New models to increase capital effectiveness

03/01/1998

New models to increase capital effectiveness

Mary Swedberg

Cell teams show results

With $2 billion fabs just over the horizon, it`s time we took a closer look at our capital effectiveness. Can we afford expenditures on this scale? The new markets that we hope to create certainly cannot afford it.

Today`s modern fabs are full of expensive, complex equipment running 24 hours/day, 7 days/week. Yet, even with working time maximized, we achieve no more than 50% overall capital effectiveness from efficient fabs.

We have planned downtime and unplanned downtime. We have setup time, qualification time, and requalification time. We have marketing throughput rates and real throughput rates. We have mistakes, uneven work flows, unbalanced lines, test wafer runs, incomplete loads, and engineering hold.

Equipment efficiency seems to be inversely proportional to complexity, and today`s equipment is very, very complex. However, with costs now accelerating, 50% effectiveness should no longer be acceptable. Why haven`t we addressed this before? We haven`t needed to. In fact, you could say we`re victims of our own technological brilliance.

This relentless march of technology has driven our production. Until now, manufacturing and technology - the "cleanroom" guys - have been the industry`s driving force. With this sort of success, the guys in the boardroom have been very happy to come along for the ride (even if that ride is sometimes bumpy).

Costs vs. revenue

Today`s $1 billion, 200-mm wafer fab has depreciation costs of $200 million/year and running costs of double that, equalling a total annual "cost" of $600 million. It produces around 360,000 wafers. Each wafer needs to generate a revenue of $3300, giving a total annual revenue of $1.2 billion simply to achieve a 50% margin (either gross, net, or manufacturing). The good news is that the industry currently achieves that.

When you look at a $1.5 billion fab, the annual depreciation and running costs rise to $900 million/year. The revenue/ wafer has to rise to $5000/wafer to achieve the necessary 50% margin, which means a total revenue of $1.8 billion/fab. It may remain possible for complex, high-margin microprocessors, but for other products it`s doubtful.

Companies need to adapt and bring in new business models to address the core challenge facing us - escalating costs. This can only really be addressed by increasing capital effectiveness.

We need to look at ways of making the current fab-based system more efficient. Though we have started to do this, we still have a long way to go and both semiconductor companies and equipment suppliers need to be involved.

Cleanroom to boardroom

There has always been a balance between cleanroom and boardroom, and as capital sums increase dramatically, the cleanroom`s demand for more investment has to be balanced by the boardroom`s need for risk management and financial returns. How do we do this?

It`s a team effort, but I believe the biggest area for potential improvement lies with equipment (e.g., upgradeable, modular designs). A new tool currently makes the old one obsolete, which makes upgrading a fab very expensive.

Integrated fab design will be equally important. Today, our fabs are equipped with tools developed in isolation for individual processes. Each equipment company is striving to achieve the maximum effectiveness for that individual tool, yet none of these tools work in isolation. We don`t actually want to maximize the productivity of any individual piece of equipment; we want to maximize the whole fab`s productivity. Equipment is integrated into a complete production line, where productivity is determined by the slowest or least efficient tool in that line. We all know the problem of expensive photo tools being limited by tracks - or is it vice versa today?

The semiconductor and equipment industries need to get together to develop an overall solution. A highly efficient integrated system will prove to be the winner. Larger companies may try to do this alone, but wise companies will form alliances. This is beginning to happen, but needs to move faster.

The following are examples of what the combination of our two industries might do. If equipment suppliers must change, then the semiconductor industry must also change to capitalize on these developments.

Modular fab designs

How about a modular fab design as an alternative to the traditional ballroom layout? Designed to simplify goods flow with dedicated process modules, its structure is highly suited to our industry. The simplified layout means less complex, therefore less expensive automation systems. The basic design consists of seven modules, where a central area (designed for distribution and general functions) is surrounded by the process modules.

For our next generation of fabs, this could mean instead of building and commissioning the whole $2 billion fab in one go, modules could be built and started as required. Production could start as soon as the first module is completed, even though the whole fab complex may eventually consist of many sets of modules, installed as demand requires.

To my mind, the beauty of this scheme is that these modular fabs can be as large or small as required. They can be based either on individual process steps, as shown, or they can be configured as minifabs. Whatever route we choose, for real progress, suppliers and manufacturers need to cooperate even more closely.

Buy process, not equipment

Philips Semiconductors is working with several of its suppliers to develop a new business model, which envisages buying the specialized step, rather than the actual tool, from the equipment supplier. So a stepper-and-track combination would sell the required number of good alignments/day.

The equipment is installed in the fab as usual, but is run and maintained by a partnership of suppliers and the semiconductor company. This leaves the semiconductor manufacturer free to concentrate on the area she needs to be good at - bringing systems on silicon to market fast and cost-effectively.

This new approach has numerous benefits for both parties. For the semiconductor company, it ensures maximum capital effectiveness, as highly complex equipment is run at optimum efficiency from day one and through succeeding generations of process technology. The equipment companies benefit because it opens up new business horizons and strengthens the bond with the semiconductor company. True partnership.

Unless we do something to combat the threat of spiralling costs, we could end up with our expenditures exceeding our income. Together, let us take up and solve the challenge of doubling capital effectiveness. Then we can continue our fantastic global success story.

Mary Swedberg is plant manager at Eaton Corporation`s Semiconductor Equipment Division; ph 508/921-0750, fax 508/927-3652.

Stuart McIntosh is COO at Philips Semiconductors, Postbus 90050, Eindhoven, The Netherlands 5600PB; ph 31/40-272-3076; fax 31/40-272-2003.