As I write these words, I have just returned from a conference organized by something called the BRAIN Initiative (BI). This Initiative and the conference itself were spearheaded by the National Institutes of Health (NIH), an organization that CIR DeepTech wouldn’t usually cover. Nonetheless, I am delighted to be able to report that BI clued me into several new quantum and AI technologies of considerable interest. Both AI and quantum are fascinating in themselves but to sustain them as commercial activities and justify massive investments they need to find something useful to do.
In this month’s CIR DeepTech I will also add another interesting application for quantum that fits into the “emerging applications” category: not neuroscience but citizens speaking truth to power companies over the matter of who pays for the elephantine centers that keep popping, especially it seems in my home state of Virginia. Here quantum, deployed in just the right way stands a better than evens chance of making a big dent in power bills.
AI and quantum technology have no shortage of extraordinary hardware, ambitious roadmaps and companies promising computational capabilities that classical computers cannot match. Quantum has lacked is a sufficiently large collection of applications capable of turning those machines into economically indispensable infrastructure.
AI has the opposite problem. It has applications multiplying faster than the industry can count them—but those applications increasingly demand enormous quantities of computing, electricity and capital.
Before I get into all this really cool stuff, I wanted all you lucky people know that I am going back into the begging business again. With this month the CIR DeepTech paywall will rise once more. For those of you who think I might have something interesting to say, there will now be, once again, a fee for finding out. And for those of you for whom CIR Deeptech has not yet become essential reading, let me tempt you to pay (just $50 per year), with the news that beyond this month’s paywall is lies information on new quantum and AI applications. But you still can get a taste of this without paying.
CIR Reports to consider buying right now: In the next week, we will have up the table of contents for our new report on Entangled Networks. The 2025 version of this report was one of best sellers and we are already selling the new version of key CIR clients. You can purchase our new Entangled Networks report here. And check back soon at the same URL soon for more details.
Meanwhile, we have two forthcoming reports that look at opportunities in (1) so-called AI Factories and (2) in data centers that emphasize the the overlap of HPC and AI chip markets . The AI Factory report examines the transformation of AI infrastructure from the server-based data centers of today into integrated production systems designed to “manufacture” computational intelligence at scale. The AI and HPC chip report begins by noting that AI has made accelerators the principal source of industry expansion, while scientific computing, simulation, national labs and weather forecasting continue to support demand for
conventional HPC systems. Meanwhile, the HPC and AI processor markets are becoming less distinct.
NeuroAI and Precision Molecular Circuit Therapies
The organizers of BI presented a four-part Neuroscience Research Roadmap. The two parts that I think should be of special interests to the readers of this Substack are a little something that they are calling NeuroAI, which they are defining as “advanced neuroscience and AI based on “theoretically grounded and brain-inspired approaches. And for those of you that may think this is a bit vague, note that the BI researchers are stating that “the limitations of conventional AI have [already] become clear in the domains of science and health. They also noted that “neuromorphic hardware are ready to enable real world implementations of brain inspired computing.”
The other part of the Roadmap that should be of some interest is Precision Molecular Circuit Therapies. This more grounded and is described “targeting neural networks (i.e., today’s AI) with molecular precision for transformative brain therapeutics.
Neural Probes and Interconnects
One of the main reasons I went to BRAIN was to hear more about neural probes and interconnects. As a boy who grew up on Asimov and Clarke, such interests were inevitable, of course. But to be honest I thought that neural interconnects had come further technologically. I imagined thin wires going smoothly and directly into the skull of some unfortunate rodent making neural interconnects distant cousins of co-packaged optics (CPO) than neurons.
But the subjects of these experiments look more like Alex in “A Clockwork Orange” with large external “head cages” than what I was imagining were the early stages of man and machine. The architecture of the neural probes that I saw were filed under “non-invasive” which made it all OK, I suppose.
Meanwhile, industry now needs hard economic benchmarks. Not “our quantum computer solved this problem.” Not even “our quantum computer solved this problem faster.” The real question is whether our quantum system beat the best available classical method and did the comparison include the entire quantum system, how much did the complete calculation cost and how much energy did it consume? And most importantly, did anybody make or save money?
Until those questions can be answered convincingly, quantum AI remains just an emerging technology rather than a revolution. AI data centers and other locations need to recognize, for example, that a computer drawing substantial power for several seconds may be economically and environmentally preferable to a supposedly efficient machine running thousands of processors for hours.
That is why nobody should presently construct an AI data center on the assumption that quantum computing is going to slash its electricity bill. The evidence isn’t there yet. If the answer eventually becomes yes, the quantum industry has acquired a market almost unimaginably larger than the quantum-computing market contemplated today.
For more insights on new and meaningful quantum and AI technologies, subscribe to the full CIR Deeptech on Substack.
