"Arm will have a bright future. As Arm's customers, we will continue to support them for decades to come. When Nvidia founder and CEO Jen-Hsun Huang said these words, he was commenting on the February 8 joint statement between Nvidia and Japan's SoftBank Group: Nvidia terminated its acquisition of Arm. But one might recall that when the acquisition was first revealed, Mr. Huang called it "the only opportunity of his life" on a conference call.
When this opportunity disappears and this megalomanacy of the century in the semiconductor industry ends in failure, will the semiconductor industry be more widely affected? Compared with NVIDIA, which has a market value of more than $600 billion, what kind of way will Arm, which is preparing to independently impact the listing, solve its development dilemmas in revenue model, technology application and other aspects?
"Holding hands" ARM dreams shattered, Nvidia "lost love"
Until the end of January this year, Nvidia still denied abandoning the acquisition of Arm, because March 2022 is the deadline for the acquisition according to the design when the acquisition plan was first proposed in September 2020. Obviously, Nvidia chose a more decent initiative to "let go".
On September 13, 2020, Nvidia officially announced the acquisition of Arm, the world's chip design giant headquartered in Cambridge, UK, and whose parent company is Japan's SoftBank, in the form of cash and stock, at the same time about $40 billion (currently worth about $66 billion). At that time, NVIDIA CEO Jen-Hsun, SoftBank CEO Son Zhengyi, and Arm CEO Simon Segars respectively expressed positive views on the deal, and Son Zhengyi commented: "For Arm, NVIDIA is the perfect partner." ”
In the nearly 18 months that followed, however, the deal, which the parties saw as a "match made in heaven," was almost opposed around the world. First, the media continued to release concerns, and then, Microsoft, Google, Qualcomm, Apple and other technology giants that rely on ARM chip design have strongly opposed Nvidia's acquisition. Since 2021, the United Kingdom, the European Union, the United States and other countries have successively opposed:
Margrethe Vestager, executive vice-president of the European Commission, said: "If Nvidia acquires Arm, it may make it difficult for Nvidia's competitors to obtain Arm's IP authorization, resulting in distortions in the semiconductor market. Holly Vedova, director of competition at the U.S. Federal Trade Commission (FTC), said: "This acquisition will reduce Arm's innovation and unfairly trap Nvidia's competitors." "In December 2021, the FTC officially announced that it would block the acquisition through an administrative lawsuit. Andrea Coscelli, head of the UK Competition and Markets Authority, said: "We are concerned that Nvidia's control of Arm will create real problems for its competitors, limiting their access to key technologies and ultimately stifling innovation in some important and growing markets. "
In fact, in the course of this game, NVIDIA and ARM have jointly or separately provided credible documents and materials to various institutions several times, trying to convince them that Nvidia's acquisition of ARM is just an ordinary business act, and that it has no willingness to attack the industry.
Because, for NVIDIA, which occupies an absolutely favorable position in the GPU field, if it can realize the acquisition of ARM, and then complete the development of integrated systems for modern industrial needs, and improve the comprehensive performance of CPU + GPU, Nvidia will ideally achieve an unprecedented business breakthrough. This is actually because, with the development of complex and complex large-scale industries such as autonomous driving, artificial intelligence, data centers, and virtual reality, more and more segments emphasize the systematic operation of CPU + GPU - the architecture of the GPU itself is suitable for accelerating computing and rendering, but the complete system needs cpu to provide startup support, such as the DRIVE series provided by NVIDIA for automatic driving, which is a complete engineering system that integrates MULTIPLE contents such as CPU, GPU, operating system, software, etc. But Nvidia's accumulation of CPUs does not exactly match its expansion ambitions, and acquisitions become the fastest option.
Eventually, SoftBank CEO Masayoshi Son announced that Arm would go public in the United States, most likely on nasdaq, and planned to complete its IPO by March 31, 2023. This is expected to become one of the largest IPOs in the semiconductor sector, and ARM will also be subject to a $1.25 billion "break-up fee" paid by Nvidia for the failure of the transaction as an architecture licensing fee. On another level, however, ARM's positive attitude actually shows that being acquired may be better for it than operating independently.
Want to go public independently, ARM is afraid of the most "hurt"?
Although ARM co-founder Hermann Hauser said in an interview: "I think this is an absolute disaster for Cambridge, the UK and Europe... It was the last European technology company with global reach, and now it's going to sell to Americans. Nvidia's initiative to let go is also difficult to say that ARM wants "freedom".
But the real dilemma comes from ARM itself. ARM is a semiconductor intellectual property (IP) provider – since the 1990s, ARM has creatively changed its business model from producing chips to licensing IP to manufacturers in various fields, allowing them to produce the chips they need themselves – that is, only designs. The advantage is that ARM has thus become a world-class chip architecture service provider, and more than 95% of mobile phones, tablets and other mobile devices in the world are using ARM's architecture, including Apple's M series chips. The downside is that this has caused ARM to lose its initiative in the semiconductor industry, because its industrial positioning has prevented large-scale cross-industry expansion. Its revenue model also does not allow, and in the six months ended September 30, 2021, ARM generated only $1.46 billion in revenue. In 2020, ARM generated only $1.98 billion in revenue, with little growth over the past five years as its underlying market is already saturated.
Therefore, whether it is listed or not, a long-term problem in ARM's development is to cross the gap between traditional areas. This involves one of the oldest competition issues in the IT industry: the advantages and disadvantages of the ARM architecture and the X86 architecture - the basic understanding of the computer industry is that the X86 has the advantages of ecology, compatibility and strong scalability, and currently has a share of about 90% of the global CPU market. However, X86's high power consumption and patent barriers have led to its loss to ARM in mobile applications, and ARM has taken advantage of this to be safe, but it is limited by basic performance problems such as the lack of instruction sets, and it cannot achieve an advantage in large-scale device competition, at least not for now.
Therefore, the industry does not have a unified view on this issue, and even the success of Apple's M chip actually only proves that ARM chips have the ability to expand in specific situations. But in areas such as high-speed data centers, the X86 still dominates the world.
Fortunately, with software ecological problems such as analog translation solved, ARM's expansion is actually promising. After obtaining arm licensing, NVIDIA has successfully launched a data center CPU based on ARM architecture. In addition, RISC-V, an open source instruction set architecture based on the principles of Reduced Instruction Set (RISC), is on the rise – and while it is a competitor to ARM, this shows that industry dynamism is amplifying. If ARM wants more development, a fully competitive environment is better than being completely dominated by traditional architectures.
After the sale was terminated, ARM named Rene Haas as the new CEO, and he said, "I don't want to belittle RISC-V, but in the process of playing against RISC-V, we saw an opportunity for growth." ”
If this big acquisition is successful, there will inevitably be an "earthquake" in the chip industry. But even if it fails, the autonomous progress of the chip industry will not stop. Bernstein analyst Stacy Rasgon said: "Nvidia does not need to acquire ARM. "And the truth is not far off, Nvidia wants it faster and better, but it is by no means necessary."
At this point, behind the "breakup", except for the compensation obtained by SoftBank and the potential return on investment after the IPO may not be ideal, everything is back to the way it is. Nvidia will not give up expanding its current industry advantages through mergers and acquisitions, and if ARM can go public as desired, it may also be a new starting point for itself.