Sept. 29, 2008 – The annual fall gathering at the Woburn, MA Hilton for an industry panel discussion hosted by Wall Street watchers took a sideroad even before it started. One of the scheduled participants, CJ Muse of meltdown-in-progress Lehman Brothers, was understandably absent [edit: phrasing changed to moderate tone. –J.M.]. Another analyst, Theodore O’Neill of Kaufman Brothers, didn’t use any slides in briefly overviewing photovoltaics. What resulted was a more open discussion forum with attendees and some energetic back-and-forth about the tense relationship between the investment world and the semiconductor industry.
Panel moderator Robert Halliday, CFO of Varian Semi. Equip. Assoc., evoked three themes resonating in the sector right now: what will happen to semiconductor capital spending in 2009, and what are chipmakers doing differently now and in the future; and the shaky state of financial markets. Some big important chipmakers simply aren’t buying tools, he noted; 2008 capex is likely to be down -35% or more, and even foundries, who also shave been spending less, are said to be feeling increasingly conservative lately. Balance sheets are “ugly,” he said, showing how Taiwan DRAM firms’ opex is now a miserable -50%, and has been in the red for more than a year. Projections into 2009 are starting to erode as well, he noted, citing a quartet of financial analysts expecting flat or slightly down capex.
However, Halliday noted he’s “reasonably optimistic” bout the longer-term picture. Samsung still plans to invest in solid-state storage devices, with a late 2010 timeframe, for example.
Taiwan DRAM companies’ estimated OpEx %, showing large losses and negative cash. (Source: VSEA)
Asking attendees to not “shoot the messenger,” Needham & Co. analyst Vernon Essi noted there’s now limited capital for “cavalier” investments fueled by cheap money and high risk/reward, and consolidation has “growth-managed” companies such as Intel, TSMC, and Samsung to spend more consistently. Capital equipment buyers want more for less (an “ominous trend,” he warned), and are pursuing lower-cost infrastructure, notably in Taiwan now that trade restrictions with China are being relaxed. The good news, from a financial perspective, is that semiconductor-industry companies “are built with hard-to-find money,” he said. The bad news: that money doesn’t come easy, and there are offshore economies where it does to others.
Essi urged attendees to prepare for what will be a “downslope of ingenuity” and a new “Dark Age of innovation” in the chip industry, where exit strategies for entrepreneurship favor aggregation and the IPO climate has dried up. Not only has Moore’s Law failed, he said, but so has scaling across most technology fronts (software, comm, and hardware), and companies will be investing a lot more on trailing technologies. Safe harbors to seek in such stormy waters, Essi suggested, include power analog, multi-load ICs using SOI approaches and BCD implementation. Renewables are “the new semi drug,” he cautioned, and RF is “not fully around the innovation curve yet,” but there are future opportunities in mixed-mode functionality. He also urged chip companies to “chum up” to OEM customers for one-off R&D projects. OEM firms are one side of the vice that’s squeezing out chip firms (Asian design/dev work is the other) — Analog firms, for example (e.g., Broadcom, Marvell, TI) are being pressed by Apple and IC designers on one end, and by Mediatek and TSMC on the other. “Their margins have nowhere to go but down. There’s no backstop to it,” he said.
Q&A: Pride in the US, semi sector
Much of the morning’s discussion was in a back-and-forth with audience members about domestic policies that impact the US financial and semiconductor sectors. All three panelists agreed that there probably won’t be a new leading-edge fab in the US except maybe as a 450mm collaboration (while Asia and even Europe continue to woo more business). But there is still great opportunity in shared R&D, noted Halliday, pointing to collaborative work being done by the “increasingly powerful” IBM-led group in Albany, NY, as well as IMEC in Europe and Japan’s Selete.
Another touchy issue is intellectual property, a “myopic” area where some overseas companies (and regions) view IP differently, said Essi. This area has become important enough that many IC companies now break out litigation expenses as a line-item, he pointed out. “The deck is stacked against the US,” when an engineer in China costs one-tenth that of a US worker, he said. O’Neill added that a more worrisome problem is what’s being manufactured in the US, and Essi reiterated the Catch-22 that many US firms have already invested in overseas manufacturing and that re-imports are “huge to consumer America.” And some firms, he lamented, are taking a disturbing next step to actually list on those foreign markets. A key problem, he said, is that the lobbying for the semiconductor industry is lousy — though it should be noted that SEMI and SIA and others have been vocal about the need for friendly legislation, e.g. federal funding for sciences and R&D, and tax issues.
All three panelists agreed that China seems interested in building up a domestic chip equipment supplier base, but there seems to be no reliable timeframe. Essi projected that 2nd-tier tool vendors likely will emerge first but only if they can be sufficiently profitable. Halliday reminded the audience, though, that chipmakers still recognize value in tools that help them make chips faster and with better efficiency — but emphasized that suppliers also take on significant pressure to not undermine a $3B fab investment with a $100M tool that screws up yields.
Market maturation: Textiles to solar
A question from the audience about promising new growth markets for chip firms (and their suppliers) drew both positive and negative assessment from both the panel and other attendees. LEDs, for example, are very silicon-intensive, with “no shortage of technology behind the scenes,” noted Essi, but quickly added that Philips has been working in this area for years with little to show for commercialization. And while there would seem to be a lot of opportunities in architecture and commercial lighting for LEDs, innovation will come in vertical areas such as FPD TV backlighting to replace CCFL lamps. O’Neill noted that at $100/lamp there’s a six-year payback for LEDs vs. a 60W bulb, and that 5B light bulbs can be replaced in the US alone. An audience member suggested that in seven years the LED market could be bigger than DRAM — but another attendee replied that LED devices also depend semiconductor-supported systems (e.g. dimmers), and pointed out that LED poster-child Philips’ sales are dwarfed by Applied Materials.
And what of today’s other emerging-market darling, photovoltaics? Halliday noted similarities to the chip industry ~20 years ago (e.g., use of technologies like diffusion), but long-term efficiency is a bigger concern than cost. O’Neill, revisiting political themes, said there are no plans to put an AMAT Sunfab line in the US now, though that will change if a “credible” US energy policy is adopted (he chided that there’s a $2k cap on refunds for residential solar installations, but no cap on commercial ones).
One attendee likened the semiconductor industry’s current environment to where the textile industry was 100 years ago — a comparison met with mumbles and snickers from the audience, which knows firsthand how New England’s former mill towns are wrestling with the economic and physical burden of those aged looming brick warehouses. Instead of textiles, Essi suggested we’re in more like what the UK found itself in post-WWII, he suggested — and reminded the audience that this required 20 years of rebuilding.
Q&A: Brain drain, and keeping the faith
The issue of skilled workers also came up; several attendees noted how the semiconductor industry is “graying,” and Halliday agreed that “this is bigger than many other issues.” Will a fresh-faced EE be drawn to a first job maintaining a chipmaking tool at a customer site, or pursuing a hot high-growth field like solar that can fulfill not only monetary goals but also feed the desire to create and innovate? Essi also blamed the US for not sowing the seeds to cultivate EEs, noting that one domestic firm is actually cultivating them in Vietnam instead. And this industry that needs help in lobbying the government hasn’t helped its own cause with lousy branding and recruitment, he added, specifically slamming Intel’s amusing but inane bunny-suit ads of the 1990s.
All the negativity raised the hackles of longtime industry figure Bill Tobey, now consultant and SST Editorial Advisory Board member, who stood up and took Essi to the proverbial woodshed for being “so negative about semiconductors” though chip technologies and devices help underpin the darling emerging growth markets in solar, bio, etc. “You [meaning Wall Street] have no faith in what we can do,” he admonished, yet “you’re investing in industries that depend on what we’ve done, and are doing!” Essi accepted the criticism, but replied that despite some top-down opportunities (e.g. in solar) there simply isn’t the same promise of profitability and success as there was in analog in the 1980s. “Where is the innovation long-term?” he asked. “Solar is not the Holy Grail for you guys.”
Halliday urged attendees to evolve and above all be positive, since this industry has a higher barrier to entry than most (e.g., complex tools, service, IP). “We’re not where we were 15 years ago” with 16% CAGR, he said, but there’s still growth at a small clip — and that’s better than quite a few other struggling industries can say. — J.M.