In many ways, the current spate of generative artificial intelligence (AI) services like ChatGPT can thank Nvidia (NVDA 1.78%) for their existence. Nvidia researches and designs semiconductors that accelerate computing time -- most notably GPUs (graphics processing units), originally used for high-end video game graphics and now being applied to train large-language models (LLMs) that ChatGPT runs on.
But a new breed of computing accelerator is being developed: quantum computing. Nvidia has announced software geared toward quantum computing researchers over the last few years. But does its latest hardware announcement, the DGX Quantum, propel the company into the world of designing quantum computing systems, too?
NASDAQ: NVDA
Key Data Points
Nvidia: the quantum computing company?
Nvidia's DGX Quantum is actually a collaboration with Israeli start-up Quantum Machines, which makes some hardware components for a quantum computing system. Specifically, DGX Quantum features a new Nvidia Grace Hopper GPU paired with an OPX+ from Quantum Machines, a "quantum control" unit that orchestrates the operations of a quantum computer.
The DGX Quantum, an Nvidia Grace Hopper GPU system paired with a Quantum Machines OPX+. Image source: Nvidia.
What does that mean? DGX Quantum isn't itself a quantum computer. Rather, it's a component in a quantum computing system, albeit an important one at this early stage of research and development for practical use of quantum computing.
According to Nvidia, a complete quantum computer requires a QPU (quantum computing processor, akin to a CPU or GPU), a system to perform operations on the QPU, a way to measure and record the resulting data, and a way to create and connect multiple QPUs that can operate with each other.
Quantum DGX is thus a means to govern the operation of this quantum computer system. And Quantum DGX will also make use of Nvidia's CUDA Quantum software stack, a hybrid platform of classical computing (CPUs and GPUs) and quantum computing (QPUs) that unifies an entire system to make it useful for researchers.
A growing number of quantum computer operators (like the National Institute of Advanced Industrial Science and Technology in Japan), quantum hardware makers IonQ, and start-up software developers (Agnostiq and QMware) have adopted CUDA Quantum in their operations.
The first deployment of the Quantum DGX hardware, the Israel Quantum Computing Center, is expected by the end of 2023.
A bet on an unknown future
At its most basic level, quantum computing is simply a means to further speed up computing times, enabling researchers to experiment with big problems that classical computers can't solve. Accelerated computing is Nvidia's specialty, so dabbling in quantum computing research is a natural fit for the semiconductor company.
But as its management told me in a recent conversation, widespread commercialization of quantum computing is still likely many years off. And when it does start to take root, quantum computing will supplement classical computing, not replace it -- much like Nvidia GPUs have accelerated the work traditionally handled by CPUs, not replaced them. Think of quantum computing and the QPUs that power them as another (future) extension of the classical computing work already done up to this point.
And given their industry-leading performance, Nvidia GPUs are helping propel this virtuous cycle of using the most advanced technology of today to develop the cutting edge technology of tomorrow.
Given that quantum computing is in its nascent stage, most investors who want to bet on it would be best suited by sticking to long-term ownership of stocks like Nvidia at this juncture.
