Reporting by XinZhiyuan
Edit: Good Sleepy Rayan David
【New Zhiyuan Introduction】Lao Huang's big business is yellow! Due to delays in approval by multinational regulators, Nvidia announced that it was abandoning its $66 billion acquisition of Arm, a deal that Lao Huang once called "the only chance of a lifetime" and eventually aborted. SoftBank, which made $1.25 billion, immediately decided to make Arm public.
Lao Huang's long-dreamed $66 billion deal officially collapsed!
On February 8, Nvidia announced that it was formally abandoning its acquisition of Arm due to "significant regulatory challenges".
And for Arm, the biggest IPO in semiconductor history is about to begin.
Although SoftBank lost a huge sum of money, Nvidia's $1.25 billion break-up fee will become its profit in the fourth quarter.
Arm will be available by March 31, 2023.
Britain, the United States and the European Union are opposed
In September 2020, U.S. graphics chip giant Nvidia agreed to buy Arm from SoftBank in the form of $40 billion in cash and stock, which should have been the largest transaction in the history of the chip industry.
At a time when demand for semiconductor chips has soared, NVIDIA's stock price has risen sharply, and the value of this acquisition has also climbed steadily, even exceeding $80 billion at its peak.
Nvidia founder CEO Jen-Hsun Wong told the media on a conference call that the deal was his "only chance of a lifetime." He also said it would take a lot of effort and time to convince regulators, but with confidence to address these challenges.
But the deal quickly drew concerns from regulators around the world and chip-making rivals.
In August 2021, the UK completed the first phase of its investigation and issued a statement saying Nvidia's acquisition of Arm raised serious competition issues.
The CMA said the combined businesses would have the ability and motivation to harm the competitiveness of their adversaries by restricting the use of Arm's intellectual property (IP). Ultimately, this loss of competition will stifle innovation in some markets, including data centers, gaming, the Internet of Things, and self-driving cars.
In October 2021, the EUROPEAN Union announced an in-depth investigation into the deal, which is scheduled to close on March 15, 2022.
The European Commission fears that once the deal is approved, Nvidia will have the ability and motivation to restrict competitors' use of Arm's technology, negatively impacting the semiconductor industry with rising prices, reduced choice and reduced ability to innovate.
EU Antitrust Commissioner Margrethe Vestager said:
"While Arm and NVIDIA do not directly compete, Arm's intellectual property rights provide significant support for products that compete with NVIDIA chips. Our analysis suggests that NVIDIA's acquisition of Arm could result in other vendors having limited or downgraded access to Arm's intellectual property, thereby having a distorting impact on the semiconductor market."
In November 2021, Uk digital secretary Nadine Dorries asked the UK Competition and Markets Authority (CMA) to conduct a second phase of "in-depth investigation" into the transaction.
The CMA has 24 weeks to complete its investigation, which involves "competition and national security."
In December 2021, the Federal Trade Commission stepped in to block it, saying it would make Nvidia illegitimately in control of computing technology and design, leading to competitors needing to develop their own chips to fight it out.
If the deal is allowed, the combined company could kill next-generation technologies, including those used to run data centers and automotive driver assistance systems, the FTC said.
In the face of the lawsuit, Nvidia has said, "We will continue to work hard to prove that this transaction will benefit the industry and promote competition".
In addition, Nvidia has proposed remedial measures to allay the concerns of regulators. These include the establishment of a fully independent licensing entity and the licensing of Arm-based intellectual property developed by Nvidia to all companies on a non-discriminatory basis.
However, after nearly 18 months of hard work, the United States, the United Kingdom, and the European Union have not approved the deal. With no hope of passing the trial, Nvidia and SoftBank could only announce their abandonment.
For this IPO, Arm will appoint a new helmsman.
Simon Segars will step down as ARM's CEO and will be replaced by another executive at the company, Rene Haas. The latter joined Arm in 2013 and previously spent 7 years at Nvidia.
SoftBank bought Arm for $32 billion six years ago, but has struggled to drive growth in the business.
As a separate public company, Arm's life will be even tougher.
Although it has sold key intellectual property to some of its most valuable companies, including Apple, Google and Samsung Electronics, its sales figures are relatively small. In the last six months ended September 30, 2021, Arm's revenue was only $1.46 billion.
While that number is up 56 percent from a year earlier, analysts expect stronger growth when Arm releases results on Tuesday. But its size pales in comparison to the larger chip industry. According to Yahoo Finance, analysts expect Nvidia's sales this fiscal year to reach nearly $27 billion.
Moreover, at the most basic level, Arm's competitive foundation is its instruction set architecture, which defines what kind of software can run on a chip. Arm needs cash to keep up with its much larger competitors.
Is M&A failure a loss to the industry?
On December 5, 1978, physicist Hermann Hauser and engineer Christi Curry founded CPU in Cambridge, England. In 1979, the company was renamed Acorn Corporation.
At first, Acorn intended to use Motorola's 16-bit chip, but found that the chip was too slow and too expensive. They turned to Intel for the design of the 80286 chip, but they refused, so Acorn was forced to develop it on its own.
In 1985, Roger Wilson and Steve Furber designed their own first-generation 32-bit, 6M Hz processor, or ARM for short. That's where the name ARM comes from.
Nvidia's acquisition of Arm, the largest semiconductor acquisition in IT history, went bankrupt. The reason for this is that there are significant regulatory issues in the antitrust space.
In September 2020, the $40 billion deal surfaced for the first time, but there was plenty of skepticism from the start. Many of the so-called reasons circulating at that time were either untrue or exaggerated.
Before saying anything else, it's worth understanding why the deal matters.
NVIDIA's core product is the image processing unit, or GPU. In the beginning, the GPU was used on the computer, in order to improve the ability of the computer to display images, like playing computer games, the picture quality will be clearer and the experience will be better.
Coincidentally, the architecture of the GPU itself is well suited for tasks that require accelerated computing, such as real-time image rendering, virtual reality, artificial intelligence, and so on. This is the first reason.
Another key reason is that only GPUs cannot run a system. It still needs a CPU to do this, which is the central processing unit. The role of the CPU is to start the system, run most of the operating system and perform some other tasks.
In the past, there were few systems other than gaming and supercomputing that required both CPUs and GPUs, but now the situation is different. There are more and more systems that can apply artificial intelligence, so GPUs are starting to be used in large quantities.
GPUs are now widely used, from self-driving cars to video terminals, digital whiteboards, to factory equipment, and so on.
The third reason is that combining CPUs and GPUs and optimizing their performance is not a simple task. This explains why the merger of Nvidia and ARM will be a significant positive for the industry.
In addition, NVIDIA has been committed to extending the value of its products from the chip field to the system field. In this regard, NVIDIA and most chip manufacturers are not the same.
For example, the NVIDIA DRIVE series is an engineered system used in self-driving cars. It contains NVIDIA's GPU, CPU, various software, operating systems, and also includes the SDK. In other words, it is enough for automakers to apply DRIVE directly, and there is no need to redesign the system.
Clara in healthcare is another example. Clara is a healthcare application framework that includes the development and deployment of AI-driven imaging technologies, genomics, and smart sensors.
The combination of CPU and GPU allows NVIDIA to better optimize engineering systems. This is good for everyone, but it can also accelerate the pace of innovation.
The reason why some people oppose Nvidia's acquisition of Arm is that they believe that if the acquisition is successful, it will definitely further consolidate Nvidia's dominance in the chip industry. This is the market chooses both sides of the coin, with pros and cons.
Although this argument is not unreasonable, Nvidia does occupy more than 80% of the market share, deservedly in the dominant position, but Arm's share in the CPU field is less than 10%, far less than Intel and AMD's CPU market share, it can only be regarded as a small role in the CPU market.
For more than a decade, Intel has been exploring the GPU field and making many mistakes, but in general, there will not be much decline. Nvidia's acquisition of Arm will allow the latter to grow, but Intel will still firmly occupy a market position. The chip industry has many secondary markets, and if these sub-markets are not clearly distinguished, the chip industry will be considered to be just a whole.
Other concerns say many tech companies are now using Arm to power their devices, and more recently, Apple has done just that. Arm allows Apple to make thinner, lighter, and more powerful products.
Sam Shead wrote in a CNBC post, "Analysts say that if Nvidia buys Arm and doesn't let other companies use Arm's chips in their semiconductors, it will definitely have a huge impact."
If Nvidia had done so, it would have ended like this. But there is no evidence that NVIDIA had this intention.
NVIDIA's value is primarily focused on the company's ecosystem. NVIDIA has maintained the highest level of openness to maximize the number of vendors applying their technology.
Some critics point out that Nvidia's acquisition of Cumulus at the time caused Broadcom's partnership with Cumulus to break down. But this was actually initiated by Broadcom, not NVIDIA.
In fact, Nvidia fully supports the cooperation between Broadcom and Cumulus. But Broadcom sees Mellanox's acquisition of Nvidia as a threat.
That's why Broadcom launched the acquisition, which forced its customers and Cumulus to make a choice to change chip suppliers. Nvidia has always believed that users should make a choice, and they believe that the best technology will win in the end.
After the acquisition, Nvidia stressed that it will continue to invest in Arm and ensure that other companies can continue to use arm-designed chips. But the deal is not a success for now.
Arm and Nvidia are both very important companies for the future development of computing and AI, and if the acquisition is completed and the merger is successful, it will definitely accelerate the pace of innovation in many industries, and ultimately everyone will benefit.
Resources:
https://www.ft.com/content/59c0d5f9-ed6a-4de6-a997-f25faed58833
https://www.reuters.com/business/softbanks-66-bln-sale-arm-nvidia-collapses-ft-2022-02-08/
https://www.wsj.com/articles/nvidia-softbank-call-off-blockbuster-arm-deal-11644297283?mod=hp_lead_pos1
https://www.zdnet.com/article/why-scrapping-nvidia-arm-deal-is-ultimately-bad-for-the-industry/
https://baike.baidu.com/item/ARM/5907