Skip to content
【☢】 BitHardware.top ▷ Hardware, Reviews, News, Tutorials, Help post

Intel stops making its own processors: challenges and plans

Immediately after activist hedge fund Third Point asked Intel’s board to explore “strategic alternatives” such as spinning off its factories and / or ditching failed acquisitions, the company’s board of directors reacted with a swift change of CEOs. , choosing Pat Gelsinger to succeed Bob Swan.

However, in his first speech to the finance board, Gelsinger pledged Intel to remain an Integrated Design Manufacturer (IDM) and yes, that could require outsourcing some of the processor manufacturing to third-party factories in the short term. While this may not appeal to some investors, it is the decision that seems the most correct at the moment, and yet it raises many questions about how to do it successfully.

Challenges for Intel to outsource processor manufacturing

The suggestion to spin off Intel manufacturing is nothing new; As the cost of operating a factory and the development of new process technologies and new manufacturing equipment has increased over the past 20 years, most semiconductor companies have outsourced most of the manufacturing, while the rest (including Intel) has adopted a hybrid strategy leveraging some foundry resources when this made sense.

Many of Intel’s acquisitions over the years already used foundry resources, particularly TSMC. However, Intel has largely stayed true to its roots, manufacturing its x86 processors and other key products in its own factories. AMD had a similar strict strategy under founder Jerry Sanders, who went on to say that “Real men have their own factories”, but due to financial difficulties AMD eventually left its factories behind and now uses multiple foundries to manufacture its products.

Instead of spinning off its factories, Intel tried to become a foundry for other semiconductor companies, but this effort (which was a good idea in theory) ended up in failure due to a number of factors, including differences between Intel’s process, tools, and terminology. with the rest of manufacturers. That made it difficult to transition from products made in other foundries’ processes to Intel’s proprietary process, even when they offered a better price.

Additionally, Intel has always been reluctant to modify its process to suit the requirements of other customers. While the company has had extraordinary success with its exact copy Fab model for its own manufacturing purposes, what is good for an x86 processor is not necessarily good for other forms of logic chips. Even Intel’s Mobile group suffered from having a manufacturing group that, at the time, was focused on having maximum performance and it is not low power but high efficiency that is what a mobile device requires.

Intel Mobile

In 2016, Intel even announced a collaboration with rival ARM’s instruction set to support the manufacture of processors based on this architecture in Intel factories, specifically at its 10nm node. Unfortunately, Intel has had trouble increasing the production of its last two process nodes (14 and 10 nm) and there was little interest in using Intel when other foundries were already used to making ARM-based products. Intel acted as the foundry for the Altera FPGA at 14nm and eventually acquired the company.

Note that Intel’s foundry effort began at a time when the company had excess capacity, and while Intel is currently increasing its new manufacturing capacity, it has been limited for more than a year in partly due to problems increasing the efficiency of its 10nm node. In fact, they even opened the door for AMD to gain share in certain markets such as desktop PCs, while also focusing its manufacturing efforts on its highest-profit segments (servers).

Trying to spin off its factories now, when the company can sell everything it can make and has exciting new products on the market, would be a disaster. Even spinning factories during a period of overcapacity would be challenging, and we saw that when GlobalFoundries found out the hard way: it took more than a decade and several management and strategy changes for GloFo to find success with its current CEO Tom Caulfield.

It is not easy to change an entity’s business model while moving from handling one demanding customer to many demanding customers at the same time, not to mention the additional financial burden of investing in new factories and new process technology without a parent company with good margins. profit on your products.

Intel’s plan could go through buying GlobalFoundries

Everything that we have explained so far is to indicate that it is without a doubt not only a bad time, but also a great challenge for Intel to abandon the manufacture of its own processors, so taking into account the benefits and margins that it has Intel, the most logical thing (and we will be direct) would be to go in the opposite direction: Intel should acquire GlobalFoundries.

GlobalFoundries

This would give Intel instant access to increased capacity in its factories for some of the product lines that it already has outsourced and even some of its own products that can be manufactured on older nodes. Global Foundries has factories in Germany, Singapore, and the United States, and while it has moved away from competition against TSMC and Samsung in its later process nodes, it is still one of the top three semiconductor foundries in the world and the most geographically dispersed, somewhat that would give Intel logistical advantages.

Right now GlobalFoundries is once again a profitable company, and that means it can no longer be purchased at a bargain price. In either case, through the Intel acquisition you would get a management team that understands how to succeed in a foundry and staff used to working with external semiconductor customers; If the GloFo team led the integration and transition to being a foundry, Intel could focus on continuing to invest in having more capacity.

At that point, when Intel is in a good competitive position and has that extra capacity it needs, it could spin off the manufacturing group while maintaining an interest as an investor and as its largest customer. This would relieve Intel of the financial and capital overhead of being a semiconductor manufacturer, while ensuring that it has ample manufacturing capacity for the foreseeable future. Intel would still have the option of leveraging TSMC as a manufacturing partner, in addition to providing the company with dual smelter capability with what would likely be the two largest functions in the world.

TSMC-vs-GlobalFoundries

The only one against this is the Intel philosophy. Intel’s culture is not kind to outsiders, and one example of this was the acquisition of Infineon’s wireless solutions business to help the company succeed in the mobile industry, a segment in which Intel fought tooth and nail for enter and that in the end came to nothing. The newcomers were not welcomed and found it difficult to implement the changes necessary for the Intel product to be successful in the marketplace. This story also holds true for many of the other companies acquired by Intel.

The new CEO, Pat Gelsinger, does not face this reaction because most of his career was at Intel and he was a well-loved executive by many at Intel before his departure eight years ago. However, they now face the challenge of changing this philosophy, since a change in strategy could go hand in hand with a change in this culture that has dragged and hampered them for decades.