Interesting article and angle. JH seems very open and speaking his mind. But over the years he has been very nimble and willing to listen to others, or so it seems to me. Personally i have a lot more respect and trust for someone who can admit to being wrong and change their mind, than someone who thinks they know everything and cannot evolve their thinking.
Do you have a view of how Quantum affects AI? And, vice versa? I would like to understand better what Quantum means for the existing markets.... for example, will one quantum computer replace millions of CPUs/servers? How do you do input and output fast enough to keep up? Or is it going to primarily used for extremely sparse but computationally intense workloads? What does the Quantum ecosystem look like? Does Quantum consume a lot of energy? Will you need 5 power plants for each Quantum computer?
Also, a big question in my mind is who is likely to be a winner? Right now you have a small bunch of start ups and the biggest players in tech, all at the same time. Do you think that Quantum is going to end up in the hands of the likes of GOOG, AMZN, MSFT?
Quantum computing isn’t going to replace millions of CPUs or become a general-purpose server alternative. Instead, it will act as a specialized accelerator—much like GPUs did for AI—targeting a narrow class of problems such as molecular simulation, optimization, and certain cryptographic workloads. It won’t train large language models or handle typical cloud compute tasks. Rather, it will work alongside classical systems in a hybrid architecture: CPUs for control, GPUs for parallelism, and QPUs for sparse but computationally intense problems. Input/output and error correction remain major bottlenecks, and while quantum systems need infrastructure like cryogenic cooling or lasers, the energy footprint is not as apocalyptic as feared—especially when compared to the energy cost of brute-force classical alternatives.
As for winners, the most likely scenario is a split between the hyperscalers—Microsoft, Amazon, Google—who control delivery via cloud platforms, and a few standout startups that solve the scalability and coherence challenges in hardware. PsiQuantum, with its CMOS-compatible photonic approach, is one such contender. IonQ and Rigetti could survive as acquisition targets or niche providers. In the near term, we’ll see quantum used more in simulation and research than mainstream enterprise applications—but long term, its positioning as an enabler for advanced AI and cryptographic workloads is clear. Quantum won’t replace the data center. It will quietly reshape its deepest layers.
Interesting article and angle. JH seems very open and speaking his mind. But over the years he has been very nimble and willing to listen to others, or so it seems to me. Personally i have a lot more respect and trust for someone who can admit to being wrong and change their mind, than someone who thinks they know everything and cannot evolve their thinking.
Do you have a view of how Quantum affects AI? And, vice versa? I would like to understand better what Quantum means for the existing markets.... for example, will one quantum computer replace millions of CPUs/servers? How do you do input and output fast enough to keep up? Or is it going to primarily used for extremely sparse but computationally intense workloads? What does the Quantum ecosystem look like? Does Quantum consume a lot of energy? Will you need 5 power plants for each Quantum computer?
Also, a big question in my mind is who is likely to be a winner? Right now you have a small bunch of start ups and the biggest players in tech, all at the same time. Do you think that Quantum is going to end up in the hands of the likes of GOOG, AMZN, MSFT?
Hi Ken,
Quantum computing isn’t going to replace millions of CPUs or become a general-purpose server alternative. Instead, it will act as a specialized accelerator—much like GPUs did for AI—targeting a narrow class of problems such as molecular simulation, optimization, and certain cryptographic workloads. It won’t train large language models or handle typical cloud compute tasks. Rather, it will work alongside classical systems in a hybrid architecture: CPUs for control, GPUs for parallelism, and QPUs for sparse but computationally intense problems. Input/output and error correction remain major bottlenecks, and while quantum systems need infrastructure like cryogenic cooling or lasers, the energy footprint is not as apocalyptic as feared—especially when compared to the energy cost of brute-force classical alternatives.
As for winners, the most likely scenario is a split between the hyperscalers—Microsoft, Amazon, Google—who control delivery via cloud platforms, and a few standout startups that solve the scalability and coherence challenges in hardware. PsiQuantum, with its CMOS-compatible photonic approach, is one such contender. IonQ and Rigetti could survive as acquisition targets or niche providers. In the near term, we’ll see quantum used more in simulation and research than mainstream enterprise applications—but long term, its positioning as an enabler for advanced AI and cryptographic workloads is clear. Quantum won’t replace the data center. It will quietly reshape its deepest layers.