This week it was reported that NVIDIA is considering backing away from its Arm acquisition due to regulatory difficulties. As I wrote back in December, this could be a far better outcome for NVIDIA than acquiring the company would be.
Now that that’s a real possibility, let’s look at what NVIDIA could do next to accomplish as much, or more, than it could have with the acquisition.
Freed from Regulatory Restrictions
Early on, it was clear that the rules that would have constrained NVIDIA post-merger would have been onerous. The combined company would have likely had to endure invasive oversight, an unusual amount of strategic transparency, and been blocked from doing anything that would have been seen as giving it a strong competitive edge. In other words, much of what NVIDIA probably wanted to do with Arm it would have been prevented from doing.
But as I mentioned in that December post, Arm is a licensing entity, and there are no restrictions on partnering more closely with a licensing entity. So, NVIDIA remains mostly free to do what it wants with the technology it licenses from Arm.
A Path to a Merged CPU/GPU
The market is currently defined by a particular model, in which the CPU and the GPU are separate parts connected by a high-speed interface. While there are parts, like AMD’s APU, that put both on the same die, none of the major vendors have created a part that blends both functions, which could potentially redefine the performance level for inexpensive computing devices that need both a CPU and a GPU and might even redefine high-end performance machines by eliminating bottlenecks that currently exist between the two very different architectures. In other words, this new blended part could step away from both architectures and create something new that outperformed any separate CPU/GPU configuration.
This part could be used for anything from a new class of personal technology devices to more advanced robots and AI systems, both of which NVIDIA is already into. Both PCs and smartphones have been in the market for decades, and it is arguably past time for both platforms to evolve. After all, autonomous robotics, vehicles and advanced AIs are all redefining where and how we compute, so the opportunity exists for someone to rethink the space. NVIDIA is ideally positioned to redefine these markets given the firm’s deep involvement with them.
As now, the market is exploring head-mounted displays and advanced digital assistants, and it’s starting to get a feel for general purpose AIs. Were a new, higher-performance, blended architecture to appear, it might create a significant advantage for the company, assuming the firm was well connected (and NVIDIA is) to all those segments.
NVIDIA is one of a handful of firms that has the bandwidth to explore something revolutionary right now. AMD lacks the breadth, Intel is still in recovery and focused on building a mega fab, and Qualcomm is tightly tied to its existing architecture and the current smartphone and PC markets. NVIDIA is the only company aggressively looking at other advanced opportunities, like flying autonomous cars and advanced robotics, where a revolutionary part might provide an even greater long-term competitive advantage.
Also read: NVIDIA Turns Its Focus to AI Inference
Good for NVIDIA and Arm
Given that most acquisitions fail and that the regulatory environment might have simply assured that failure, the loss of the Arm acquisition is arguably better for NVIDIA. The company will remain free of the operational and regulatory headaches that a foreign merger of an underfunded subsidiary would have caused.
Arm going public is also likely better for Arm because it won’t have to deal with a merger. The potential financial upside to employee options should allow it to better retain and acquire talent than when it was under an acquisition cloud and the related incentives were less certain. Granted, this would depend on how Arm’s IPO goes. We are looking at a harsh stock market now.
In the end, though, I think the failed acquisition opens NVIDIA’s door to something even more interesting and potentially powerful by providing the opportunity to solve the company’s lack of CPU in a more creative way. That way could be a matter of rethinking the system in order to create a revolutionary SOC (System On Chip) that could redefine NVIDIA’s dominance in the segment.
In short, losing Arm could end up being the most successful path for both NVIDIA and Arm.
Read next: NVIDIA’s Huang Gets Overdue Recognition