Download the latest issue of the Electronic Systems Design Engineering
and subscribe to receive future issues and the email newsletter
Published on February 27th, 2013
Intel is expanding its efforts in the foundry business. Earlier this week, Altera entered into an agreement for the future manufacturing of its FPGAs, based on Intel’s 14nm tri-gate transistor technology. Intel will provide foundry services for the FPGA giant.
Until now, Intel has been a bit player in the foundry business. Now, the chip giant is on a direct collision course in the foundry business against the likes of GlobalFoundries, Samsung, TSMC and UMC.
The move could change the landscape in the foundry business, in which Intel will likely become a much bigger player in the arena. Plus, Intel is already in production with finFETs at the 22nm node and the foundries are still stuck in the planar transistor era.
Still, there are some questions regarding Intel’s foundry efforts. First, does the chip giant have the staying power to remain in the foundry business? The foundry business is different than Intel’s current model. The foundry segment is a service business and customers are fickle and demanding.
Intel is known for entering new markets. Then, when the market fails to pay dividends, Intel has a habit of exiting various businesses. The list is countless.
In addition, IDM foundries also have a spotty track record. When the IC market is slow, the IDM foundries are aggressively pursuing business. But then, when demand is hot, they tend to allocate fab space for their own products, leaving foundry customers in limbo.
On the other hand, the current leading-edge foundries fear Intel and for good reason: It adds another big and formidable competitor to the mix. And Altera could become the first in a wave of new foundry customers for Intel.
In fact, Intel gave Altera a good deal, which could lure other customers. “Under the agreement terms, Altera gets exclusive access to 14nm tri-gate for manufacturing FPGAs for a limited period of time. Altera receives a 12-year product support commitment from Intel with last-time buy rights and substantial termination fees and Intel is restricted from acquiring any other FPGA company,” said John Vinh, an analyst with Pacific Crest Securities.
“Why Altera decided to ultimately move forward with Intel as a foundry partner? 14nm tri-gate provides Altera a way to catch up, particularly at the high end prototyping, after falling behind Xilinx, as initial production will focus on high-end FPGAs. Intel is significantly further ahead of TSMC on finFET, and a dual-sourcing strategy mitigates the risk to finFET at 16 nm with TSMC,” Vinh said.
Hans Mosesmann, an analyst with Raymond James, said: “From the current 28nm portfolio of FPGAs, Altera will migrate its high-end FPGAs to Intel 14nm and we believe its mid-to-lower end portfolio to TSMC’s 20nm. In our view, Altera was able to get what appears to be quite the sweet deal with 14nm exclusivity—(as) Xilinx is locked out--and a 12-year end-of-life support commitment from Intel.
“The move will likely come as a surprise to the street, as Altera’s relationship with TSMC had long been considered strategic while Xilinx’ was more recent and thus tactical. We do believe the semi industry generally will struggle with finFET technology as it adds ~40 manufacturing steps, and in this framework John Daane’s move may end up looking brilliant. We expect that Xilinx will eventually have access to Intel technology (after 14nm), although this remains to be seen. The impact to the financials will likely be minimal in the near-term, as Altera will likely sample 14nm in 2014 and won’t see revenue well into 2015/2016,” Mosesmann said.
“Strategically, the more formal entry into the foundry business is an implicit admission that Intel will not be able to fill the ~$10 billion fabs with x86 processors,” Mosesmann said.
Risto Puhakka, president of VLSI Research Inc., has a slightly different viewpoint. Altera gains access to leading-edge technology, but the volumes won’t necessarily fill up Intel’s fabs, Puhakka said. “The volumes are not that big,” he said. “We’re talking about expensive parts.”
Putting the news in perspective, Doug Freedman, an analyst with RBC Capital Markets, said: “On the Altera side, the company is now on -track to reach sub-20nm before competitor Xilinx. We believe the news is an incremental negative for TSMC as more companies could transition toward Intel's advanced processes.
“We believe the incremental opportunity for Intel could be ~$70 million in 2014, ~$200 million in 2015 and ~$300 million in 2016. We note that this does not even represent 1% of Intel’s estimated sales. We believe that foundry sales need to exceed $2 billion a year before having a material impact on Intel's business model as this would represent just less than 5% of sales,” Freedman said.
Mark LaPedus has covered the semiconductor industry since 1986, including five years in Asia when he was based in Taiwan. He has held senior editorial positions at Electronic News, EBN and Silicon Strategies. In Asia, he was a contributing writer for Byte Magazine. Most recently, he worked as the semiconductor editor at EE Times.