It's been 16 months since Nvidia announced a $40 billion acquisition of Arm. It seems that the hope of passing regulatory approval for the semiconductor industry's largest merger has become increasingly slim. Perhaps for NVIDIA founder and CEO Jen-Hsun Huang, the acquisition of Arm is like a mirror flower, a very beautiful but extremely difficult dream to achieve.
Chip industry sky-high M&A
In September 2020, Silicon Valley chip giant Nvidia announced a wholly-owned acquisition of British chip design company Arm from Japanese technology giant SoftBank in the form of cash and stock. If the deal is successfully completed, it will be the largest deal in the history of the semiconductor industry. At that time, NVIDIA's market value exceeded $300 billion, and it was already the highest-value semiconductor company.
The second-largest acquisition in the semiconductor industry was the $37 billion acquisition of Broadcom by Avago (formerly HEWLP's semiconductor business) in 2015. It is worth mentioning that Chen Fuyang, a Chinese CEO from Malaysia, with superb capital operation and adventurous spirit, has continuously annexed several large and small chip companies such as LSI, Brocade and Broadcom in just a few years, and successfully built the world's fifth largest semiconductor company in the way of "snake swallowing elephant".
Of course, these two acquisitions are not the highest offers in the chip industry, and the two acquisition invitations with higher prices have finally collapsed under the veto of the regulatory authorities. In 2018, Chen Fuyang once again tried to stage a capital miracle, and successive quotations were intended to forcibly acquire Qualcomm. But the sky-high $120 billion acquisition was eventually rejected by the U.S. government on national security grounds. The Trump administration's veto reason is that Broadcom's acquisition of Qualcomm may affect Qualcomm's technological innovation and damage the advantages of the United States in 5G and the chip industry.
Although Qualcomm relied on the U.S. government to avoid a power of attorney dispute with Broadcom, in the same year, they were unable to obtain approval from the Chinese government because of the delay, and after two consecutive deadlines, they finally had to abandon the acquisition of NXP Semiconductors (NXP) in the Netherlands, and paid up to $2 billion in damages. The $44 billion deal, which was finally quoted, required review by nine major regulators around the world, and the only thing Qualcomm didn't wait for was the green light from China's Ministry of Commerce.
Chips have always been the most core area of the technology industry, and they are more directly related to the core competitiveness of a country. Therefore, the major mergers and acquisitions activities in the chip industry are not only related to the future prospects of both parties to the transaction, but also may trigger a chain reaction and weight adjustment of the entire industry, and may even affect a country's right to speak in the global economic system, thus triggering the most sensitive review nerves of regulatory departments in various countries.
Nvidia initially expected it to take 18 months to obtain regulatory approval to complete the transaction, but now 16 months have passed. With the current regulatory landscape, the prospects for the completion of the deal are increasingly slim. If Nvidia can't complete the acquisition of Arm, then the biggest merger in the semiconductor industry is still Avago's acquisition of Broadcom in 2015.
Bloomberg reported this week, citing unnamed sources, that Nvidia has been hopeless within it of obtaining regulatory approvals to complete the deal to buy Arm. Once Nvidia abandons the acquisition, SoftBank will push Arm to conduct an initial public offering. This report seems to validate the market's long-standing speculation. Perhaps Huang had to face the reality that obtaining regulatory approval was beyond his reach.
Nvidia's ambitious planning
The acquisition of Arm is the most important transaction for Jen-Hsun Huang in his plans for NVIDIA's future. Nvidia also offered very generous terms of the acquisition as a result. Under the acquisition agreement between Nvidia and SoftBank and Arm, Nvidia plans to pay $21.5 billion worth of stock and $12 billion in cash to acquire more than 92 percent of Arm's equity, with a $1.25 billion deal on a void term.
If Arm meets the set performance targets, Nvidia will also pay up to $5 billion. In addition, in order to retain talent, Nvidia will also pay $1.5 billion to Arm employees. In 2016, SoftBank spent $32 billion to acquire Arm. SoftBank currently holds a 75% stake in Arm, and SoftBank's Vision Fund holds the remaining 25%.
Why did NVIDIA buy Arm? Although Arm's valuation is only $40 billion, the annual revenue is less than $2 billion, and it is only a medium-sized enterprise in the chip industry, the strategic value of this company is not measurable by data, and it is not even an exaggeration to say that Arm is a pillar company in the global semiconductor industry.
The British chip design company does not produce its own chips, but only licenses the chip design architecture. Their chip architecture completely dominates the entire smartphone industry, whether it is Apple or Qualcomm, or Samsung, MediaTek and Huawei, for the time being, they are inseparable from Arm's chip design. Nvidia's acquisition of Arm controls the foundation of the smartphone industry.
However, Nvidia bought arm for the purpose of not a smartphone chip. At the time of the acquisition, Jen-Hsun Huang made it clear that they would not change Arm's current open licensing model to maintain a global customer neutrality position. Arm will remain as a subsidiary of NVIDIA and its headquarters will remain in Cambridge, UK.
NVIDIA's real purpose in acquiring Arm is to take its AI strategy to new heights, especially to open up the vast market of cloud computing. According to Huang Jenxun's plan, arm chip design in the future will run through AI technology, which will not only directly affect the smartphone industry, but also affect the data center market.
Disrupt future chips with Arm and AI
Compared with the smartphone industry, which has become saturated, the data center and smart car markets with huge growth space may be the areas that NVIDIA values more. Huang Renxun also made no secret of his ambitions in this field. The data center business is already the biggest driver of NVIDIA's performance, and the growth rate has even exceeded the rapid growth of the gaming business during the epidemic. The most recent third-quarter financial report last year showed that NVIDIA's data center revenue increased by 55% year-on-year, and the proportion of revenue has reached 41%.
Just six months before announcing the acquisition of Arm, Nvidia had just completed a $7 billion acquisition of server hardware company Mellanow. The acquisition was approved by regulatory authorities in the United States and China. Myros primarily provides end-to-end solutions for servers and storage, and the deal is considered a key move for NVIDIA's big push into the data center market.
Just after announcing the acquisition of Arm, Nvidia introduced DPUs (Data Processing Units) for future data center processors. Their solutions include multi-core CPUs based on arm architecture, GPUs that handle applications such as AI, and high-speed network transmission interfaces derived from the acquisition of SmartNIC technology from Mylos.
Although the data center market is still dominated by Intel's x86 architecture, low-power chips are increasingly becoming a key factor in the data center. In Jen-Hsun Huang's view, "both the data center and cloud computing markets are expecting Arm's processors." Power consumption directly affects compute performance, compute throughput, and configuration service costs. "Amazon, Qualcomm, and Huawei have all launched their own Arm-based products in the data center space, and Google has also launched its own TPU processor.
More importantly, with the maturity of machine learning technology, the data center processors required for future cloud computing have gradually shifted from the previous multi-core CPUs to parallel processing and the coexistence of different types of processors, including the GPU core of Nvidia. If Nvidia relies on Arm, AI and GPUs to make a qualitative leap in the field of data center chips, the valuation growth brought to NVIDIA can far exceed the $40 billion it paid to acquire Arm.
Obviously, heterogeneous computing will become the mainstream of future data center processors. In this regard, both Nvidia and AMD share the same strategic thinking. Nvidia has a housekeeper's GPU, first for network transmission technology to buy Mylos, and then for the CPU to buy Arm. If the acquisition of Arm is completed, Nvidia will complete the layout of trinity in the field of data centers and have the strength to challenge Intel in the future.
Just three weeks after Nvidia announced its acquisition of Arm, AMD announced an all-stock bid of $35 billion to buy FPGA chip giant Xilinx. The deal is clearly an industry chain reaction following Nvidia's acquisition of Arm. Xilinx can directly strengthen AMD's core competitiveness in the data center field and further compete for market share from Intel. Moreover, AMD and Xilinx have a highly consistent user base – telecom equipment vendors and data centers.
Compared to the huge response caused by NVIDIA's acquisition of Arm, AMD's acquisition of Xilinx has been much smoother. Just yesterday, China's State Administration for Market Regulation (SAMR) has conditionally approved AMD's acquisition of Xilinx, imposing a number of specific technical licensing restrictions on the business development of the two companies after the completion of the transaction. The two companies are expected to complete the merger in the first quarter of this year.
May repeat qualcomm mistakes
Although Huang Ishih was well aware of the industry implications of the acquisition of Arm and made it clear that he did not intend to change Arm's patent licensing model, he may have underestimated the difficulty of antitrust review of the merger. In a way, Nvidia's acquisition of Arm and Qualcomm's acquisition of NXP have many similarities.
When Qualcomm acquired NXP, the two companies did not overlap. Qualcomm chips focus on mobile computing and communications, while NXP chips focus on smart cars, Internet of Things, mobile payments and security. After Qualcomm's acquisition of NXP, it is equivalent to spanning many computing industries. Although Qualcomm promised to continue to license NXP technologies to the outside world with the same standards in the future, it ultimately did not resolve the concerns of Chinese regulators.
In contrast, Nvidia's acquisition of Arm has sparked more controversy. In the view of Bharat Kapoor, an analyst at market consultancy Kearney, the acquisition was unlikely to be completed in the first place. Acquiring a cornerstone technology and platform that the entire industry relies on is a deal that is difficult for any company to swallow.
Although Nvidia's target for arm acquisition is not the smartphone industry, they currently have almost no mobile chip field, and Nvidia has promised not to change Arm's patent licensing model in the future, but this does not change the regulatory and technology industry's skepticism of Nvidia's control chip cornerstone company.
Whether it is the regulatory authorities in the United States, the United Kingdom, China, or industry giants such as Google, Qualcomm, and Intel, they have expressed concerns about the core competitiveness of the industry controlled by Nvidia after the acquisition of Arm, or explicitly opposed it. Qualcomm has made it clear from the beginning that it opposed Nvidia's acquisition of Arm, fearing that Nvidia would control the core technology of the mobile chip industry through this transaction.
Arm co-founder Hermann Hauser has publicly said the deal will be "a disaster that will destroy its business model." "Arm was the last European technology company with global relevance that was sold to Americans," Hauser said. The deal, announced overnight, would destroy ARM's business model as the "Switzerland of the semiconductor industry," while Nvidia competes with ARM's customers. ”
There is little hope of regulatory approval
The first to stand up against it was arm's British government. Last April, the British government asked the Competition and Markets Authority to investigate the deal, citing national security, which launched a formal investigation into the Arm deal last November. Nvidia also submitted a 29-page report to the U.K. government outlining the potential benefits of the deal for Arm and the U.K.
It is worth mentioning that the British antitrust agency asked Facebook to sell its previously acquired GIF platform Giphy in December, arguing that the deal would hurt market competition in the social media sector. Just this month, the UK's Competition and Markets Authority launched an antitrust investigation into Microsoft's $16 billion acquisition of Nuance. In addition, in November last year, the UK government enacted the National Security Investment Act, which plans to authorize the government to intervene in foreign investment and M&A transactions with UK companies on the grounds of national security. Nvidia's acquisition of Arm is likely to be the target of a new bill for the British government.
Strengthening antitrust supervision in the technology industry and preventing giants from enhancing their strength through mergers and acquisitions has become the consensus of global regulators. Last December, the U.S. Federal Trade Commission (FTC) formally sued Nvidia, demanding that it reject Nvidia's acquisition of Arm, arguing that the transaction would hurt the innovation of next-generation chip technology in many areas, including data centers and smart cars.
The FTC clearly stated that NVIDIA's acquisition of Arm will directly affect innovation and competition in three major technology areas: DPU SmartNICs used in data center servers, ADAS (Advanced Driver Assistance Systems) used in smart cars, and Arm architecture CPUs used in cloud computing servers.
The FTC said in the lawsuit that the reason for blocking the largest acquisition in the history of the semiconductor industry is to avoid chip giants stifling the innovation path of the next generation of technology. Future technological innovation depends on maintaining the competitiveness of the current chip market. Nvidia's acquisition could interfere with Arm's innovation in the chip market, causing unfair damage to Nvidia's competitors. The lawsuit is intended to send a clear signal that positive action must be taken to protect critical infrastructure markets from illegal vertical M&A transactions that will have a profoundly devastating impact on future innovations.
The FTC said it worked closely with antitrust agencies in the European Union, the United Kingdom, Japan and South Korea to investigate NVIDIA's acquisition of Arm, and the committee unanimously agreed to reject the deal in a 4-0 vote. Hearings on the lawsuit will begin on August 9 this year. This also means that Nvidia is unlikely to complete the approval according to the previous schedule.
More importantly, Chinese regulators have yet to issue a formal opinion on the deal, and Nvidia will almost certainly face greater skepticism than the U.S. and British governments. Considering that semiconductors have become the core strategy of China's industrial upgrading and technological innovation, the acquisition of Arm, the cornerstone of the global chip industry by US technology giant Nvidia, is obviously not in line with China's national security strategy, especially in the context of the US government's continuous sanctions against Chinese technology companies such as Huawei through entity lists and other means.
Nvidia's deal to acquire Arm will expire two years later, in September. However, even if Nvidia is eventually forced to abandon the acquisition of Arm under regulatory pressure from various countries, it will not affect their current market competition pattern and will not affect their product impact in the field of data centers and artificial intelligence. Nvidia already has the license for the Arm chip.
Perhaps Nvidia lost only the $1.25 billion in break-up fees it paid earlier, and their market value has grown by more than $200 billion over the past year and a half. Even if NVIDIA's stock price has rebounded sharply with the US stock market in the past month, the current market value is still as high as $550 billion.
An NVIDIA spokesperson replied by email: "We remain committed to the views detailed in our previous submissions to regulators, and this transaction presents opportunities to accelerate Arm (development) and drive market competition and innovation." ”
Source: Sina Technology
![There is no "neutral country" in the semiconductor world[fig]](data:image/gif;base64,R0lGODlhAQABAIAAAP///wAAACwAAAAAAQABAAACAkQBADs=)