Intel today announced in its Q2 2020 earnings release that it has now delayed the rollout of its 7nm processors by six months from its previously scheduled release date, undoubtedly causing significant delays in company roadmaps. Intel’s press release also states that its 7nm process outputs are now twelve months behind internal company targets, meaning the company is not currently on track to produce its 7nm process in an economically viable manner. The company is now claiming that its 7nm processors won’t debut until late 2022 or early 2023.
Here’s the excerpt from Intel’s press release:“The timing of the company’s 7nm-based CPU product is about six months out of previous expectations. The main driver is the efficiency of Intel’s 7nm process, which, based on recent data, is now about twelve months behind the company’s internal target. “During the call for results, Intel CEO Bob Swan said the company had identified a “fault mode” in its 7nm process that caused performance degradation issues. As a result, Intel invested in “contingency plans,” which Swan later defined as including the use of third-party foundries. The company will also use external third-party foundries for its upcoming 7nm Ponte Vecchio GPUs, the company’s first graphics chips. Swan noted that GPUs would arrive in late 2021 or early 2022, suggesting a delay beyond the initial timeline for a 2021 launch of the Aurora exascale supercomputer. Ponte Vecchio is a chip-based design, and Swan clarified that the production of some of the chiplets (tiles) will be outsourced to third parties.
Intel’s first 7nm server processors (Granite Rapids) will arrive in 2023, which is later than indicated in previous roadmaps that called for a 2022 launch. This timeline is of concern given the continued execution of ‘AMD with its EPYC data center chips – AMD roadmaps outline its 5nm Genoa processors will hit the market before the end of 2022. Swan also said that Intel’s first 7nm processors will debut in the client market, i.e. chips targeting either desktop or laptop computers. Intel’s first 10nm desktop processors, Alder Lake, will arrive in the second half of 2021.
For perspective, rival foundry TSMC plans to be on the 3nm node around the same time as Intel’s new schedule for 7nm. Intel is clearly not happy with its performance on the 7nm node, with a beleaguered Swan commenting that “I’m not happy, I’m not happy, with our 7nm performance” at the end of the day. ‘call after murderous question and answer session with analysts. Swan also said, “We caused the root [7nm] problem and think there are no fundamental barriers ”and that the company would provide further updates on an upcoming Architecture Day.
Swan said the company has a buffer built into its roadmap to account for process node delays. This hosting is the result of hard-learned lessons from the company’s relentless 10nm delays. Intel says it will use its advanced packaging technologies, which allow it to mix and match components produced from external sources with its own chips, to help balance the six-month lead time of its 7nm processors with the one-year delay in its internal 7nm performance projections. . In the past, Intel has said that it will also allow newer architectures to be portable to older nodes, so it’s plausible that Intel could resort to backporting certain architectures as part of its plan. emergency.
The 7nm lag reflects another setback, as Intel still struggles to overcome multi-year performance issues with its 10nm process. These delays have allowed rivals, such as AMD, to wrest the leading position of the process node from Intel for the first time in the company’s history. This sparked a price war in the market as Intel battles a real x86 competitor with a better node, not to mention Amazon’s new Graviton 2 ARM chips based on TSMC’s 7nm node. Apple also recently announced that it is switching from Intel chips to its own ARM-based 7nm silicon. The 7nm delay also exacerbates recent news that rock star chip architect Jim Keller, who was a key part of a team effort to revitalize the company’s roadmaps, has left the company. business.
Intel CFO George Davis previously indicated that the company’s process technology will lag its competitors until the 7nm arrives in 2021, and the company will take the lead with its process. 5 nm at an undefined time:
“So we’re bringing a lot of capabilities to our customers, in addition to the processor, and we feel like we’re starting to see the acceleration on the process side that we’ve been talking about to get back to parity in the 7nm and find the leadership in the 5nm generation. ”
This plan to regain a competitive position has now obviously changed due to the 7nm delay. Intel had planned to speed up delivery of its 7nm node to compensate for the underperforming 10nm, which it said would not perform as well as other nodes. At the time, Davis noted that the company was trying to be clear with investors about the impacts of 10nm on the company’s gross margins: “… but the point is, I wanted to be clear on what was going on. during the 10nm generation. in fact, this knot will not be as strong as what people expect from 14nm or what they will see in 7nm. ”
Regarding the 10nm node, Davis commented, “As we said on our analyst day on May 19: listen, this won’t just be the best node Intel has ever had. It will be less productive than 14nm, less productive than 22nm, but we are excited about the improvements we are seeing and plan to start the 7nm period with a much better performance profile compared to that at the end of 2021. . ”
“In addition, we were at a time when, to regain process leadership, we had to accelerate the overlap between 10 nm, 7 nm, then 7 nm and 5 nm, therefore the cost that you absorb, from 2021 in particular, You got that intersection of 10nm performance, the investment in 7nm, and were also well on the way to starting the investment in 5nm: all of these combine to have an impact on gross margin. ”
Today Intel announced plans to increase its 10nm chip shipments by 20% from its previous forecast, so it looks like the company’s 10nm plans have changed out of necessity. Intel’s new plan focuses on getting another ‘full node’ of performance from its current 10nm node, which means 10nm could have longer legs than the company expected. when it announced last year that it would ramp up production to 7nm. Intel has achieved a similar feat with its 14nm processors through a series of “+” revisions that added incremental performance improvements, so it has a history of successful inter-node improvements that could help it stay competitive until ‘so that he can correct the issues with his 7nm process.
Intel has also traditionally used third-party factories, currently at around 20% of its production, for low-margin, processor-less products built on trailing edge nodes. Intel’s new plans to more aggressively exploit external fabs could cause it to use other fabs for its core logic, like processors and GPUs, which the company has not done in the past. past. As Swan noted, this will present challenges in maintaining attractive ASPs for Intel products, especially given the scale of its production needs. Ultimately, Intel could also face significantly squeezed margins if it outsources a significant portion of its manufacturing of high-margin products, like processors, to third parties. Relying on an external supplier for advanced node production also carries more risk in terms of supply assurance, as Intel could be forced to compete with competing semiconductor companies, such as Apple, Nvidia and AMD. , among other things, for production capacity.