Intel CEO Pat Gelsinger announced today at the company’s Intel Accelerated webcast that the company is rethinking how it releases — and brands — its semiconductor innovations. The announcement includes broad strokes of Intel’s processor roadmap for the next half-decade, new chip and packaging technologies, and a promise of a “annual cadence of innovation,” with the ultimate goal of seeing Intel reclaim its processor leadership by 2025.
Future Intel products (beginning with its upcoming 12th Gen Alder Lake chips later this year) will no longer use the nanometer-based node nomenclature that the chipmaking industry has used for years. Instead, Intel is introducing a new naming scheme that the company claims will provide a “more accurate view of process nodes across the industry” and how Intel’s products fit into that landscape.
In practice, this means that the new third-generation 10nm chips will be referred to as “Intel 7,” rather than some 10nm-based name (as were last year’s 10nm SuperFin chips).
At first glance, it appears to be a cheap marketing ploy designed to make Intel’s upcoming 10nm chips appear more competitive in comparison to AMD products, which are already on TSMC’s 7nm node, or Apple’s 5nm M1 chips. While this is technically correct, the comparison is not as unfair as it appears.
Node names in modern semiconductors no longer refer to the size of a transistor on a chip: this hasn’t been the case since 1997, thanks to advances like 3D packaging technologies and the physical realities of semiconductor design. Along with the announcement of its process roadmap, Intel also announced two major updates to its Foveros chip-stacking packaging technologies (the second generation of which will debut in Intel 4’s Meteor Lake in 2023). Foveros chip stacking combines multiple hardware elements into a single die, similar to Intel’s Lakefield chips, which combine five CPU cores, an integrated GPU, and DRAM into a compact stack to save internal space when compared to a traditional design.
Foveros Omni will increase the variety of stacked chips by making it easier to mix and match tiles of different sizes — for example, allowing a base tile that is smaller than the top tile in a stack. Furthermore, Foveros Direct will enable direct copper-to-copper bonding between components, reducing resistance and bump pitches. Both new Foveros technologies are scheduled to go into production in 2023.
Intel’s new names may help the company better contextualize its current and future products in comparison to the competition, but the fact remains that Intel is trailing. Even if the Intel 7 is on par with 7nm products from other foundries, those foundries have already moved on from 7nm to 5nm hardware. This means that companies that rely on external foundries, such as Apple, AMD, Nvidia, Qualcomm, and nearly every other major tech company, can still obtain chips that are more advanced than Intel’s best work. Apple’s superlative M1 Macs, for example, already use TSMC 5nm chips — and easily outperform Intel’s comparable products. AMD is also rumored to be working on 5nm Zen 4 processors for as early as 2022, which could pose similar concerns for Intel from its already encroaching competitor.
Even with its ambitious annual cadence for its roadmap, Intel is playing catch-up; it does not expect to catch up to the rest of the industry until Intel 20A in 2024. And it does not expect to reclaim semiconductor leadership until 2025 with Intel 18A. All of this assumes that Intel does not experience any further delays or manufacturing hiccups like the ones that slowed both its 10nm and 7nm processes (which arguably put the company in its current situation in the first place).
Despite years of setbacks, it’s clear that the resurrected Intel isn’t going down without a fight. However, whether its efforts are sufficient in the coming years will be determined in the coming years.