Intel eyes ARM after Nvidia acquisition fails
07-03-2022 | By Robin Mitchell
Now that the acquisition of ARM by Nvidia has failed, Intel has set its sights on the intellectual core developer. Why is ARM of interest to Intel, will such a deal succeed, and is ARM doomed to be a neutral company forever?
Why is Intel interested in ARM?
For the past year, the media has been obsessed with Nvidia’s attempt to acquire ARM due to the many ramifications that such a purchase could bring. While Nvidia themselves do not create processors, they produce GPUs that are essentially coprocessors, and GPUs are essential in most computing applications.
This close link between CPUs and GPUs means that Nvidia is dependent on ARM developments, and being able to steer the direction of future ARM directions in favour of Nvidia architecture would harm competitors Nvidia. Furthermore, Nvidia having access to the latest ARM technology before other customers would also have given them an unfair technological advantage.
Finally, having control of ARM would give Nvidia the power to decide who can and cannot license the core. Nvidia being a US company, would also put ARM under the control of US legislation, and this could see some countries suddenly be banned from using ARM products (such as Russia and China). This would transform ARM from a neutral CPU architecture that anyone can use to a restricted architecture.
Now Intel has indicated its interest in acquiring ARM via a consortium of other companies. Interestingly, this consortium had been established before Nvidia announced its plans to acquire ARM in September 2020. Still, it is only now that the failure of the Nvidia acquisition combined with the desire for SoftBank, the current owners of ARM, to make ARM a public company.
While Intel designs its own processors, ARM could play a significant role in Intel’s future business as Intel opens up its foundry services to outside companies. Many of these customers would likely use ARM cores in their designs, and Intel having access to such designs directly (and being the part-owner of them) would not only be able to develop ARM technology but also directly offer it to customers. This would give Intel a major competitive edge against manufacturers such as Samsung and TSMC.
Would such a deal between Intel and ARM
It is more likely for Intel to have some partial ownership via a consortium than for Nvidia to outright own ARM for the simple reason that a consortium of different companies would have to share the rights and profits of ARM. In theory, this would help prevent any single controlling interest from taking over ARM and imposing its will.
However, the idea of Intel offering ARM cores in its fabrication services raises serious concerns. The whole concept of ARM being neutral is that anyone can approach ARM, pay for a license, and then use an ARM core in their silicon design that is then manufactured. However, if a manufacturer has ownership of ARM licensing, it could refuse to sell licenses to those who do not use its manufacturing service. This could also be realised by offering discounts for using an ARM license with an Intel foundry service that would encourage customers to use Intel instead of its competitors.
For this reason alone, Intel may struggle to acquire ARM, even partly, unless it makes explicit arrangements to not integrate ARM services into its foundry services. Even then, Intel may have the discretion to refuse licenses to some customers, and Intel is a US company that would put ARM potentially under the rule of US law.
Is ARM doomed to be a neutral company off-limits to everyone?
SoftBank seems desperate to sell ARM, and despite ARMs widespread use, there are not many takers. Those who want to acquire ARM seem to have some potential hidden agenda that would benefit them.
ARM has already made it clear that they will take the company public next year. A public company whose shares are owned by a diverse range of shareholders would not be beholden to any individual shareholder. This would allow ARM to raise funds via share selling and improve its market position. Still, the strong resistance of many governments worldwide in its acquisition clearly demonstrates that no one wants ARM to be bought outright.
Currently, ARM is operating on a profit margin of just 10%, and this could continue to fall depending on market factors and the rise of RISC-V. It could turn out that in the next decade, ARM ceases to be relevant as it is replaced with other technologies, and at that point, it won’t matter who owns it.