• Sat. Sep 19th, 2020

Dimancherouge

Technology

What Nvidia buying Arm means for your next smartphone

Nvidia announced on Sunday evening that it would acquire computing giant Arm for $40 billion. During a press call held early on Monday morning, Nvidia CEO Jensen Huang and Arm CEO Simon Segars shed more light on what the deal means for existing markets, such as smartphones, and what the future holds for the two companies.

The most important revelation is that Arm’s core business strategy will remain independent. Nvidia is purchasing Arm partly for access to its huge range of licensing partners and these companies can continue to use Arm intellectual property as before. Nvidia plans to introduce its own graphics and AI IP into the chain, greatly broadening its reach into various computing markets around the world.

Here’s what Nvidia buying Arm means for the mobile market and the tech industry at large.

Nvidia buys Arm: The ins and outs

Pending regulatory approval, Nvidia is purchasing Arm from SoftBank for a total sum of $40 billion. This will be paid in the form of $12 billion in cash and $21.5 billion in stock. Regulatory approval could take around 18 months. During that time Arm and Nvidia continue to operate as separate entries and will be sticking to their current public roadmaps. If the deal goes through, Arm will become a business unit of Nvidia.

Nvidia — currently a customer of Arm — notes that Arm will retain its operational independence. In the sense that it will continue to operate as a subsidiary company rather than being incorporated into Nvidia at large. In other words, the deal won’t chop Arm up for intellectual property or staff. This is important, as Arm’s core IP licensing business will remain as is. Various partners currently licensing Arm technology, such as mobile giants Apple and Qualcomm, can continue to do so. However, we don’t yet know if Nvidia will end up utilizing Arm technology for future proprietary designs.

Latest Arm processors: Arm Cortex-X1 and A78 | Arm Mali-G78 and G68

The big picture is that Nvidia bought Arm in part to access its licensing business model and Rolodex of existing customers. Huang wants to expand Arm’s IP licensing portfolio with Nvidia’s world-leading GPU and AI technology. This is part of Nvidia’s plan to bring its AI compute power to a much broader range of devices, from tiny internet-of-things processors to server farms. Oddly enough, Nvidia will continue to use RISC-V, a competing CPU architecture, for a number of small processors used in its GPU and other products.

Arm’s IP licensing business will remain as is, Nvidia wants its Rolodex of existing customers

The other elephant in the room is the implication of US-based Nvidia buying UK-based Arm. Particularly when it comes to deals with partners based in China. Arm’s Simon Segars notes the majority of Arm’s IP is held in the UK and much of the remainder is based outside of the US, so doesn’t fall under US export rules. This won’t change following the Nvidia purchase.

Nvidia also wants to assure its commitment Arm’s UK base by investing in a new AI data center in the city of Cambridge. The site will include an Arm-powered supercomputer and facilities to serve AI researchers, scientists, and startup companies. The overarching message is that it will be business as usual for Arm, but we won’t know for certain until the ink is dry.

What this means for future smartphones

ROG Phone 2 Front shot of gaming playing Minecraft

All current smartphone SoCs are built using Arm CPU and other technologies, so this deal has major implications for future mobile processors. Arm’s list of high-profile partners includes Apple, Qualcomm, Samsung, and more. As the IP situation will remain unchanged, these companies can continue to access existing and future component designs from Arm, as well as their architecture licenses. It’s business as usual for mobile SoC designers. Although whether these companies are happy with profits funneling to their rival remains to be seen.

What’s perhaps more interesting is that Nvidia stands to gain a major foothold in the PC CPU space. With Apple making the switch to in-house Arm PC processors and the growth of Windows on Arm, Nvidia adds another string to its bow in the battle with Intel and AMD in PCs and servers. If we’re on the precipice of a major shift in the Arm vs x86 PC dynamic, the timing of Nvidia’s purchase is impeccable.

Arm vs x86: Instruction sets, architecture, and all key differences explained

Nvidia is best known for its graphics processors and so the question on mobile enthusiasts lips is whether future SoCs could feature GPUs designed by the green team. The Tegra chips powering the Nintendo Switch already showcase Nvidia graphics running at 7.5-15W TDP, and Nvidia’s Jetson Nano board brings its Maxwell architecture as low as 5W. The improved efficiency of newer Turing and Ampere architectures might run well within the 5W requirements of smartphones. So yes, smartphones powered by Nvidia graphics are definitely possible.

Don’t bet on seeing a smartphone with an Nvidia GPU any time soon.

That said, Huang confirmed that Arm’s existing Mali graphics solutions aren’t going anywhere. Mali serves a wide range of use cases, from TVs to smartwatches, and Nvidia won’t be forcing existing Arm customers to switch to a new IP. However, it does plan to introduce new options for customers who want access to Nvidia’s rich ecosystem of accelerated computing, gaming, and other tools. Just don’t bet on seeing a smartphone with an Nvidia GPU any time soon.

It’s also worth noting that Qualcomm has its in-house Adreno GPU team and Samsung has a deal with AMD for future chipsets. So there’s no obvious large-scale high-end partner for smartphone SoCs featuring Nvidia graphics right now. We’ll just have to wait and see what Nvidia brings to the table for sub-5W use cases in the coming years.

AI everywhere: A shared company vision

NVIDIA Xavier NX development kit see from the top

While Nvidia wants to convince pundits, partners, and regulators that it will mostly be business as usual for Arm, the company isn’t just interested in banking some of the world’s most important intellectual property. Just like SoftBank, Nvidia is so interested in Arm because of its current position and further potential for powering the vastly interconnected world of the internet-of-things.

Arm’s low power RISC processor design and developer ecosystem have made it a key processor player from power-hungry data servers to tiny battery-powered edge devices. Nvidia wants to leverage Arm’s ecosystem to integrate its industry-leading AI processing capabilities into this range of product segments. Their combined IP, licensing experience, and investment potential positions Arm/Nvidia as a one-stop-shop for the AI technologies of tomorrow.

See also: Nvidia Jetson Xavier NX review: Redefining GPU accelerated machine learning

Of course, this is a longer-term outlook that will play out over the next decade. In the meantime, Nvidia will have to balance the expectations of Arm’s existing customers with its pursuit of cornering the future AI market. That may not be such an easy task if existing partners decide to reevaluate their options in light of this purchase. It’s now over to the regulators to decide if this match is in the tech industry’s broader interests.

Source Article