I find myself in front of a quantum computer fashioned from atoms and light at the UK’s National Quantum Computing Centre, located on the fringes of Oxford. Spread across a lab table, a sophisticated array of mirrors and lenses encircles a cell the size of a Rubik’s Cube, where 100 cesium atoms are held in a grid layout by a meticulously adjusted laser beam.
This cesium atom configuration is so compact that I could lift it, transport it out of the lab, and place it in the back seat of my car to bring home. However, I’d likely not get very far. It may be small, but it packs a punch—and is thus extremely precious. Infleqtion, the Colorado-based firm that possesses it, hopes the machine’s capabilities will clinch $5 million next week at an event to be hosted in Marina del Rey, California.
Infleqtion is among six teams that have reached the final phase of a 30-month quantum computing competition known as Quantum for Bio (Q4Bio). Organized by the nonprofit Wellcome Leap, it seeks to demonstrate that current quantum computers, despite being messy and prone to errors and far from the large-scale devices engineers aspire to develop, could indeed impact human health positively. Achieving this would mark a considerable advancement in validating the potential of quantum computers. However, for the time being, this potential appears connected with leveraging and enhancing the performance of conventional (or classical) computers collaboratively, forming a quantum-classical hybrid capable of surpassing what classical machines alone can accomplish.
There are two categories for prizes. A $2 million reward will be granted to any teams that can successfully execute a significantly beneficial health care algorithm on quantum computers with 50 or more qubits (where a qubit is the fundamental processing unit in a quantum computer). To secure the $5 million grand award, a team must effectively implement a quantum algorithm that resolves a critical real-world health care issue, utilizing 100 or more qubits. Winners must fulfill stringent performance standards and resolve a health care problem that conventional computers are incapable of addressing—a challenging task.
Despite the daunting nature of the challenge, the majority of teams believe some of this prize money could be theirs. “I feel we’ve got a good chance,” remarks Jonathan D. Hirst, a computational chemist at the University of Nottingham, UK. “We are well within the criteria for the $2 million prize,” states Stanford University’s Grant Rotskoff, whose team is exploring the quantum characteristics of the ATP molecule that energizes biological cells.
The grand prize seems to be a bit less certain. “This is genuinely at the very limit of what’s feasible,” notes Rotskoff. Insiders indicate that the challenge is so formidable, considering current quantum computing technology, that a significant portion of the funds may remain with Wellcome Leap.
With most of the Q4Bio research unpublished and safeguarded by NDAs, and with the quantum computing field already fraught with assertions and counterassertions regarding performance and accomplishments, only the judges will have the authority to determine who is correct.
A hybrid solution
The foundational concept of quantum computers is that they leverage small-scale entities adhering to the principles of quantum mechanics, such as atoms and photons, to emulate real-world processes too intricate to replicate on our usual classical machines.
For decades, researchers have endeavored to create such systems, which could yield insights for the production of new materials, the advancement of pharmaceuticals, and the enhancement of chemical processes like fertilizer manufacturing. However, grappling with quantum entities like atoms is exceedingly challenging. The most ambitious applications necessitate enormous, sturdy machines capable of resisting environmental “noise” that can easily disturb sensitive quantum systems. Such machines are still not available—and it remains uncertain when they will be.
Wellcome Leap aimed to discover if the smaller-scale machines currently at our disposal could be utilized to achieve something—anything—beneficial for health care while we await the advent of robust, large-scale quantum computers. The organization commenced the competition in 2024, providing $1.5 million in funding to each group among the 12 selected teams.
The six Q4Bio finalists have adopted a variety of strategies. Importantly, they have devised clever methods to navigate the limitations of quantum computing. Confronted with noisy, restricted machines, they have sought ways to delegate much of the computational burden to classical processors executing newly formulated algorithms that often surpass previous standards. The quantum processors are then utilized solely for those aspects of the problem where classical methods fail to scale adequately as computations grow larger.
For instance, a team led by Sergii Strelchuk from Oxford University is employing a quantum computer to analyze genetic variability among humans and pathogens via complex graph-based structures. These structures will—hope the researchers—reveal concealed connections and possible treatment routes. “You can view it as a platform for addressing challenging problems in computational genomics,” suggests Strelchuk.
The associated classical tools encounter significant difficulties when attempting to upscale to extensive databases. Strelchuk’s team has created an automated pipeline that helps determine whether classical solvers may struggle with a specific problem, and how a quantum algorithm could be crafted to render the data solvable on a classical computer or manageable on a noise-prone quantum one. “You can accomplish all this before starting to incur computing costs,” states Strelchuk.
In collaboration with Cleveland Clinic, Helsinki’s Algorithmiq has utilized a superconducting quantum computer developed by IBM to simulate a cancer medication activated by specific light wavelengths. “The concept is you take the drug, and it’s present throughout your body, yet it’s inactive, merely lying in wait, until light of a certain wavelength is applied,” explains Guillermo García-Pérez, Algorithmiq’s chief scientific officer. At that moment, it becomes a molecular missile, targeting the tumor exclusively at the site in the body where that light is directed.
The medication with which Algorithmiq initiated its research is already undergoing phase II clinical trials for treating bladder cancers. The simulation, enhanced by quantum computation and refining classical algorithms, will facilitate its redesign for addressing other health issues. “It has remained a specific treatment as it cannot be simulated using classical methods,” remarks Sabrina Maniscalco, Algorithmiq’s CEO and co-founder.
Maniscalco, harboring hopes of winning from the competition, believes the techniques employed to construct the algorithm will show broad utility: “What we’ve accomplished during the Q4Bio program is something exceptional that can reshape how chemistry is simulated for health care and life sciences.”
Infleqtion’s submission, operating on its cesium-driven machine, endeavors to refine the recognition of cancer signatures within medical data. In collaboration with teammates from the University of Chicago and MIT, the company’s researchers have devised a quantum algorithm that analyzes vast data sets such as the Cancer Genome Atlas.
The objective is to detect patterns enabling clinicians to ascertain elements like the likely source of a patient’s metastasized cancer. “It’s crucial to understand the origin because it can influence the optimal treatment,” states Teague Tomesh, a quantum software engineer who leads Infleqtion’s Q4Bio project.
Regrettably, those patterns lie hidden within data sets so extensive that they overwhelm classical solvers. Infleqtion utilizes the quantum computer to unearth correlations in the data that can simplify the computation. “Then we return the condensed problem to the classical solver,” Teague explains. “I’m essentially aiming to leverage the strengths of both my quantum and classical assets.”
The team from Nottingham, on the other hand, is employing quantum computing to pinpoint a drug candidate that could treat myotonic dystrophy, the most prevalent form of adult-onset muscular dystrophy. One team member, David Brook, was involved in identifying the gene responsible for this condition back in 1992. Now, over 30 years later, Brook, Hirst, and others in their group—which includes QuEra, a Boston-based company developing a quantum computer using neutral atoms—have calculated a quantum method in which medications can form chemical bonds with the protein linked to the disease, inhibiting the mechanism triggering the issue.
Measured expectations
The confidence of the entrants may be high, but Shihan Sajeed’s is considerably lower. Sajeed, a quantum computing entrepreneur from Waterloo, Ontario, is the program director for Q4Bio. He believes the error-prone quantum machines the researchers must utilize are unlikely to fulfill all grand prize criteria. “It’s extremely challenging to achieve something with a noisy quantum computer that a classical machine can accomplish,” he articulates.
That being said, he has been taken aback by the advancements made. “At the program’s inception, people were unaware of any use cases where quantum could tangibly influence biology,” he notes. However, the teams have uncovered promising avenues, he adds: “We now recognize the areas where quantum can make a difference.”
Furthermore, the advancements in “hybrid quantum-classical” processing that the participants are employing are “transformational,” claims Sajeed.
Will it suffice to persuade him to part with Wellcome Leap’s funding? That depends on a panel of judges, whose identities are a closely guarded secret to ensure that presentations aren’t tailored to suit a specific style. Yet, we won’t learn the results for some time; the victor or victors will be revealed in mid-April.
If it ultimately turns out that no winners emerge, Sajeed offers some reassurance for the contestants. The fundamental aim has always been to run a useful algorithm on an existing machine, he emphasizes; failing to meet the target does not imply that your algorithm won’t be beneficial on a future quantum computer. “It simply means the machine you require isn’t available yet.”
