Big Picture: Intel has ambitions to create a foundry business by making chips for other companies. This is an important strategic initiative that the company will need to recoup the huge investments it is now making in factories around the world.
Most analysts agree that the proposal doesn’t make sense until Intel catches up with its manufacturing process. And rightly so, without this, Intel Foundry Services (IFS) would not have serious competitive differentiation (yes, packaging, but this is not enough). And we’ve also warned that the company lacks customer service muscle after the entire history of the fabs team running the show.
Guest Author Jonathan Goldberg is the founder of D2D Advisory, a multifunctional consulting firm. Jonathan has developed growth strategies and alliances for mobile, networking, gaming and software companies.
But beyond that, there are even more concerns.
When we talk about the production of semi-finished products, we tend to view the various foundry assemblies as interchangeable, as if a customer could easily trade Samsung for Intel for TSMC. This is not accurate. In fact, each factory has its own way of doing production. The Intel process was developed for the production of processors, and these processes are different from those required for other types of chips. For example, mixed-signal chips that need to process digital versions of analog signals (such as cell phone signals) as well as perform pure digital logic require a completely different set of manufacturing steps—different machines calibrated differently, in a different order. To put it mildly, Intel doesn’t have much experience in manufacturing its own mixed signal parts.
And then there are the tools. The bulk of the foundry work is the software used to communicate between the factoryless customer and the foundry. It’s not as easy as sending a couple of files via email. All major foundries have invested heavily in integrating the tools they use to control production with the software tools used by their customers.
This is much of the competitive stranglehold EDA vendors like Cadence and Synopsys have on the industry. To be clear, Intel has its own set of tools to manage the production of its own chips, but they are surprisingly proprietary to Intel. From what we’ve heard, even Intel employees don’t like the experience. Does IFS expect clients to master these tools? Most likely, Intel will have to invest heavily in software in order to create an entirely new set of tools that will be convenient for customers.
This all leads to the question: if Intel can get its production back on track (the big “if”), what kind of customer would want to switch to Intel? Even if Intel can beat the TSMC process by 2025, as they claim (or will it be 2026?), it will take a significant period of time (measured in years) for non-factory customers before they feel comfortable submitting actual production orders to IFS . .
The industry is full of horror stories of companies getting into trouble because their foundry process didn’t go according to plan.
It is important to keep in mind that every time a company sends a chip to a foundry for production, that company takes on the risk. There’s a reason they call them “risk starts.” For established foundries, this risk is measured by yield, the percentage of chips produced on each wafer. But for a new process, let alone a new foundry, there is a real risk that the process may not work. The industry is full of horror stories of companies getting into trouble because their foundry process didn’t go according to plan.
Given all this, it is unlikely that any of the largest chip companies will rush to sign up for IFS. Does Apple Risk Breaking the iPhone Cycle Due to Foundry Unpreparedness? Just going from 5nm TSMC to 3nm TSMC is scary enough. If Intel can fix its process (again, a big “if”), large customers will definitely seriously consider IFS, but that serious look will entail years of meetings and deep diligence.
That being said, there is one group of customers who might be very interested in rolling the dice at IFS – the developers of RISC V chips. These companies don’t have many good options today. They can access TSMC, but they are generally very small and therefore do not receive “team” service or prices.
RISC V chips also have the advantage of being open from scratch. They are so new that no one has much experience in fine-tuning the foundry processes to produce them. Remember back in February when Intel announced they were doing big investment RISC V? At that time, we noted that Intel had no intention of switching their processors from x86 to RISC V, but they were interested in the wider ecosystem.
The investments they talked about were mostly in building the tools and processes needed to attract factory-less RISC V developers. It’s worth noting that the Intel RISC V team appears to be largely seated inside the IFS organization.
This is not a bad strategy. Under normal circumstances, we would praise the wisdom of Intel. RISC V companies can easily become a springboard for IFS, a whole group of fairly dependent
Guinea pigs clients on which they can experiment hone your customer service capabilities. The trouble is that all this is still very far away. We recently told a friend that we didn’t expect IFS to be truly operational until 2030, and he replied, “Oh, so you’re an optimist.”
Intel is on the right track, it’s just a very long way to go.
Editor’s note: Shortly after this column was written, Intel Foundry Services President Randhir Thakur told EE Times that the US Department of Defense is the “No. 1 client” of IFS under the US Department of Defense SHIP program. Intel’s choice may have some obvious political implications, but it still means Intel has yet to prove itself as a foundry taking on customer projects.