Driven by strong 30% unit growth and a moderate 8% price increase, the world chip market passed the $200 billion mark in 2000, growing at 38% over 1999. Remember, it was $50 billion in 1990, $10 billion in 1980, $2 billion in 1970, and just $1 billion in 1960. What a race!
Clearly, a new era of growth has started, propelled by the e-society, mobility, and connectivity, with digital set-top boxes (STBs), DVD, and digital TV replacing the PC motherboard as our industry drivers.
When all the regional data eventually comes in, the 2000 chip market will have expanded 43% in the Asia-Pacific region, 42%
in Japan (a nice recovery after decreasing many years), 37% in North America, and 32% in Europe.
By industry segments, communication products, consisting of digital networking and mobile ICs, lead the pack at 60% growth, while PCs, despite third- and fourth-quarter warnings from the industry, should enjoy a healthy 32% growth, to $100 billion.
The increasing use of chips in cars last year resulted in a healthy 25% gain for chip products going into the auto industry. In Europe, electronic content per car now stands at 18% of total cost-close to the mechanical-part proportion.
Driven by communications, the flash-memory market should post revenue of $10 billion in 2000. DRAMs, ASSPs, and ASICs, the core engine of the equipment market, should grow 36% to represent 33% of the entire chip market vs. 21% for MPUs and MPRs. A major change compared with five years ago.
It is clear that the U.S. GDP growth sharply decelerated during the third quarter of 2000 from previous unsustainable levels. It is also clear that the oil shock will cut consumer spending in Europe and the Asia-Pacific region in the next quarters.
But market fundamentals, such as rising productivity, low inflation, and the budget deficit, are safe in the Western world. In 2001, GDP growth in North America and Europe should be 3.1%, Japan 2.3%, and Asia-Pacific 5.2%. The bottom line is 4.3% global GDP growth for 2001-another record following the 4.9% boom of 2000.
Chip capacity in 2001 is the wild card, however. Chip industry capex in 2000 should grow at least 65%, to $55 billion, which gives a ratio to chip sales of 26% -- still within the safety zone. In fact, during the 1990-2000 time frame, four years were far below the average capital spending due to the downturns of 1990-1991 and 1997-1998. Moreover, the industry is in a technology transition from 0.25-micron to 0.18- and 0.15-micron, and this requires more spending.
Last but not least, 2000 was the year companies started spending on new fabs, and not only capacity upgrades, which happened in 1999. Since it takes at least 18 to 20 months to bring a fab to full production, we forecast no overcapacity in the near term.
The chip market works like a moon rocket. The first stage is the PC (35% of total) with an average growth trend between 15% and 17% per year. With such growth, the rocket takes off. The rocket booster is the communication segment (30% of total), which a few years ago started a 10-year cycle, driven by the Internet and the mobile society, and is contributing 25% to 30% growth per year to accelerate the space vehicle. The overall share of emerging applications such as digital consumer equipment (STB, DVD, e-appliances) is small as a percentage of the total chip market, but with 40%-plus growth, it is quite good to launch the satellites.
The chip market appears ready to blast off, but first let's check the three stages:
First, PC demand will start growing again in 2001, fueled by the new MPU and DRAM generations. Second, mobile communications, with projected sales of about 550 million digital handsets in 2001 and the strong momentum of digital networking for Internet access, could easily propel communications-IC revenue growth to between 35% and 40% year-over-year.
Finally, after one or two soft quarters of over-inventory correction, other segments, such as industrial electronics and distribution, could see growth resuming at the end of the second quarter.
On the capacity side, 2001 will be a "digestion phase" following the huge meal taken in 2000. A 10% capital increase is forecast, limiting any risk of overcapacity in 2001 and stretching out this cycle to 2002.
In 2001, we forecast 25% revenue growth for the world semiconductor market. The Asia-Pacific region should, as usual, exceed worldwide average revenue growth by 10 percentage points, fueled by organic growth and the continuous transplant of manufacturing facilities to the area. The United States should enjoy about average growth, along with Japan and Europe.
Productwise, let's assume good growth of at least 30% for
ASSPs and ASICs and up to 60% for flash.
In sum, the world semiconductor market entered a mild growth-adjustment phase during the end of 2000. Growth will resume this year, and off we go to the moon.
Jean-Philippe Dauvin is a group vice president at STMicroelectronics,Geneva.