Intel Bets Big on Chipmaking Future

New CEO Pat Gelsinger outlines strategy for Intel manufacturing

We outlined several challenges facing Intel’s incoming CEO Pat Gelsinger when he was appointed in January (see Instant Insight: Intel Appoints Pat Gelsinger as New CEO). We talked about the opportunities and obstacles ahead for the company, and specifically, the crossroads it’s at with its manufacturing strategy. On 23 March, Mr Gelsinger set out his Integrated Design and Manufacturing 2.0 strategy to address exactly this question.

The announcement consisted of five parts:

  • Intel said its 7 nm process is “progressing well”, with designs for its 7 nm Meteor Lake CPU expected to be finished in the second quarter of 2021.
  • A new business, called Intel Foundry Services, aims to become a leading provider of semiconductor manufacturing services in the US and Europe.
  • Manufacturing will be expanded with a $20 billion investment to build two new fabs in Arizona.
  • Intel plans to enter a new research collaboration with IBM.
  • Intel Developer Forum will return in 2021, with an Intel Innovation event set to be held in San Francisco in October.

Intel’s choice of strategy doesn’t come as a surprise. Activist shareholders had put pressure on the company to divest manufacturing entirely and pursue a fabless model akin to that of Qualcomm or MediaTek, for example, but doing so would break away from Intel’s heritage. Moreover, that idea ignores the significant advantages that Intel still enjoys despite the loss of leading-edge process leadership to TSMC and Samsung. We agree with Mr Gelsinger’s comment that “Intel is the only company with depth and breadth of software, silicon, platform, packaging and process with at-scale manufacturing”.

These assets in combination remain highly valuable. Intel’s strategy revision is essentially an extension of what Mr Gelsinger’s predecessor Bob Swan started, but in a more coordinated, structured and disciplined fashion, led by someone with a background in Intel manufacturing.

Intel is adapting its model and acknowledging rivals’ strengths, yet raising its bet on areas where it leads in scale and innovation, such as its 3D packaging technology. It’s a pragmatic strategy borne of necessity. Closing the gap with TSMC on leading-edge chip manufacturing will be extremely difficult but equally, achieving parity isn’t essential if Intel can remain competitive at scale and form partnerships for leading-edge technologies where it matters.

This raises the question of the extent to which Intel plans to rely on external foundries, and its decision remains unclear. Doing so will increase flexibility to “optimize Intel’s road map for cost, performance, schedule and supply”, but such a move isn’t without risks. Manufacturing internally has been the foundation of Intel’s success and delivered advantages in scale and economics.

Using external foundries not only implies that it has ceded leadership in process technology, but also risks putting pressure on margins. In the longer term, the more Intel uses third parties, the harder it will be to run its own manufacturing at capacity. It also risks eroding the margin advantage it has over AMD and others.

This is where the Intel Foundry Services business model makes a lot of sense. By opening its manufacturing capability for customers seeking dedicated capacity in the US and Europe for x86, Arm and RISC-V architectures, Intel hopes to achieve two main things: to diversify revenue and to use outside business to ensure its fabs run at or close to capacity.

This is a shrewd move, particularly the support for Arm and RISC-V, but it will also be viewed with suspicion because Intel has made statements of being “open for business” in the past. The company will need to prove that it can alleviate the concerns of potential customers it also competes with, and successfully position its strength in manufacturing as well as the attraction of its US and European locations.

The response to the announcement was overwhelmingly positive, reflecting Mr Gelsinger’s inherent understanding of manufacturing and how Intel needs to adapt to be competitive. The company’s clear direction and commitment to invest a further $20 billion in US manufacturing, coupled with its progress in 7 nm processes, undoubtedly played a big role too. The return of Intel Developer Forum as the Intel On event will also have helped bring the Intel faithful on board.

Mr Gelsinger has set a high bar and a clear basis on which customers and investors can judge Intel’s progress with its latest transformation.