Back to Basics

News comes that 英特尔 has lost the design win for Sony's PlayStation 6 to AMD (yet again) . It's another sign that Intel needs to think more like a startup, less like a Wallstreet darling of decades past. As with most well-established companies, they need to get back to basics. <https://www.reuters.com/technology/how-intel-lost-sony-playstation-business-2024-09-16/>

As a recap, Intel is in a fight for relevance. Once the most important company in the world and 2+ generations ahead of everyone else in the MOST advance silicon technology, they're now 2+ generations behind. One saving grace has been the US CHIPS Act, which encourages US companies to on-shore cutting-edge fabrication and retake the lead. As the biggest silicon producer in the US, Intel benefits disproportionately from this funding (to the tune of tens of billions US$). I won't argue the validity of the CHIPS Act here- it's a policy decision, not a technical one. I've previously mentioned that Intel did not lose leadership in manufacturing due to lack of money (see my note here: https://www.dhirubhai.net/pulse/only-paranoid-jay-diamond-snqcc/?trackingId=PCQQsv5pSq6kxvKI4G0YpQ%3D%3D), but the infusion definitely gives them an advantage in a marketplace in which capital expenditure for a factory is unfathomably large, among the largest in history.

Along with that expenditure comes a necessary massive expansion in capacity.

If you don't build huge volumes as silicon dimensions shrink, you simply cannot compete profitably. That in turn mandates that you'd better scale sales and marketing efforts to move that product. Making things worse, each generation brings smaller transistors AND larger wafers, generating a multiplicative effect and a colossal sales issue.

Historically for Intel, that's been easy. As inventor and popularizer of the microprocessor, with a founder first understanding that you could quadruple capacity every two years, and with the fundamental architecture that ran the vast majority of modern computers, Intel would simply produce a next generation chip and everyone would flock to buy it. That oversimplifies things slightly, since there was a lot of work on the part of a lot of brilliant people to produce that next-gen chip, but true nevertheless. We had faith in the "Field of Dreams" sales strategy: If we build it, they will come. And they did, over and over. Until they didn't.

Andy Grove famously said "the last one commoditized wins", meaning that if you could stay at the front of the pack while all of the smaller chipmakers fight for the low-margin silicon business, you reap the vast majority of the rewards. Intel did that for 50 years until the wheels fell off. With SO MUCH CAPACITY in modern semiconductor plants and the rapid advancement of the technology, factory depreciation results in vast last-generation capacity. The latest tech is pricy, but the last generation gets very cheap, very quickly.

Intel has been frantically trying to scale their silicon foundry business to regain technology leadership, which as I've previously discussed is a cultural problem for a company notoriously obsessed with its own products. Historically they haven't had a good success supporting customers, with 苹果 being a notable exception (in their laptops until they built internal design capacity to compete). I witnessed Intel breaking long-held cultural norms to win and keep business when they were forced to meet the demands of an important customer.

Intel famously walked away from the iPhone because it wasn't profitable enough, Intel would have been forced to learn, adapt, and could have owned the vast majority of silicon in the handset market. Currently they having no share of the handheld market despite billions in investment.

Intel's foundry problem is starting to smell like old fish.

They've had the foundry product in the market for a while, but nobody is buying. Nobody is now buying because nobody has bought previously. It stinks of outdated failure.

Like an aging shopping mall, Intel needs an anchor tenant. They think it's the US government, with officials encouraging onshoring and CHIPS Act investment as great reasons for procurement. That's a solid long-term client, but one unimpressive to the private sector, which is why the PlayStation news is troubling.

Intel has long forgotten that sometimes you need to lower your margins or, gulp, LOSE money to prove that you can do the job, win business, and cleanse that rotting fish smell.

PlayStation could have been that anchor tenant, filling fabs for a decade and forcing Intel to actually please a customer. Worse than losing the business, they are sending the signal that they aren't willing to win the business, or more appropriately your business.

TSMC, ARM, and anyone playing in cutting-edge semis knows that to sell the fab slots, you also need design support. With many thousands of brilliant engineers, Intel could absolutely do that and it's a massive unstated value that Intel is ignoring.? Instead, Intel is reserving those resources so that they continue to compete on last generation's cash cow. They have not internalized the desperate need to break from the Field of Dreams sales strategy and "not invented here" stigma.

Concurrently competition is increasing their in-house design capabilities/expertise and has spent the last decade securing IP libraries to quickly deliver anything a potential customer could want. With every customer loss, Intel falls farther behind and the foundry business becomes more of an albatross.

Sony is certainly not the only possible anchor tenant for Intel Foundries, but as revenues degrade, time is running out to secure the premier customers that will save this once great company.

Interested in taking your business to the next level? I consult and coach on technology, operations, marketing, business development, and overall business strategy.

Talk to me!