Nvidia opened Pandora's box, deciding to buy ARM. If the deal is approved.
Last week, Nvidia announced that it intends to buy ARM for $40 billion. The amount is considerable even for the technology sector, but besides the fact that $40 billion is a tenth of Tesla, in this transaction is not particularly clear. How will this event affect the computer industry? What does artificial intelligence-related development have to do with it? What will happen to mobile devices, most of which use ARM architecture? Below we will try to answer these questions.
Let's just say ARM Holding is an ARM Holdingcompany, but it's also the processor architecture that ARM has developed. Its full name is Advanced RISC Machines, by the way, hence the acronym "ARM". It develops, designs and licenses the ARM-architecture of processors. And companies such as zualcomm, Apple and Samsung are developing their ARM-based processors.
ARM Holding offers two types of licenses for its technology. The first concerns the design of chipsets: the manufacturer pays ARM money, develops its processors, using most of the ARM technology, including its design. This is the easiest option, which is used by many smartphone manufacturers. The second option involves using only the ARM architecture.
Most technology companies are ARM licensees
That is, another corporation has the right to make its own processor with its own design and other features. That's how few people work, one of those companies is Apple, which makes IPhone and iPad processors based on ARM, and now chips for Macs (Apple Silicon).
So you have no idea how important this company is from London with an uncomplicated three-letter name. But why did Nvidia need ARM?
During the discussion of the deal, Nvidia representatives said that ARM's core business strategy will remain independent. Nvidia buys ARM in part to access a huge number of licensing partners, and these companies can continue to use ARM's intellectual property as before. But did Nvidia just need ARM licensees? Of course not.
Contrary to popular belief, Nvidia also produces graphics chips. It also supplies chips for data centers, data centers and computing systems. The company has adapted GPUs to handle large data sets and has largely outperformed its competitors, including Intel.
Nvidia will be able to use ARM technology seamlessly while integrating its own solutions. Nvidia plans to build a huge data center for artificial intelligence research, which can be used in various fields after the completion of the transaction. It will have a supercomputer based on ARM processors.
The project of the new Nvidia data center, where there will be a supercomputer with an ARM processor
However, Nvidia's goals go far beyond AI. Buying ARM guarantees its dominance in the global cloud data centermarket. Nvidia has only one competitor left there- Intel, and with the help of ARM it will be able to move it. How? There are at least two reasons:
There is no need to go far for an example: TSMC, a processor manufacturer that switched to ARM, has become one of the industry leaders, leaving Intel behind.
Nvidia has already bypassed Intel and AMD by capitalization, and this is just the beginning
Buying ARM solves all Nvidia's problems in the fight against Intel and the takeover of the server market. The company will no longer be dependent on Intel and AMD and will be able to seamlessly combine high-performance ARM processors with its own GPUs.
Nvidia, currently an ARM customer, notes that the company will retain its operational independence. In the sense that it will continue to operate as a subsidiary rather than be included in Nvidia. In other words, the deal will not hurt other arm companies or staff. This is important because ARM's core intellectual property licensing business will remain the same. Various partners who currently license ARM technology, such as mobile giants Apple and zualcomm, can continue to do so. However, we do not yet know whether Nvidia will eventually use ARM technology for future proprietary projects.
Another question is whether this deal will take place at all. Although in fact the American company buys British, it is necessary to take into account the interests of China. ARM has many partners in China, and if the company becomes American, the U.S. in theory will be able to impose sanctions andto deny Chinese vendors access to ARM technology. Yes, they can do it. Therefore, the deal must be approved by the antitrust authorities of the United States, Great Britain and China.
So far, ARM is owned by the Japanese investment company SoftBank. ARM Holding was acquired in 2016 for $32 billion. The demand for ARM chips increased in the following years, by the end of 2016 more than 95 billion of these chips were sold worldwide (17.7 billion in 2016 alone). As of the end of March 2020, the total number of ARM chips sold exceeded 160 billion. There was no conflict of interest when the company was owned by SoftBank. Nvidia is a different matter.
All modern processors (systems-on-chip, a.s.a. SoC)for smartphones are built using ARM and other technologies, so this deal will have serious implications for future mobile processors. The list of high-ranking Arm partners includes Apple, zualcomm, Samsung and many others. These companies can continue to gain access to existing and future ARM component projects as well as their architecture licenses. But where is the guarantee that at one point Nvidia will not change his mind and will take away all the technology itself? In this case, the smartphone market is waiting for a huge blow. This scenario is unlikely, but it cannot be completely ruled out.
Nvidia is best known for its GPUs, so many in the question of whether the processors of future smartphones can contain Nvidia graphics chips. So far, it is known that the existing arm graphics solutions - Mali, will not go anywhere. ARM Mali serves a wide range of devices, from TVs to smartwatches, and Nvidia won't force existing ARM customers to switch to something new. So long.
Nvidia will get at its disposal a company that keeps half of the industry in its fist
It's also worth noting that the company has its own Adreno GP team, and Samsung has struck a deal with AMD on future chipsets. So there is currently no obvious partner for the release of smartphones with Nvidia graphics.
GeForce will develop separately for the time being
ARM and GeForceare likely to remain separate areas of Nvidia's business. Perhaps we will feel the impact of the ARM purchase on the overall budget of the corporation, as Nvidia seems to be hoping to use its GPUs to license large-budget customers who are now ARM partners. But this potential financial injection is likely to happen only when the dust settles and ARM starts to recoup its investments. Until Nvidia says that buying ARM will somehow affect .
Of course, this is a longer-term perspective that will be realized over the next decade. In the meantime, Nvidia will have to balance the expectations of existing ARM customers with its desire to conquer future markets, from data processing to AI and graphics in smartphones. This may not be an easy task if existing partners decide to renegotiate their arrangements in light of this purchase. Regulators now have to decide whether the purchase is in the broader interest of the technology industry.