 25 years, with worldwide revenue totaling $226 billion in 2009, an 11.4 percent decline from 2008, according to Gartner.
Here's the 2009 sales projections for Wal-Mart alone: $409.39 billion!
Once upon a time, a semi exec claimed the IC industry would be a $500 billion industry. What happened? We're did we go wrong?
We need new growth engines? Agree or not?
Sign in to Reply
Dylan_EET
1/5/2010 3:05 PM EST
Mark- bold and insightful predictions, as always. I must dare to disagree with your assessment of the FPGA space, though. I do believe that FPGAs are benefitting from the trend of declining ASIC starts and should grow faster than the IC industry overall over the next few years. A study released by the Linley Group earlier this year projects that FPGA revenue will grow to nearly $3.5 billion in 2013 from $2.55 billion in 2009, a faster growth rate than is anticipated for the broader IC market.
The big FPGA vendors won't show growth for 2009, but who will? Few. As for the startups, they still face an uphill battle in many ways, but most if not all are bringing something innovative to the table. Some are already sampling and say they have design wins under their belt. SiliconBlue CEO Kapil Shankar told me his company had between 40 and 50 design wins (and that was back in August) and the company is promising to show at CES how its products are being used in handheld mobile consumer electronics devices. Achronix also claims a number of design wins and announced last month that its first Speedster device, SPD60, is now in volume production at TSMC. I don't have any info pertaining to chipchap's comments on the CEOs in this space, but like you I would love to hear more.
Sign in to Reply
dextercath96
1/5/2010 8:31 PM EST
I wonder if it will happen... what is your top 25 predictions?
Thanks for the forum you have posted. It is great. And it gives
me a lot of information and knowledge about this.
Sunpowerport Solar Generator for sale. $450. New. Works
great. Will run most small electrical appliances. Great for
camping or emergency backup. No noise. No fumes. For more
details see http://www.sunpowerport.com.
[url="http://www.sunpowerport.com"] Sunpowerport Solar Generator
[/url]
Sign in to Reply
Mark LaPedus
1/6/2010 6:39 PM EST
In terms of solar hype, I would not blame the technology providers. First Solar lists its panel manufacturing costs as less than $1/Wp, compared to "grid parity" of $1/Wp installed. Adding in markup and installation, maybe that is $3/Wp installed. Applied has repeatedly shown a learning curve for shallow junction silicon with $1/Wp at 100,000 MW cumulative production, which they project reaching around 2020. I am very interested in what NanoSolar's production costs are now that they are actually in production. They have the potential for very low costs. So I would say that the major solar cell vendors of all technologies have demonstrated a good learning curve, delivering on their promises, and all are projecting a while until grid parity.
But as the CTO of Applied has pointed out, there is a large market where solar is already cost competitive with existing electrical generation, and falling solar costs are rapidly growing this market.
Lower cost != higher efficiency. Increasing silicon cell efficiency will only contribute a modest amount to the learning curve. Similarly for other photovoltaic technologies. They aren't going to find a
factor of two there. If you really want efficiency, you'd go with solar thermal, which potentially can achieve 30%, and can store molten salt to generate power at night, but needs full sun, expensive
heliostats, must be large enough to amortize the balance of plant, and
doesn't seem to have much of a learning curve.
So assuming incentives stay about the same as now (except for the blip in Spain), I expect prices to fall to "parity" in around 10 years, and this projection is in line with major vendor statements.
Duncan M. (Hank) Walker
Professor and Graduate Advisor
Dept. of Computer Science and Engineering Texas A&M University
3112 TAMU
College Station TX 77843-3112
Office Tel: (979) 862-4387
Office Fax: (979) 847-8578
Office Email: walker@cse.tamu.edu
Grad Advisor Tel: 979-845-4087
Grad Advisor FAX: 979-862-3684
Grad Advisor Mail: grad-advisor@cse.tamu.edu
URL: www.cse.tamu.edu/faculty/walker
Sign in to Reply
Hillol
1/10/2010 6:10 AM EST
There is no coverage for Analog and RF Business units.
Hillol Sarkar
CEO
AgO
www.ago-inc.com
Sign in to Reply
kjdsfkjdshfkdshfvc
1/15/2010 12:54 AM EST
One thing I hope to see in my lifetime is the revealing of the true nature of Steve Appleton of Micron. He is truly a crook as the many price fixing lawsuits that MU has been party to continue to wind their way through our legal system. If RMBS can ever over come the trial by media and dirty tricks that it has endured for the past 10 years we will see a whole pile of MU execs take the 5th, and M Sadler will again be called a Lying Sack of Sh*t by a jury forewoman (it has happened before! Look it up). Clearly MU has been the ringleader in many nefarious dealings but has managed to use its considerable political influence to acquire amnesty agreements. Yeah, hopefully in my lifetime we will see what a terrible corrupt executive suite can do to a good company. It happened to Nortel and Lucent. Micron is probably next.
Sign in to Reply
JACWY
2/6/2010 5:03 PM EST
I certainly am not going to address the venomous levels of BS as proffered by someone with the handle of a spammer. However I will say that Steve Appleton has shepherded Micron through some very tough times. I admire his ability to lead this American company against intense competition frequently backed by foreign governments. If it were not for Steve Appleton the United States would not even be a factor in this business. He makes tough calls, which is what leadership is all about. Keep up the good work Mr. Appleton!
Sign in to Reply