Intel is in a very bad position and they have to admit it.
Editor’s opinion: After 30 years of dominance, the industry has come to view Intel as a giant in trouble. We don’t think this is the right way to look at the company, and it creates mental blind spots that prevent us from assessing what the right next steps are for the company. But it’s true, Intel is in a very bad position. This must be recognized, especially internally. We’re not predicting doomsday, but we think it’s time to recognize that Intel will never be the power it once was and probably won’t be for a long time.
In fact, Intel is not an industry giant. Intel’s total share of manufacturing capacity is around 10%, it’s not a giant who stumbled, it’s a niche player and has been around for years. True, they occupy an expensive, high-price niche, but nevertheless it is a niche.
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.
The best analogy we can come up with here is cars. Mercedes sells about 10% of cars in the US, while Intel owns about 10% of the industry’s capacity. Now imagine if Mercedes somehow lost its brand – maybe a massive recall or a series of high-profile accidents caused by vehicles. Not only will they lose market share, but the entire value of their brand, setting off a long-term downward trend in sales that will be very costly to exit. Intel is a luxury semi-finished products brand, and suddenly their machines don’t move fast. We’ve tormented this analogy enough, the point is that Intel really isn’t in the strategic position we all thought it was.
Intel is a luxury semi-finished products brand, and suddenly their machines don’t move fast.
After their latest set of results, especially their 2023 prediction, we’re increasingly leaning towards the view that Intel has no options. They predict they will burn $15 billion in cash next year, which is a huge amount even for a company with $34 billion in net cash on its balance sheet.
After them catastrophic roadmap event Last month, we were forced to question the company’s ability to accurately predict its business. In fact, we have many more examples of systematic errors in their predictive abilities, but none of them are as public as this event. So we have little confidence in the company’s $15 billion forecast, it could easily be much higher. Add to that the need to continue to fund their production needs and their cash needs are huge.
It is also unclear whether 2024 will be a better year. Deep down we have always argued that the company has one task, and this is an existential task – it must catch up in production. The earliest they predict to achieve this result is the end of 2024, which means that it will likely not be included in the results until 2025. By then, the company’s bank balances will be dangerously low.
Moreover, if Intel can somehow achieve process parity in 2025, it will still have to rebuild its business. This leads to obvious questions about Intel Foundry Services (IFS). The only way Intel can ever get a bigger share of the industry’s capacity is for IFS to start a real business and poach some of the big customers away from TSMC.
We have argued privately, and now publicly, that IFS will not be able to become a real business before the end of the decade, there is too much work to be done in the first place. Even Intel’s existing processor business should be treated with some caution. By the time Intel can become truly competitive in the processor market, the processor market will be very different. As recently as five years ago, ~90% of data center silicon spending was on processors (and memory), and that figure is already declining. And while the growth of the cloud will likely mean that overall CPUs will continue to grow, they will no longer play a key role in the data center. This is a massive shift in the market. Add to that all the domestic silicon and Arm processors entering the market, and it’s clear that the post-2025 CPU market won’t be as exciting as it used to be.
And while Intel is fighting for its life, the rest of the industry is moving forward – all of its peers are taking big steps to adapt to the world of custom chips and heterogeneous computing. With the latest round of cuts, Intel will be far behind in serving these markets. Intel has now exited most of the networking and memory markets, abandoned most of its RISC V efforts, abandoned its Mobile Eye automotive ambitions, and will likely retire FPGAs soon. If he catches up with production, the company will mainly produce semi-finished products from a single product.
What is most frustrating is that they do not have a clear alternative course.
Many people think that Intel should be divided into two parts: an engineering company and a foundry company, as AMD and GlobalFoundries did ten years ago. We see the logic in we spoke out for that in the past. We assume that Street, as well as some members of the Intel board of directors, strongly support this approach. But there are indeed problems with this.
First, it took AMD and GloFo almost the entire next decade to stabilize and return to functionality. Critically, there is a very real problem of how to finance the factories. GloFo abandoned advanced manufacturing processes many years ago. Would a standalone IFS do the same? They start life with just one client, the design side of Intel, and that client depends on advanced processes. The amount of money Intel needs to catch up is huge and seems incalculable at this point – $50 billion? It is difficult to imagine a structure that would attract investors to participate in this financing. The size and seemingly bottomless nature of the obligation is too large even for the largest private equity funds.
This probably means that the only viable alternative is the one they are currently trying to use: raise as much money as possible, ask the government for more, ignore the Street and run like hell to the Intel 20A.
In all of this, we think it’s important to assign blame very carefully. The current executive team is not the root cause of the problem and probably has the only viable solution. It’s a decade of bad decisions that brought Intel here. At the same time, we believe that it is time for the leadership to acknowledge the depth of its plight. Dividend cuts, obviously. Going further, Pat Gelsinger shouldn’t have taken a 25% cut in his pay, he should have taken a $1 cut and a big bunch of expensive stock options. And for them to earn money, they need to change the narrative, throw off the leadership further, kitchen wash it. Then start handing out milestones that matter. Set up trackable metrics for manufacturing—not just a roadmap, but R&D and production milestones—show the street what needs to be done, and then report back on those accomplishments. We have passed the point of half-measures and bizarrely optimistic TAM forecasts.
The industry needs to recognize this, and most importantly, the company needs to internalize it. There have been too many public comments from Intel middle managers who strong smell of denial or forgetfulness. Intel is too big to completely disappear, but in a few years it will turn into a completely different company, and it will probably have new owners. We’re not saying doomsday is coming, but we think it’s time to face the fact that Intel will never be the power it once was, and probably won’t be for a long time.
Head credit: Eric McLean