It's time to plan for the upturn

Jim Morgan has an excellent view into the global semiconductor industry. As head of Applied Materials, the largest supplier of chip-making equipment in the world, his company's machines are used somewhere along the manufacturing process of virtually every chip produced.

This provides Mr Morgan, who has been leading the company since 1977, with an insight into the production plans and processes of all chip companies.

Applied Materials has close relationships with all the large chipmakers, working with them on research and development, and sharing technology road maps and production plans. Although the global semiconductor market seems to be experiencing a fragile recovery from its 2001 downturn, long-term prospects for the industry are very good. "We are working on opening the information age for the next billion people," Mr Morgan says.

As increasing numbers of people around the world acquire PCs, cell phones and other electronic products, and these devices become linked into a global information network through the internet, there will explosive demand for chips of all kinds.

But this future scenario faces an obstacle that is slowing the expansion of the information age: the lack of broadband connections between consumers and the internet.

It is a topic that Mr Morgan often speaks about in public appearances. "Broadband is probably the single most significant technology capable of expanding the chip market, and in turn, the broader technology markets," he says. "Greater availability of broadband would certainly stimulate the overall economy."

Broadband access would create the ability for large data applications such as streaming video to be delivered to consumers and business users. This, in turn, would drive increased purchases of communications equipment and servers to increase communications network capacity, and also increase the number of devices capable of accessing data through wired and wireless connections.

But broadband deployment is moving ahead at a painfully slow pace. And Mr Morgan believes that governments must become involved in helping to build this vital infrastructure because of the many political issues surrounding its deployment. However, political processes are slow by nature, which likely means there will be no fast solution.

While the chip industry waits for the extra demand that broadband will bring, there is plenty of work to be done in transitioning to smaller chip geometries, using larger 300mm wafers, and using new types of chip making materials that are essential in manufacturing complex advanced chips. Applied is very closely involved in all of these technology issues, through research and development, and working with customers. "The difference between the recent downturn in business and previous down cycles is that we are in the middle of a technology transition," he says.

But this technology transition has been slowed by the downturn in chip markets. At Applied, Mr Morgan's philosophy is that "the best time to build a company is in tough times, because when times are good, you are running around trying to keep up with demand."

But not all chipmakers are following this strategy - and Mr Morgan notes that there has been substantial under-investment in production capacity in virtually all sectors of the chip industry. This is especially true in building state-of-the-art "fabs" (fabrication plants), where capacity is very tight and which make up only 5 per cent of total global production capacity.

When chip markets turn around, there will be a rise in demand for chip-making equipment, but it takes at least six months to two years to add production capacity, he says. By that time, the highly cyclical industry might be heading into another downturn and chipmakers could be saddled with expanded production capacity but lower demand.

Key trends

There are several key trends emerging, says Mr Morgan. One of these is the outsourcing of certain steps in the chip production process. As chips - and the chip production process - become more complex, "customers are seeking to outsource some of this manufacturing complexity". This takes the form of Applied and others, offering services for the operation and maintenance of complex chip-making equipment.

Hundreds of Applied's employees work in customers' fabs, operating the equipment. This also helps chip equipment makers generate additional revenues through the sale of the equipment, and then the revenue from services sold around that equipment.

Applied has been a pioneer of several chip production processes. Its latest venture is the introduction of manufacturing process "modules" which combine two or more chip-making machines that are integrated together and come with guaranteed output. This helps customers enhance production levels more rapidly-essential when chip fabrication plants cost more than $2bn apiece.

Looking around the world, the regional market with the brightest promise is China. "There are tremendous opportunities there," says Mr Morgan. Only about 25 per cent of the chips used in China are manufactured inside the country. And there are several large foundry projects underway to increase domestic production. Applied and its rivals are hoping for a strong market in China as the country builds state-of-the-art foundries to compete with those in Taiwan and Singapore.

Applied has spent more than 18 years establishing relations with Chinese universities and building a large presence in China.

As fabs rise in expense, there is a change in how the industry organises its chip production. Securing large amounts of capital is one problem. The other is building, equipping, ramping and operating the fabs to ensure that the production lines are full and that yields are high.

Fabs are the most complex factories of our industrial age and fewer and fewer companies have all the skills necessary - which is why foundries, where chips are made for other companies under contract, are a growing sector. "Foundries and joint ventures between chip-makers are ways to fill the capacity of fabs and share costs, knowledge and production experience," says Mr Morgan.

Although fabs are increasingly costly, Mr Morgan points out that on a cost per function basis, rather than cost-per-wafer comparison, the trend is down, not up. "We are working hard to reduce the cost-per-function by harnessing Moore's law." For example, by moving to larger wafers and smaller geometries, chip-makers lower the cost of making each chip because they can produce more on each wafer.

The chip industry is also preparing for the next technological step, "nano-chips". These are semiconductors that have feature sizes of 100 nanometers or below, or twice as small as today's chips. Making nano-chips will require new types of production processes but it will also enable the production of new generations of advanced chips.

While Mr Morgan recognises that there are short-term challenges, he is confident of the long-term growth prospects for the chip industry - and in Applied's vital role in enabling the continued expansion of the worldwide information age.