How nanotechnology enables AI, CRISPR, and quantum computing

(Nanowerk Spotlight) As we close the final chapter of 2025, this Nanowerk Spotlight article marks a significant milestone in our coverage. We find ourselves standing at the midpoint of the decade: the ‘watershed year’ where the speculative promises of the early 2020s have matured into the industrial infrastructure of the late 2020s.

The first half of this decade was defined by a rapid acceleration in digitalization and the proof-of-concept phase for several frontier technologies. Since 2020, we have witnessed the dramatic rise of generative AI, the clinical validation of mRNA vaccine delivery and lipid nanoparticle drug delivery platforms, and the first industrial-scale qubit arrays. What was once considered the fringe of material science has now reached a commercial turning point.

As we pivot toward 2026 and the second half of the decade, the focus is shifting from discovery to convergence. The articles we featured in our 2025 Nanowerk Spotlight series have increasingly highlighted this trend, showing that the most profound breakthroughs no longer happen in isolation but at the collision points between disciplines. In this final feature of the year, we explore how this Convergence Engine is not just a trend for 2025, but the blueprint for the remainder of the decade.

In late 2023, a robotic laboratory at Lawrence Berkeley National Laboratory accomplished something that would have been unthinkable a decade earlier. The A-Lab, developed in collaboration with Google DeepMind, synthesized 41 novel inorganic compounds in just 17 days, operating around the clock without human intervention. Machine learning algorithms predicted which materials might be stable, proposed synthesis recipes by parsing decades of scientific literature, and robotic arms executed the experiments. When a synthesis failed, the system analyzed why, adjusted parameters, and tried again.

Casgevy is an ex vivo therapy: a patient’s stem cells are extracted, edited outside the body using electroporation, and then reinfused. This approach works because hematopoietic stem cells can be safely removed and returned. But for the vast majority of tissues—the brain, lungs, liver, heart—extraction is impossible. Reaching those cells requires in vivo delivery: injecting gene-editing machinery directly into a living patient and guiding it to its target.

These two breakthroughs, one in autonomous materials discovery, the other in genetic medicine, share a common thread: nanotechnology has transitioned from a vertical research discipline into a horizontal enabler. It is no longer simply about making things small; it is about providing the physical substrate on which other frontier technologies depend.

The relationships differ in character. In the case of AI, the connection is synergistic: machine learning accelerates nanomaterial discovery, while nanoscale hardware enables more efficient computation. For biotechnology, nanotechnology plays a foundational role: LNPs and other nanocarriers are the delivery mechanisms without which gene therapies cannot function. And for quantum computing, nanotechnology serves as the construction discipline: qubits are nanoscale devices, and the path to millions of them runs through semiconductor nanofabrication facilities.

In the following sections we explore each of these convergence modes and why understanding them matters for researchers, technologists, and anyone trying to make sense of where nanotechnology is heading.

The relationship between artificial intelligence and nanotechnology is bi-directional, characterized by a loop between digital optimization and physical innovation. AI serves as a catalyst for the discovery and precise tuning of novel nanomaterials, while these newly engineered materials provide the physical basis for non-silicon computing architectures that run AI models far more efficiently than traditional processors. This creates a feedback loop where breakthroughs in material discovery directly enable the hardware required to amplify the next generation of AI.

The A-Lab team has disputed these conclusions, providing additional spectroscopic evidence they argue supports the original synthesis claims.

These findings do not invalidate the promise of autonomous materials discovery, but they underscore that AI systems require significant refinement before they can reliably replace expert human analysis in solid-state chemistry. The demonstration of closed-loop autonomous synthesis nonetheless remains a landmark in the field, pointing toward a future where such platforms could dramatically accelerate materials discovery.

This creates the feedback loop: AI designs better nanomaterials, including the materials used to fabricate more efficient AI chips, which in turn enable more powerful AI systems for materials discovery. The virtuous cycle is now accelerating.

COVID-19 mRNA vaccines demonstrated the clinical viability of lipid nanoparticle delivery at unprecedented scale. The Pfizer-BioNTech and Moderna vaccines relied on ionizable lipid nanoparticles to protect fragile mRNA molecules and facilitate their uptake into cells. That success validated a platform that researchers are now extending to more complex payloads.

Casgevy, the CRISPR therapy approved in late 2023, uses an ex vivo approach: a patient’s stem cells are extracted, edited outside the body, and reinfused. But the next frontier is in vivo editing: delivering gene-editing machinery directly to cells inside a living patient. This is where LNPs become essential.

When delivered via tissue-selective LNP formulations, the system achieved 16-37% genome editing in the liver and lungs of mice with a single intravenous injection. The work also demonstrated editing of the disease-causing SFTPC gene in lung tissue at 19% efficiency, representing a major advance toward treating pulmonary genetic disorders.

One of the central challenges in LNP therapeutics is organ targeting. Intravenously administered LNPs tend to accumulate in the liver, where they are taken up by hepatocytes and cleared by the reticuloendothelial system. For liver diseases, this is advantageous. For diseases affecting the brain, lungs, or immune cells, it is a barrier.

These results illustrate a broader principle: the combination of CRISPR’s precision with LNP’s delivery capabilities enables therapeutic interventions that were previously impossible. The nanoparticle is not merely a carrier, it is a targeting system that determines which cells receive the genetic payload.

This has significant implications. Diraq’s CEO, Andrew Dzurak, noted that achieving these benchmarks in a foundry environment “opens a cost-effective pathway to chips containing millions of qubits.” Unlike superconducting qubits (used by IBM and Google) that require specialized fabrication, silicon spin qubits leverage 60 years of semiconductor manufacturing development and trillions of dollars of existing infrastructure.

The work is early-stage and practical quantum computers based on this approach remain years away. But it demonstrates that the cooling requirement is a materials and fabrication challenge, not a law of physics. If nanoscale engineering can stabilize quantum states at higher temperatures, the path to scalable quantum systems becomes substantially clearer.

The transition from tens of qubits to millions is fundamentally a manufacturing problem. Current error correction schemes require thousands of physical qubits to produce a single error-corrected logical qubit, which is the kind capable of performing reliable computation. Achieving this at scale demands not only high-fidelity individual qubits but also precise control over their interactions, uniformity across large arrays, and integration with classical control electronics.

French startup Quobly secured €21 million in 2025 to industrialize a 100-qubit silicon processor built on 300mm FD-SOI wafers, targeting production readiness by 2027. Quantum Motion, a London-based company, partnered with GlobalFoundries to fabricate chips with over 1,000 quantum dot sites using commercial 22nm FD-SOI processes.

These efforts share a common assumption: the semiconductor industry’s nanofabrication capabilities are not obstacles to quantum computing but assets. The question is whether qubit performance can be maintained as manufacturing scales up, and the 2025 results from Diraq and Intel suggest the answer is increasingly yes.

The three convergence pairs examined here represent different modes of interaction, but they share a common thread: nanotechnology operates at the scale where the abstractions of other fields meet physical reality.

Machine learning algorithms can predict material properties, but the materials must still be synthesized, characterized, and optimized at the nanoscale. Gene-editing systems can rewrite DNA with remarkable precision, but the editing machinery must physically enter cells, navigate the cytoplasm, and reach the nucleus. Quantum algorithms can theoretically solve problems intractable for classical computers, but qubits are physical objects whose coherence depends on atomic-level fabrication quality.

This substrate role has implications for researchers, investors, and policymakers. Understanding where nanotechnology fits in the technology stack helps identify both opportunities and bottlenecks. The stunning results from autonomous laboratories mean little if the materials they discover cannot be manufactured at scale. CRISPR’s therapeutic potential remains constrained by delivery, and the pace of progress in LNP engineering may ultimately determine which diseases become treatable. Quantum computing roadmaps hinge on nanofabrication yields that have only recently begun to be demonstrated.

Electricity did not disappear when it became ubiquitous; it simply stopped being remarkable. The internet did not become less essential when we stopped calling applications ‘web-enabled’; the descriptor became redundant because connectivity became assumed. The same transition is underway for nanotechnology. The questions researchers ask have shifted from ‘can we work at the nanoscale?” to ‘which nanoscale approach best solves this problem?” That is the clearest sign of a maturing technology: when it stops being a specialty and becomes infrastructure.

In 2020, lipid nanoparticles were a niche delivery technology. By 2025, they had delivered vaccines to billions of people and begun crossing the blood-brain barrier. In 2020, autonomous materials synthesis was an academic curiosity. By 2025, the A-Lab had demonstrated closed-loop discovery at industrial pace. In 2020, silicon spin qubits achieved high fidelity only in academic cleanrooms. By 2025, they exceeded 99% fidelity in commercial foundries.

These are not predictions. They are the record of the first half of the decade—and the clearest evidence that nanotechnology has become the common denominator.