Seen but not heard at SPIE: it's very quiet out there

The SPIE Advanced Lithography conference, officially a very technical event tightly focused on photolithography, has become one of the annual watering holes for the semiconductor process front-end-of-line community—a who’s who of lithography, handling and plumbing equipment, materials, consumables, and other such experts. As such, it is usually big and bustling. But this year, it gives the distinct impression of an industry knocked unconscious.

The exhibit floor has only seven or eight isles, leaving about half of the one ballroom at the San Jose Convention Center open for lounges, meal service, or perhaps impromptu bowling competitions. There is no competition for the seats in the attendees lounge, nor even a shortage of bandwidth on the free WiFi. Most worrying of all, the exhibitors still have plenty of the expensive Dove chocolates in the little glass jars at their booths.

It can’t be a surprise to anyone who has followed the news in recent months that the equipment industry is in severe recession. Semiconductor manufactures are running a horrifyingly low capacity utilization, and have accordingly cut their capital budgets to the point of putting time-locks on the office-supply closets. Nor should it be a big surprise that a lithography event in a country that is systematically divesting itself of semiconductor manufacturing might be declining. Knowing all this is one thing, but seeing it face-to-face is quite another.

A couple of interesting points do emerge from the quiet hall, however. One is that capital expenditures come and go, but consumables are for ever. And in times like these, a vendor who can make a good argument for switching to his materials to save operating costs is going to get a hearing.

Another interesting point is that the same recession that hamstrings the vendors of incremental improvement may enable the evangelists of real change. People who want to talk about doing things quite differently in order to make a major improvement in NRE, time-to-production, or yield may have trouble getting a heard over the din of progress when everything is going well, and companies are raking it in without taking chances. But comes a time when big companies have to seriously ration their design starts, small companies simply can’t tape out, and no one can afford the multiple spins of conventional yield learning, and all of a sudden those voices that were crying in the wilderness seem to be singing our tune.

One major beneficiary of this phenomenon may be the cluster of companies attempting to improve the results of lithography—such as Synopsys, with one of the larger booths on the exhibit floor—and companies exploring restricted-pattern physical design—such as Tela Innovations. Another example, a sleeper that may or may not be too long-term to benefit from their time having arrived, is the cluster of companies applying direct-write e-beam technology to wafer production. But more about them later.