In the technological map, are there still borders that separate quantum computing and artificial intelligence?
The comparison came swiftly, and the question is now spreading worldwide: is the fast development of Artificial Intelligence (AI) leaving quantum computing behind? At a time when the effective use of quantum machines for solving real problems remains uncertain — with some predicting at least two more decades for this to happen -, and when AI is advancing rapidly with an impact on different sectors of activity (with others stating that it may even solve issues that quantum machines would solve), the boundaries that separate them seem increasingly blurred. Are we facing a competition or an opportunity for collaboration with benefits for both domains?
January 7, 2025. Las Vegas, United States of America. Jensen Huang, founder and CEO of NVIDIA, gives a presentation on the advances of AI at CES — one of the major technology events in the world, which brought together more than 141,000 attendees this year. Before an audience of thousands, he projected the effective use of quantum machines to be two decades away.
This prediction by Jensen Huang is soon to take place in quantum computing, with losses for companies like Rigetti Computing, D-Wave Quantum, Quantum Computing or IonQ. Losses of around 40% are estimated, namely millions in market value. A blow to an industry that had, just over two months ago, benefited from Google’s announcement of the willow chip, which seemed to shorten the path to widespread use of quantum machines.
One thing’s for sure: Google was quick to reply, introducing a new timeline: five years. In an interview with Reuters, Hartmut Neven, head of Quantum at Google, said he was optimistic about the emergence of real applications of quantum computers within five years — applications like building superior batteries for electric cars, developing new drugs or alternative energies.
From the debate over how long it will take for a quantum machine to solve a useful problem, a comparison with AI emerges — whose advances are undeniable. Is AI outpacing quantum computing? Could it ever surpass or even replace it? Or does one’s progress benefit the other’s advances? Could using quantum machines to generate data for training LLMs (Large Language Models) be a path to collaboration? Could the fact that quantum machines are not yet being used effectively to solve real-world problems open the door for AI to do so instead? There are still many questions.
A few months before Jensen Huang’s statements, Demis Hassabis, Nobel Prize winner in Chemistry and CEO of Google DeepMind, explained why he believed classical computers are capable of much more than initially imagined. He stated that traditional systems, when used correctly, could potentially model much more complex systems — perhaps even quantum systems, in a counterintuitive way. He doesn’t say this without acknowledging that it’s a controversial idea, but he emphasises the need to seriously consider the possibility that traditional systems could model quantum systems.
After all, is AI gaining ground? An opinion piece published in the Financial Times, written by editor Richard Waters, raised the question of whether the fast development of AI is leaving quantum computing behind, and concluded with uncertainty about how the two might be integrated. In this edition of INESC TECWatch, we invited three INESC TEC researchers — Susana Vitória Marques, Alexandra Ramôa and Luís Paulo Santos — to analyse the evolution of the two areas, the borders that still separate them and what the future may hold. Competition or collaboration? That’s what we’re going to look at next.
Are they close or far apart?
Alexandra Ramôa, researcher at INESC TEC and PhD student at the University of Minho, believes that AI and quantum computing do not have much in common apart from being up-and-coming technologies — “and as any other promising technologies, they are more likely to fruitfully coexist than to compete with one another.” The opinion is shared by Susana Vitória Marques, who considers that the technologies do not need to be in opposing fields. According to the researcher, who is also a PhD student at the University of Minho, “quantum computing could enhance AI rather than displace it, supplying speed-ups or data-generation techniques that, when combined with AI’s advanced algorithms, yield transformative results. For now, however, AI remains the technology reshaping the world. Unlike quantum computing, it doesn’t require patience — it delivers results today”, she said.
But that wasn’t always the case — the results weren’t always immediate. As explained by Luís Paulo Santos, researcher at INESC TEC and professor at the University of Minho, the evolution of machine learning has been going on for decades — since the 1950s, more specifically -, having originated the expression “IA Winter”. “IA Spring has arrived, with recent developments, as impressive as they are, building on top of the exponential growth of affordable classical computing power,” he stated. According to the researcher, and despite the tremendous impact deep learning already demonstrated, some unknowns still hang on the horizon: will these results scale? Or will further capabilities require exponentially growing resources and/or data? Is Artificial General Intelligence (AGI) within reach of the current approach to AI or will we stay imprisoned within the realms of sophisticated pattern matching machines?
Alexandra Ramôa recalled that natural language models have been around since the 1950s, but have only captured mainstream attention due to recent advances and increased accessibility — they are more common and user-friendly. “The core principles behind AI are decades old, with the recent explosion fuelled by quantitative improvements: increased computing power, access to massive datasets, advances in algorithms, among others. While AI is undeniably interesting, much of its recent hype stems from its immediate appeal to the general public-not necessarily a sound (or consistent) metric for merit.”
Even so, the role that this technology currently plays in our daily lives seems undeniable, as Susana Vitória Marques recalled: “unlike experimental innovations that remain confined to research labs for decades, AI is already woven into daily life.” The researcher also mentioned that AI’s influence extends far beyond chatbots and virtual assistants, presenting important contributions it powers self-driving cars, transforms medicine through personalized treatments, detects financial fraud in real time, and revolutionizes industrial automation. “At its core, AI is fulfilling a universal human desire: making work easier, more efficient, and, where possible, unnecessary. Across industries, it is boosting productivity and reshaping the way we live and work. Speculating about AI’s future is challenging, but one thing is certain: it isn’t standing still,” explained the researcher.
Given the evolution of AI, is Quantum Computing really lagging behind?
“All incremental innovations are alike; each paradigm-shifting innovation is paradigm-shifting in its own way. Transformative technologies bring both unique challenges and unique rewards — and reaping the latter requires patience towards the former. This is definitely the case for quantum computing, which some claim lies in the shadow of AI’s recent ‘explosion’. As a recent piece on Financial Times put it, ‘it’s hard to focus on the more distant and unproven benefits of quantum machines when the headlong rush of AI dominates the headlines’. But the reason for this distance between us and practical quantum machines is no less than their groundbreaking nature; and it is in this very nature that their potential resides,” — said Alexandra Ramôa.
According to the researcher, more than comparing quantum computing to the recent evolution of AI, quantum computers can be compared to the advent of digital computers, at a time when there were significant doubts about their true potential. “Throughout history, revolutionary advancements have often been met with scepticism”, she mentioned, highlighting that developing technology unlike anything that has been done before takes longer than evolving pre-existing technology.
One cannot deny that quantum computing evolved over the last few years, although its application to real problems remains uncertain. “Quantum computing has also witnessed tremendous developments, but is still far from demonstrating practical results, which clearly separate it from results that would be achievable using classical computing resources. It is still unknown whether this quantum advantage is possible. And even if it is, whether it is practical and economically worthwhile. This question can probably be understood as being at the same level as whether AGI is achievable,” said Luís Paulo Santos.
According to the researcher, “if and when quantum computing becomes practical, then machine learning is one of the areas that are expected to benefit. Can quantum machine learning learn models that are in some sense more effective or efficient than current classical models? Can these models be more compact? Can they generalise better? Can they be trained more efficiently or even in circumstances where classical ones would not learn?”
“The problem is not whether AI can solve problems that are believed to be efficiently solved by resorting to the quantum computing paradigm. The problem is whether quantum computing can be made practical to meet some, or all, of its promises. And then, how can it contribute to the evolution of many problems faced by humankind, including the development of so called intelligent systems,” mentioned Luís Paulo Santos.
Alexandra Ramôa also presented a question: “But what can quantum computing do that AI cannot? Quantum computing, in contrast to AI, is not a by-product of classical computing, but rather a fundamental departure from it. Quantum machines operate on an entirely new regime of physics, which can exploit entirely new phenomena to unlock computational speed-ups for certain tasks.”
Still, Susana Vitória Marques recalled that AI is already capable of solving certain problems that, at first, required more advanced computational capabilities — and shows no signs of slowing down. “While some predict that quantum computing will usher in a new era of computational power, that assumption rests on AI remaining static,” she mentioned, claiming that the true potential of AI is only beginning to be grasped. “What we see today — language models capable of near-human conversation, autonomous systems making split-second decisions, and AI-driven tools optimizing entire industries — is just the surface,” stated the researcher
In her opinion, technology isn’t merely advancing: it’s accelerating, constantly redefining the boundaries of what machines can accomplish. “While today’s AI is impressive, its future promises breakthroughs that could far exceed even our boldest expectations,” she concluded.
Returning to the initial question, the boundaries seem to be — for now — well-defined, with different technologies advancing at their own pace and with many questions still unanswered as to what the future may bring to each of them, knowing that their impact and potential are undeniable — as well as the idea that collaboration, more than competition, could mean gains for both technologies.
