Today there is more attention on quantum computing (QC) than there was on generative AI the day before ChatGPT launched in November 2022. That should make everyone stop and think about what is possible for QC in 2025. Let's dig into the topic.

Investments in artificial intelligence and the push to keep advancing QC reveal two very different paths.

Agentic AI and QC have dominated the headlines in the first days of 2025. On one side, there is a widespread consensus that 2025 is already the year of AI. On the other, skepticism persists about whether QC can become commonplace and profitable in the near future, especially after NVIDIA CEO Jensen Huang said at CES 2025 that "very useful" quantum computing is still 20 years away.

Both views, however, face their critics. Leading AI experts such as Gary Marcus and Rodney Brooks have argued that AI's short-term potential is overrated and that its return on investment is hard to achieve, criticisms that remain relevant in 2025. That is why agentic AI has gained traction: it narrows the scope of AI applications and focuses on specific tasks to maximize the gains of today's technology, rather than striving for a more ambitious general AI. By contrast, several QC vendors have challenged Huang's 20-year forecast for useful QC. They say QC is already commercially viable for specific operations and that work toward broader applications continues steadily. D-Wave Quantum CEO Alan Baratz, for example, said Huang is "dead wrong" about QC's readiness. In an interesting move, NVIDIA also announced its first Quantum Day, scheduled for March 20 during GTC 2025, which further underscores its own stake in QC.

This tension raises hopes of faster progress in QC and adds uncertainty about its timeline. Fundamentally, that uncertainty does not necessarily signal delays; it could suggest that breakthroughs may arrive sooner than expected.

There is a commercial and technological tension between classical-computing AI and QC

There is a lot to unpack here. First, Huang's comments caused a notable drop in quantum-related stocks, which prompted a backlash from the QC community. Daniela Herrmann, cofounder of Dynex, a quantum-as-a-service company, told The Observer: "Huang's extended timeline reflects the broader understanding of the challenges in achieving fully developed quantum systems, but it does not fully acknowledge the progress that existing quantum-inspired applications are making." The Observer further noted: "Herrmann highlighted that Huang's comments seem to keep the focus on initiatives like NVIDIA's CUDA-Q, which has yet to solve real-world problems at scale."

Despite Huang's caution, NVIDIA remains actively involved in QC, mainly through collaborations and supporting technologies. NVIDIA's CUDA software platform integrates with QC efforts, fostering a hybrid model that combines classical and quantum computing. At CES 2025, Huang said: "Almost every quantum computing company in the world is working with us now… We're extending CUDA to quantum, and they use us to simulate the algorithms, simulate the architecture, create the architecture itself." The intricate relationship among NVIDIA, classical-computing AI and QC reflects a dynamic in which Huang's long-term prediction can coexist with evidence of tangible near-term progress.

Original source: Forbes · Luis E. Romero