A few weeks ago, NVIDIA said that it was postponing its acquisition of ARM Limited because of “severe regulatory hurdles.” Because NVIDIA has been the subject of intense scrutiny since Huang first unveiled the agreement, this revelation comes as no surprise. Several regulators and organisations have expressed concern about the company’s $40 billion bids for ARM Limited. Even though Huang was excited about the purchase, negotiations were relatively slow.
The acquisition of ARM by NVIDIA has been officially suspended, raising worries about the company’s future projects and paths. In an interview with Venture Beat, Jensen confirms that NVIDIA will continue its 20-year licence arrangement with ARM Limited for usage in processors, graphics cards, and DPUs, with no changes to the roadmap made public.
VentureBeat: What is your post-Arm strategy? Do you have to communicate your strategic direction in light of [the Arms deal being called off]?
Jensen Huang: Not anything. Because we never finished combining with Arm. So any strategies that would have come from the combination were never discussed. And so, our approach is the same. We do accelerated computing wherever there are (central processing units) CPUs. And so we’ll do that for x86. And we’ll do we do that for Arm. We have many ARM CPUs and system-on-chips (SoCs) in development. And we’re enthusiasts. We do all that. We have a 20-year license to Arm’s intellectual property. And we’ll continue to take advantage of all that and all the markets. And that’s about it. Keep building CPUs (graphics processing units), GPUs, DPUs (data processing units).
Jensen continues to add that NVIDIA will take benefit of the three prevalent architectures (x86, ARM, RISC-V) in their development and will determine which technology takes precedence for each design.
VB: So it’s your three-chip strategy? Would you consider RISC-V now that the Arms deal is not happening?
Huang: We use RISC-V. We’re RISC-V users inside our GPUs. We use it in several areas. For system controllers, inside the Bluefield GPU, there is a RISC-V acceleration engine, if you will, a programmable engine. And we use RISC-V when it makes sense. We use Arm when it makes sense. We use x86 when it makes sense.
With the termination of the ARM purchase, NVIDIA loses the potential to develop a more comprehensive roadmap for the future use of the ARM architecture. This shift in ARM’s acquisition will negatively impact NVIDIA because the two firms continue to collaborate well, and NVIDIA is poised to become the leader in high-efficiency power computing and networking technologies.
Also Read:
