The failures of some are the victories of others. And in the fiasco of the non-repurchase of ARM by Nvidia, there is an actor who not only made his voice heard, but who in addition was legitimate to speak about it: doctor Hermann Hauser, who is nothing less than the co-founder of ARM. This British tech serial investor gave us an exclusive zoom-to-zoom interview.
Read also: What Nvidia’s failed ARM takeover will change
When the deal between Nvidia and ARM was formalized in mid-2020, among the few voices that had risen to denounce the 40 billion dollar operation, it was his voice that carried the most weight. If he was speaking from his seat as co-founder of Amadeus Capital Partners – a fund that invests in European technologies – it is also and above all as founder of ARM that Mr. Hauser’s word had such a special echo.
Read also: “We must save ARM”: The co-founder of the company refuses the takeover by Nvidia (Sept 2020)
Born in 1948, this Austrian citizen, who received a doctorate from the prestigious English university of Cambridge, founded a company in the United Kingdom which gave birth to the most popular processor architecture in the world. And he confides to 01net.com that “ the creation of ARM was even an exception in the field of technologies: while many ideas and concepts were budding in Europe before being industrialized in the USA, we did the opposite. The RISC architecture was invented in the United States and it was Europe that finally gave birth to the company that carried it! “, still has fun the dashing septuagenarian.
A technology and semiconductor man, who defines ARM’s success as follows: The reasons for ARM’s success are two major advantages: when we started, we didn’t have a lot of money and few employees. The first chip is arguably the only one to have been designed entirely by just two people. So we had to keep the design and instructions as simple as possible. “.
And one thing leading to another, this architecture very energy efficient” ended up being at the heart of the smallest microcontrollers, in smartphones and even in the servers of Amazon (AWS) or even Fugaku, the most powerful supercomputer on the planet. A domination that did not happen overnight. ” It took more than 20 years of hardware and software ecosystem creations for ARM to take hold and eventually threaten Intel in PCs. And now in the servers “.
ARM, a “virtuous monopoly”
With over 95% market share in smartphones, not to mention controllers, microcontrollers, etc. ARM is the dominant architecture. ” Normally, a monopoly situation is bad for business and innovation “. No longer owning shares or decision-making power in his former “baby”, Mr. Hauser speaks here as an investor – even if, he admits ” the heart also speaks a little! “.
Read also: From smartphones to supercomputers, can ARM processors dominate the world?
” ARM’s situation has nothing to do with a classic monopoly. When I participated in the separation of the computing activities of Acorn Computer and the processor architecture (initially “Acorn RISC Machine” became “Advanced RISC Machines” which gave the name ARM, ndr) in the 90s, it was because I thought that no one wanted to buy a competitor’s processor. By making ARM just a technology provider that anyone could buy, I ensured its success. “.
And it is precisely this balance that the deal Nvidia/ARM was jeopardizing.
“Nvidia had no technical reason to acquire ARM”
When Mr. Hauser is asked about the initial reasons for Nvidia’s takeover of ARM, he immediately dismisses the technical reasons. ” ARM’s business model allows anyone to develop their own chips. By using a very high-level license such as the ARM Architecture License, they could have continued to develop cutting-edge chips. Nvidia had no technical reason to acquire ARM insists Hauser.
And it is enough to look at the example of Apple to realize that his words hit the mark: without owning ownership of the ARM instruction set, Apple can, with its high-end license, develop unique chips, which are clean.
” I see only two reasons for Nvidia’s approach: by owning ARM, the company would not only have been the only user not to pay for an ARM license – while making the others pay. And mastering the development of the architecture would have enabled it to ensure that its developments would have been, for example, more favorable to its own GPUs, to the detriment of competition such as AMD or Intel. “.
Because whether it’s with its Tegra or Mellanox chips – state-of-the-art network chips found in data centers – Nvidia is already developing quality chips with ARM cores. And with success, since it is a Tegra X1 which propels the Nintendo Switch for example.
A return to the stock market with a government blocking mechanism, “the only solution for ARM”
After the failure of the takeover offer of 40 billion ARM by Nvidia, Mr. Hauser sees only one solution which is already the one envisaged by SoftBank: the reintroduction on the stock market. But on this operation, Mr. Hauser is precise: “ If SoftBank respected ARM’s neutrality, the Nvidia case shows the limits that a future takeover would entail. ARM must therefore be listed again not only in the United Kingdom, but also in the United States. But in addition, I suggest that the British government invest to the tune of, say, a billion pounds, and in return obtain a ‘golden share’ of the company which gives it the power to block future external acquisition. “, he explains.
Also see video:
Also see video:
And this not out of love for the old company he founded, but out of love for his continent. The man, who is preparing a book on technological sovereignty, is convinced that he “ we must ensure our technological independence! In the future, there will be three technological blocks: the USA, China and Europe. And it is absolutely necessary not only to develop our semiconductor production capacities, but also to ensure our sovereignty over the intellectual property that will be put into them. (like the ARM architecture, ndr). This is why I welcome the European plan EU Chips Act “, he concludes.
Assuring that ” if RISC-V is a very good architecture, it still lacks what is the strength of ARM today: a mature and rich software ecosystem. And building it takes decades “.