by James Montgomery, news editor
September 21, 2010 – The semiconductor industry’s legendary volatility keeps everyone on their toes, from chip suppliers to investors. And it’s the reason that a company like ATREG came into being, has been profitable for much of the past decade, and is now a spinoff from parent company Colliers to spread its wings.
| Stephen Rothrock, managing partner, ATREG |
ATREG got its start about 10-12 years ago under the umbrella of real estate firm Colliers; its first deal was selling a 200mm fab shell for Matsushita. Taking into account the costs of decommissioning and reinstalling equipment, etc., "we realized if they’d left the tools in place, we could have doubled their money," recalls Stephen Rothrock, ATREG managing principal, in an interview with SST. That deal led to referrals to Fujitsu (which did result in an entire site sale), NEC, Sony, and Komatsu, and from there IBM, Freescale, and others.
But in the past five years or so, the business has involved much more complexity normally seen in broader M&A negotiations, from IP to supply agreements. The ATREG unit has led Colliers’ businesses in growth and profits for several years, with sales north of $3B — but now fewer than 30% of its business real-estate oriented, Rothrock said, and "priced as a commodity, as on the real estate side, doesn’t do it justice." The company, newly based in Seattle, has roughly a dozen people currently involved with 10+ assignments, roughly mixed 70% dispositions (sellers) and 30% acquisitions (buyers).
Among ATREG’s recent deals:
– Qimonda’s North American 300mm line, sold to TI (the shell was later sold separately, to be used for a data center). ATREG also has been helping seek a buyer for Qimonda’s 300mm campus in Dresden.
– NXP’s 150mm integrated passives device unit in Caen, France, sold to management-led startup IPDIA.
– Atmel’s ASIC wafer manufacturing operation in Rousset, France, to Lfoundry, and its 200mm fab in Texas to Maxim.
Still on the block:
– IDT’s fab in Hillsboro, OR. Rothrock calls this deal "close to a conclusion."
– Sumco’s 200,000 sq. ft site in Cincinnati, OH. "They thought it would be an asset strip," Rothrock said, "but we brought them multiple buyers they didn’t think were there." This deal also is still open.
– A number of Freescale 150mm operations worldwide: two former sites in the UK (now "final stages"), a fab in France, and a site in Japan. The latter two are being sold as going concerns.
Other deals in ATREG’s past portfolio include projects for AMI, On Semi, LSI Logic, and NXP. And the company was involved in the original deal to sell Altis, the Infineon/IBM JV, to a Russia-connected group — a deal that "fell apart when Putin invaded Georgia," Rothrock noted.
In today’s environment where capacity is so tight, and foundry customers are being put on allocation, "we’re seeing a move back toward a desire to own [a fab], to control their destiny," Rothrock said. That’s a shift from the traditional mindset of setting up a brand-new line, installing and qualifying and calibrating equipment, and running the fab as a self-built entity.
A business like ATREG exists entirely due to the volatility of the market — up or down, Rothrock acknowledges. "Cyclicality is part and parcel of our business," he said. "Right now we’re doing more acquisitions than in the past several years." The company’s sweetspot is in deals ranging from $10M up to ~$300M; below that it’s a wash of time/resources and payback, and deals bigger than that are tricky too, typically more complex and longer to establish. Most deals take a year or two to close, and Rothrock said it’s ideal to open talks two or three years ahead of that, to help understand and guide decisions about tooling, inventory, finding extra value, etc. Competition tends to be internal corporate M&A teams of the divesting companies, who understand the complexities of their own business and technology but may not have the broader picture of the value of these assets to the market.
He also emphasized a desire to balance the backoffice analysis with keeping the operation and assets viable — the manufacturing keeps humming to keep the technology viable, and that helps keep the workplace intact and diminish millions in liabilities, he explained. "With Qimonda, we didn’t get there early enough," he said. He emphasized a desire to help solve the problem of seeing Europe and the US "stripped of this stuff, and going to Asia for pennies on the dollar." And he suggested the TI deal to buy Qimonda’s 300mm tools for its analog fab is a new trend, and suggested another trend is in more of a consortium purchase of larger-scale manufacturing sites, suggesting such a deal could be linked to Qimonda’s as-yet-unsold Dresden campus.
Rothrock projects a $300B market for such transaction advisory services targeting "cleanroom technology manufacturing," and wants ATREG to triple in size in the next 10 years. Right now about 75% of its deals are in the semiconductor area, with 15% in data centers (Colliers counts big names such as Microsoft, Cisco, Amazon, AT&T, BT, etc.) and 10% in solar; 75%-80% are disposition "advising the seller" and 20%-25% acquisition-oriented. Ultimately he’d like to balance that out to about 50% business in semiconductors and the rest split among solar, data centers, "potentially pharmaceuticals and bio," and "alternative energy" e.g. large battery manufacturing.