This post was originally published on this site
Pat Gelsinger took the helm of Intel a year ago, and there has been no lack of ambition about the semiconductor company’s road back.
The chipmaker provided a path to process leadership. If Intel
INTC,
can execute flawlessly, something that has evaded the company over the past several years, it could return to leadership by 2025. It will be expensive, costing as much as $85 billion in capital expenditures over the next few years.
Last week at Intel’s investor day in San Francisco, I spent an hour with the CEO, which included an in-depth interview on The Making Markets Podcast, and we covered a lot of ground.
With competitor AMD’s
AMD,
market value recently passing Intel’s, it felt like a good moment to dig into the tough questions. Not that the market hasn’t been tough on Intel, because it has.
Read: Intel’s Gelsinger lays out turnaround plan, and investors are in for a long haul
Gelsinger’s presentation at the event was on point. He focused on the company’s strategy broadly and gave acute details on how it aims to meet the opportunities that it had struggled to take advantage of in recent years, such as a plan to build a $10 billion discrete graphics business (GPU) by 2026 to compete with Nvidia
NVDA,
and AMD.
While there was much for Intel investors to be encouraged about in manufacturing, process, network/edge, discrete graphics and automotive segments, the standout opportunity for me was Intel Foundry Services (IFS). The business isn’t only a short-term growth opportunity but also a necessity to the U.S. and global economy.
Intel’s foundry business is a targeted geopolitical play that has the chance to solve the damage of our broken supply chain, driven by disproportionate offshoring while bringing jobs and manufacturing leadership back to the United States.
In 2020, 77% of semiconductor manufacturing was done in Asia, led by Taiwan Semiconductor Manufacturing
TSM,
Samsung, United Microelectronics
UMC,
and Semiconductor Manufacturing International
688981,
while the U.S. and Europe accounted for 21%. The U.S. has seen its output diminish from 37% of the world’s chips in 1990 to a woeful 12% in 2020.
Intel’s “moonshot” seeks to move the balance of semiconductor manufacturing to 50/50 between Asia and the rest of the world. Intel is seeking to capitalize on what will be a trillion-plus dollar market for semiconductors by the end of the decade. Its leading-edge capabilities will be essential as that market is growing approximately six times faster than the lagging edge, which is the predominant manufacturing capability in U.S. foundries today.
This ambitious goal will be a challenge and will prove expensive, but the potential returns will be significant in both growth and economic output.
Intel will continue to aggressively add capacity with new fabs in the U.S., including its recent mega-fab announcement in Ohio. The company also intends to announce a new chip facility in Europe soon. Of course, these sites take years to build and bring online, so they will not fix any residual short-term issues.
However, they represent a part of the longer-term fix that has proven to be much needed as we head into another year of supply-and-demand discrepancies fueled by the pandemic and amplified by a lack of supply-chain resiliency. And, like most infrastructure projects, these fabs will bring economic value to local communities and our broader economic output, including an estimated 10,000 jobs in the construction and ongoing operations of the Ohio facility.
Policymakers in the U.S. who are looking to win points with their constituencies should certainly see Intel’s efforts as material. Intel, especially with its planned acquisition of Tower Semiconductor
TSEM,
has a near-complete portfolio of chips manufacturing and foundry capabilities. From lagging-edge processes to leading-edge nodes across varied architectures, including x86, Arm and RISK-V, all will be offered by IFS.
As it scales its foundry service, Intel will not only improve its manufacturing capabilities, supporting its own IDM 2.0 strategy, but should be a buffer in capacity for many of its fast-growing competitors such as Qualcomm
QCOM,
Marvell
MRVL,
Nvidia and AMD. Intel will have to prove it can win support from those companies. Still, there is little reason to think these companies wouldn’t want to take advantage of the opportunity to produce bigger volumes based on excessive demand.
Furthermore, as far as U.S.-based companies that can offer the manufacturing of semiconductors at scale, Intel sits in a great place to provide a solution. This effort should be seen as valuable for economic reasons and national security matters. It has become abundantly clear that not only supply-chain resiliency but cyber resiliency and connectivity are only going to grow as matters of national security.
With Russia attacking Ukraine, creating an energy-supply snarl, we have to keep in mind how fragile our supply chain and economy can be. The potential for increased tension between Taiwan and China could be even more harmful to semiconductors and tech. Developing and manufacturing in the U.S. and Europe must be prioritized.
And while we should be pleased to see TSMC and Samsung continuing to invest and build fabs here in the U.S, it will be good to have a U.S.-based company taking greater responsibility for meeting demand and supporting the political and economic agenda and priorities, especially in reducing the disproportionate dependence we have on Asia for semiconductors.
The road to foundry prominence will be challenging for Intel, and I would be wary of anyone saying it is a sure thing. However, it would be odd not to cheer for Intel’s foundry ambitions. Of course, there are many other U.S.-based chip manufacturers and foundries, including Micron
MU,
On Semi
ON,
Texas Instruments
TXN,
and GlobalFoundries
GFS,
(U.S.-based but incorporated offshore).
However, as it pertains to a truly comprehensive offering to compete with the likes of TSMC, it would be hard to argue there is a U.S. company that is more capable than Intel. And if the goal of the legislation, including the recent Chips Act, is to take greater control of the manufacturing of semiconductors and the broader implications this has on a bevy of industries, it would behoove leaders to continue to back Intel in its endeavors. No matter which side of the aisle you find yourself on.
Daniel Newman is the principal analyst at Futurum Research, which provides or has provided research, analysis, advising or consulting to Nvidia, Qualcomm, Microsoft and dozens of other companies. Neither he nor his firm holds any equity positions in companies cited. Follow him on Twitter @danielnewmanUV.
