The U.S. Government Crushes Intel Corporation’s Chinese Supercomputer Chip Sales

By Leo Sun
April 15, 2015

The U.S. government, however, recently blocked Intel from selling its new Xeon chips to China, citing concerns about their role in nuclear research and simulations. Under U.S. export regulations, technologies used in the “design, development, or fabrication” of nuclear weapons can be legally blocked. The U.S. also claims that the Chinese supercomputers could act “contrary to the national security or foreign policy interests” of the United States.

Let’s take a look at whether or not that ban was justified, and how it could impact Intel’s growth in China.

Should Intel have complied?
Intel didn’t dispute the government’s ban. However, investors should be aware that there are some double standards at play.

The U.S. Department of Energy has been testing nuclear weapons with supercomputers since 1992, the year live nuclear weapon tests were banned worldwide. The DoE boasted that Sequoia, its weapon-testing supercomputer, was the fastest in the world up until Tianhe-2 topped it in June 2013. Therefore, blocking Intel exports to China can be considered a hypocritical way to throttle Tianhe-2’s growth.

Missing out on the Chinese supercomputer upgrades will prevent Intel from receiving between $1 billion to $1.3 billion in revenue. By comparison, Intel’s Data Center business generated $14.4 billion in revenues last year.

However, new deals at home could offset those lost overseas revenues. Intel and Cray(NASDAQ: CRAY  ) recently signed a deal with the U.S. government to build the Aurora supercomputer in the Argonne National Laboratory in Illinois. That machine, which will be powered by Intel’s upcoming 10nm “Knights Hill” processors, is expected to crush Tianhe-2 with peak performance of 180 petaflops.

China might not need Intel
The ban on Intel’s exports might throttle the growth of China’s supercomputers, but it won’t cripple them.

China’s homegrown Loongson chips already power several Chinese supercomputers, including the Sunway BlueLight MPP and Dawning 6000. Beijing-based BLX IC Design designs the Loongson microprocessors, while Swiss chipmaker STMicroelectronics (NYSE: STM  ) fabricates and markets them. Loongson processors run on the MIPS instruction set — which is different from Intel and AMD‘s x86 and ARM instruction sets.

That design limits their appeal for mainstream PC or mobile OEMs, but they remain a lucrative choice for servers. Speaking to MIT Review in December 2012, MIPS Technologies’ Jennifer Bernier-Santarini stated that Loongson chips had “an increasing opportunity in servers and high-performance computing”, since the MIPS instruction set had “fewer instructions and less complexity” than x86 and ARM-based designs.

BLX’s key weakness is that its Longsoon chip designs are considerably less powerful than Intel’s. Its newest server chips are built on the 28nm process, while Intel’s new chips will likely reach 10nm — an industry first — by 2017. Unless the Chinese government invests heavily in its own foundry, it’s unlikely that Loongson chips will match Xeon-level performance anytime soon.

Another move in the technological arms race
Investors should realize that tech giants like Intel, Microsoft, Google, and Apple are being increasingly affected by the U.S. government’s foreign policies.

The exposure of PRISM already made foreign countries wary of American tech firms. In response, Apple started hosting its Chinese user data on Chinese soil. Microsoft refused tosurrender emails on an Irish server to the U.S. government. Google started laying fiber optic cables under oceans to shield users from the NSA’s prying eyes. In all of these cases, the U.S. arguably treated tech companies as extensions of the government.

With Intel, it’s no different — the U.S. is interfering with the growth of its international business to bolster its position against China in the technological arms race. Strategies like these could force powerful countries like China to invest more in homegrown tech solutions, which would adversely impact U.S. tech firms’ overseas revenues.