10nm Race Heats Up

The 10nm process and foundry race is heating up, as Intel announced its 10nm technology at its annual conference.

As part of the multi-pronged announcement, Intel’s foundry unit forged a major partnership with ARM. Specifically, ARM will make its physical intellectual-property (IP) available on Intel’s 10nm process. Intel, in turn, will offer the IP for foundry customers.

And on top of that, Intel Custom Foundry announced two new customers—LG Electronics and Spreadtrum.

Meanwhile, on the process technology front, Intel disclosed some details about its 10nm process, which is based on a scaled version of its finFET transistors. Intel did not disclose the specs of the technology, although the company maintains that it is still ahead of its competitors at 10nm.

In fact, TSMC and Samsung are separately readying their respective 10nm finFET processes. Samsung hopes to move into 10nm production by year’s end. Meanwhile, TSMC plans to ramp up the technology in the first quarter of 2017, with 7nm slated for 2018.

Foundry customers will soon have some new but confusing choices at those nodes. For one thing, there is some confusion between 10nm and 7nm.

“Not all 10nm technologies are the same,” said Mark Bohr, a senior fellow and director of process architecture and integration at Intel, in a recent interview. “It’s now becoming clear that what other companies call a ‘10nm’ technology will not be as dense as Intel’s 10nm technology. We expect that what others call ‘7nm’ will be close to Intel’s 10nm technology for density.”

Intel plans to stay on the traditional gate pitch scaling curve, while others tend to have what Bohr calls “looser gate pitches.” This, in turn, “gives Intel a density advantage,” Bohr said at a press event at the Intel Developer Forum (IDF) in San Francisco.

Not long ago, Intel reiterated its capital spending budget of $9.5 billion, plus or minus $500 million, in 2016. Meanwhile, TSMC raised its capital spending budget for the year. “Overall, the demand environment is good for equipment companies entering 2H ‘16, led by strong 3D NAND and 10nm logic/foundry activity” in the overall memory and foundry space, said Weston Twigg, an analyst with Pacific Crest Securities, in a recent report.

Separately, ARM and Intel Custom Foundry announced an agreement to accelerate the development of ARM’s IP on Intel’s 10nm process. The physical IP involves ARM’s POP technology. The initial POP IP will involve two future ARM Cortex-A processor cores. It involves either ARM big.LITTLE technology or stand-alone configurations. IP development is already underway and ARM physical IP will be available to support designs beginning in 2017.

The move is part of Intel’s ongoing efforts to expand its third-party EDA tools and IP in the foundry arena, according to Zane Ball, vice president in the Technology and Manufacturing Group at Intel and co-general manager of Intel Custom Foundry. “Having leading IP providers in our portfolio will accelerate ecosystem readiness while providing greater flexibility and time-to-market advantages to our customers,” Ball said in a blog posting on Intel’s site.

Related Stories
To 7nm and Beyond
GlobalFoundries’ top technologists open up on next-gen FD-SOI, the economics and challenges of 7nm, and what lies ahead.
10nm Versus 7nm
The economics and benefits of moving to the next process node are not so obvious anymore.
Optimization Challenges For 10nm And 7nm
What will it take to optimize a design at 10nm and 7nm? The problem gets harder with each new node.