Bloomberg Technology: Live From Nvidia's GTC Event

Key Concepts

Quantum Computing, Accelerated Computing, Supercomputers, AI, Qubits, Error Correction, Calibration, Quantum Annealing, Gate-Based Quantum Computing, Blockchain, Hashing Functions, Proof of Work, Quantum Volume, Quantum Advantage, CUDA, DGX, Material Discovery, Chemistry Applications, Cryptocurrency Mining.

NVIDIA's Role in Quantum Computing

NVIDIA does not build or sell quantum computers. Instead, NVIDIA provides architecture, software (like CUDA), and services to quantum computing companies. Their vision involves integrating large-scale quantum processors into existing data centers, similar to today's complex computing infrastructure with storage and interconnects. NVIDIA focuses on helping quantum tech companies develop their technologies and setting up the infrastructure for error correction and device calibration. These tasks are computationally intensive and suited for accelerated supercomputers. NVIDIA is working with over 160 groups in quantum computing, providing technology to accelerate their work across a wide range of applications.

The Convergence of Quantum and Accelerated Computing

Historically, quantum computing and accelerated computing (supercomputers for AI) have been distinct fields. However, these worlds are increasingly converging. AI models can be used to control and error-correct larger, more capable quantum devices. Conversely, quantum computers, which are essentially physics experiments modeling quantum physics, can potentially provide data to train AI models for understanding phenomena within the computer itself.

Defining "Useful" Quantum Computing

Defining "useful" in the context of quantum computing remains a challenge. Tim Costa (NVIDIA) emphasizes the importance of narrowing the focus and defining specific problems to solve, allowing for a clear definition of success. There's a general consensus that early applications will emerge in chemistry and biochemistry, where modeling quantum physics within the device aligns with the requirements for accurate chemistry simulations.

NVIDIA's Research Center in Boston

NVIDIA is establishing a research center in Boston to develop error correction technologies and integrate quantum devices with GPUs, building early versions of quantum-accelerated supercomputers. This center will facilitate collaboration between NVIDIA and its partners.

IonQ's Perspective (Peter Chapman)

IonQ's CEO, Niccolo Tomasi, replaced Peter Chapman (now Executive Chairman) a week prior to the event. Chapman stated that the purpose of the event was to "take back" Jensen Huang's earlier comments about quantum computing being a decade away from usefulness. IonQ demonstrated a 20x improvement in performance using NVIDIA and AWS, and a 12% increase using their quantum computers on a product normally run on GPUs. While these improvements aren't yet disruptive, they are remarkable given IonQ's 36-qubit system. IonQ uses DGX clusters for designing quantum computers, but not for error correction, which they handle internally. They also use GPUs for simulating smaller qubit counts to verify hardware functionality. Chapman emphasized the need to find early applications to generate revenue and fund R&D, citing chemistry applications as a promising area. He downplayed a recent short report on IonQ's stock, dismissing it as an attempt to cast doubt on the company for financial gain.

D-Wave's Perspective (Alan Baratz)

D-Wave takes a different approach to quantum computing than the rest of the industry, focusing on quantum annealing. Alan Baratz (D-Wave CEO) maintains that D-Wave already supports useful applications today. He cited a paper published in "Science" demonstrating the ability to compute properties of magnetic materials that cannot be computed classically, enabling new material discovery platforms. D-Wave is also applying its technology to blockchain architecture, using quantum computers to compute hashing functions for proof of work, potentially reducing energy consumption in cryptocurrency mining. D-Wave has a prototype blockchain running on four of its quantum computers, aiming for a full commercial blockchain in a year or two. Baratz stated that D-Wave's quantum computers do not have the same error correction requirements as gate-based quantum computers, and therefore NVIDIA's error correction solutions are not as important to them. He also stated that D-Wave does not need NVIDIA's help to calibrate consistency. He believes that the event was not helpful to the industry or to D-Wave.

Key Statements

• Jensen Huang (NVIDIA CEO): "I'm a public company CEO, and every so often someone asks me a question... Some of the time, I say something right. And sometimes it comes out wrong." (Referring to his earlier comments about quantum computing's timeline).
• Tim Costa (NVIDIA): "What we are focused on is really helping quantum tech companies who are building those technologies to better develop those technologies, because we are interested in solving the problems it will be able to solve."
• Peter Chapman (IonQ): "What Jensen said, one area we agree on is you need to find a set of applications early on that you can start to make money on, and build that firewall to be able to power your R&D."
• Alan Baratz (D-Wave): "We are able to support useful, important applications today... That seems pretty useful to me."

Technical Terms Explained

• Qubit: A quantum bit, the basic unit of information in a quantum computer.
• Error Correction: Techniques used to mitigate the effects of noise and errors in quantum computations.
• Calibration: The process of tuning and optimizing quantum hardware to ensure accurate and reliable operation.
• Quantum Annealing: A quantum computing approach used to find the minimum of a function, often applied to optimization problems.
• Gate-Based Quantum Computing: A quantum computing approach that uses quantum logic gates to manipulate qubits.
• Blockchain: A distributed, decentralized, public ledger used to record transactions.
• Hashing Functions: Mathematical functions that convert input data into a fixed-size string of characters (hash).
• Proof of Work: A mechanism used in blockchain to validate transactions and prevent double-spending.
• Quantum Volume: A metric used to measure the performance of a quantum computer, taking into account the number of qubits, connectivity, and error rates.
• Quantum Advantage: The point at which a quantum computer can solve a problem that is intractable for classical computers.
• CUDA: NVIDIA's parallel computing platform and programming model.
• DGX: NVIDIA's line of servers designed for AI and high-performance computing.

Logical Connections

The discussion flows from NVIDIA's role as a provider of infrastructure and software for quantum computing to the perspectives of two quantum computing companies, IonQ and D-Wave. The conversation highlights the differing approaches to quantum computing (gate-based vs. annealing) and the varying levels of agreement on the current state and future direction of the field. The discussion also touches on the importance of defining "useful" quantum computing and finding early applications to generate revenue and drive further development.

Synthesis/Conclusion

The video presents a multifaceted view of the quantum computing landscape, highlighting NVIDIA's supporting role and the diverse perspectives of quantum computing companies. While challenges remain in defining "useful" quantum computing and achieving quantum advantage, the industry is making progress in developing hardware, software, and applications. The convergence of quantum and accelerated computing, along with the exploration of new application areas like material discovery and blockchain, suggests a promising future for quantum technology. The event served as a platform for NVIDIA to clarify its position and for quantum computing companies to share their progress and perspectives, even if disagreements persist.