Nobel laureate Martinis sets date when quantum AI will start solving unsolvable problems
At the TiEcon 2026 conference, John Martinis, creator of Google's first proof of quantum supremacy, stated that the convergence of quantum computers and AI is happening now, paving the way for solving a new class of problems.
Quantum AI on the threshold: Nobel laureate John Martinis on the convergence that blurs the boundaries of the possible
Introduction
At the TiEcon 2026 conference, held from April 29 to May 1 at the Santa Clara Convention Center (California), a statement was made that could well become a historical milestone. John Martinis — winner of the 2025 Nobel Prize in Physics and the man who gave the world the first proof of quantum supremacy — announced that the fusion of quantum computing and artificial intelligence is no longer a theoretical abstraction. According to him, this convergence is happening right now and opens the way to solving problems that were previously considered fundamentally unsolvable. To understand why this statement made the tech community hold its breath, it is necessary to view it in the context of the scientist's forty-year career, which has twice revolutionized quantum physics.
Event details and timeline
TiEcon 2026 was held under the motto "AI & You: Human Centered, AI Powered," bringing together over 156 speakers — from representatives of OpenAI and Google to Morgan Stanley. However, the central event of the business program was John Martinis's keynote on the main stage. Organizers announced his participation as "epoch-making": for the first time in the conference's history, Nobel laureates took the stage, and Martinis was one of two scientists of such caliber.
The format of his talk was dedicated to the convergence of quantum computers and artificial intelligence — a topic that the scientist himself calls a natural continuation of the trajectory that began in the 1980s. It was then, as a graduate student at the University of California, Berkeley, that Martinis and his colleagues experimentally proved that quantum effects work at the macroscopic level — a discovery that half a century later earned them the Nobel Prize.
A timeline of key milestones in Martinis's career helps to understand the weight of his current statement:
1980s — experiments with macroscopic quantum states on superconducting circuits, laying the foundation for all modern superconducting qubits. "There was a question of whether a macroscopic variable could escape quantum mechanics, and for me, a young student just learning quantum mechanics, it seemed like a fantastic experiment," the scientist recalled in a recent interview.
2014–2019 — leading the Google Quantum AI team, which created the 54-qubit Sycamore processor. In October 2019, Sycamore performed a computation in 200 seconds that, by estimates, would have taken the most powerful classical supercomputer Summit 10,000 years. This event went down in history as "quantum supremacy" and, in the words of then-CEO Sundar Pichai, was "the hello world moment we've been waiting for."
October 2025 — awarding of the Nobel Prize in Physics to John Martinis, John Clarke, and Michel Devoret for "demonstration of macroscopic quantum tunneling and energy quantization in an electrical circuit."
2024–2026 — founding of QoLab, where Martinis is reinventing quantum computing for the third time. This time, through a radically new approach to manufacturing quantum chips, which he describes as "abandoning 1950s technology" in favor of modern semiconductor industry methods.
Concurrently with TiEcon 2026, Martinis confirmed his participation in the DAC 2026 conference (July, Long Beach), where he will deliver a talk titled "From Fundamental Science to Building a Superconducting Quantum Computer."
Impact and significance
Martinis's statement about the convergence of quantum computers and AI is important not so much as a prediction, but as a diagnosis of the current moment. A scientist who understands the physics of every element of a quantum system — from microwave electronics to cryostats — asserts that the barriers separating classical AI and quantum computing are disappearing right now.
From a technical standpoint, the merger is happening on two levels. The first is hardware: quantum processors are beginning to be designed and manufactured using the same approaches as modern semiconductor chips for AI workloads. The second is algorithmic: quantum computers naturally solve problems in quantum chemistry and materials science, which are "stumbling blocks" for classical supercomputers.
"I'm really interested in using a quantum computer to solve problems in quantum chemistry and quantum materials," Martinis noted in a February interview with New Scientist. "This quantum problem is hard to solve on a classical supercomputer due to fundamental difficulties in quantum mechanics. But it is, of course, fundamentally solvable by a quantum computer — you just map the quantum problem onto a quantum computer."
The difference between his approach and the mainstream discourse on quantum AI is telling. While many talk about optimization problems and quantum machine learning in the subjunctive mood, Martinis prefers to rely on problems for which the advantage of the quantum approach is mathematically proven: "The theory behind applications in materials science and chemistry is much more definite. We know what size quantum computer is needed. It's a machine that, I think, we can build — both in terms of size and speed."
For the industry, this means shifting focus from "let's try and see if it works" to engineering-feasible projects. For society, it means the prospect of accelerated drug development, new materials, and catalysts — which Martinis sees not as a distant fantasy, but as a goal achievable by solving specific hardware problems.
Reaction of key players
The organizers of TiEcon 2026 presented Martinis's talk as the conference's central event. The program of the track "Powering the Future: Where Nations, Academia & Deep Tech Converge" was built around the intersection of quantum technologies, AI, and geopolitics — with participation from diplomats, academic pioneers, and tech company leaders from the US, India, Japan, and Europe. This reflects a growing understanding that quantum AI is becoming not only a scientific but also a geostrategic category.
The DAC 2026 conference, which announced Martinis as one of three key speakers (alongside Qualcomm's CTO and a UC Berkeley professor), confirmed that chip design is entering an era "from silicon to systems," where quantum accelerators are considered part of the overall computing architecture.
The professional community itself continues to debate the realism of timelines. In an interview with New Scientist, Martinis admitted that his new approach at QoLab met with "surprisingly much resistance and skepticism," but quickly added: "From my decades of experience in physics, that means we have a good idea." He consistently criticizes the "naivety" of many teams trying to scale quantum computers without rethinking fundamental approaches to manufacturing and qubit wiring. In his view, the "jungle of wires" problem is the main barrier, and the solution lies in integrating all microwave control directly onto the chip.
Forecast and conclusions
Martinis's words carry weight precisely because he does not predict the future — he has already created it twice. In the 1980s, his experiments laid the foundation for all modern superconducting qubits. In 2019, his team proved that a quantum computer can do what a classical one cannot. Now he asserts that the convergence of quantum computers and AI is happening right now — and this is not a prediction, but a statement of the process he observes.
What does this mean practically? In the short term (one to three years) — the emergence of hybrid systems where quantum processes solve quantum chemistry problems and classical AI processes the results. Martinis particularly highlights the application for enhancing nuclear magnetic resonance in chemistry — an "initial application" that, in his estimation, requires a computer of a realistically achievable scale.
In the medium term (three to seven years) — the emergence of quantum computers built using the QoLab method: with chips manufactured on modern semiconductor equipment rather than mid-20th-century technology, and with integration of all microwave control onto the chip. This should solve the "wires" problem — the main barrier to scaling, according to Martinis.
Long-term (seven years and beyond) — quantum accelerators for problems that today are not even attempted due to computational inaccessibility: design of new materials, optimization of catalysts for the chemical industry, modeling of biological processes.
Martinis does not promise a miracle — he appeals to engineering logic. And in this, perhaps, lies the main difference between his talk and many futuristic predictions: when a person who understands the physics of every element of a quantum system says that convergence is already underway, it is worth listening. The coming years will show whether Martinis's third revolution will be as successful as the first two. But already it is clear: the conversation about quantum AI has moved from the realm of "is it possible" to "when and in what architecture."
— Editorial Team
No comments yet.