Building next-generation quantum computers
Throughout its 65-year history, computing has enjoyed a steady increase not only in processing power, but also in its useful application. From the first semiconductor transistors of the 1950s to today’s billion-transistor chips, more speed has meant more function—right up to climate-modeling supercomputers and power-efficient wearables. But quantum computing enjoys no such benefit. It is not enough for a quantum computer to work on its own terms; to be truly useful, it must outperform a classical supercomputer. “Quantum computers will do amazing things once they are built, but they have not been built yet,” says Tim Menke, COO of Atlantic Quantum. Atlantic aims to change that, by tackling the singular challenge of the field: building a scalable, fault-tolerant quantum computer.
Emerging out of MIT and Sweden’s Chalmers University of Technology, Atlantic was founded on a key insight: the way people are building qubits today is not scalable. Today’s most commonly used transmon qubits are simple to build, but need high-frequency control sources—which, practically, means each qubit has to have its own classical microprocessor. The result is a physically complex and error-prone infrastructure, with control wires bridging room-temperature and cryogenic environments. Like the gargantuan early vacuum tube computers, these transmon qubit quantum computers can only scale with a brute-force approach to engineering—requiring endless racks of equipment. “In order to tackle meaningful applications, we’re going to need to get to the million-qubit regime, but a million classical computers for one quantum computer just doesn’t work,” says CEO Bharath Kannan.
“We have this fundamentally different idea of how to do it,” says CTO Simon Gustavsson. Atlantic replaces the transmon qubits most commonly used by today’s aspiring quantum computer builders with novel fluxonium qubits, which are simpler to control and simpler to scale. They add slight complexity to the quantum circuit, but in doing so reduce its operating frequency—opening the door to streamlined control infrastructures, as well as lower error rates. Quantum computing depends on error rates low enough to support error correction algorithms, which in practice requires high “coherence,” or the time it takes for information to degrade. Atlantic’s fluxonium qubit lengthens that coherence time by an order of magnitude beyond today’s transmon qubits, creating a new roadmap towards the creation of more efficient and scalable logical qubits. “What we're building solves the error rate and scalability problems, not individually, but simultaneously,” says Kannan.
Just as the transistor’s replacement of the vacuum tube opened the door for decades of continuous innovation in classical computing, Atlantic’s paradigmatic hardware shift opens a new path to building quantum computers powerful enough to meet their extraordinary potential—for healthcare, finance, logistics, and other fields.
Atlantic Quantum’s core IP emerges from one of the world’s top quantum computing labs, William Oliver’s Engineering Quantum Systems (EQuS) group at MIT. The result of nearly two decades of work by Oliver, along with co-founders Gustavsson and Jonas Bylander, the Atlantic team has demonstrated its fluxonium qubit technology, setting a new standard for qubit quality, with record gate fidelities and coherence times. They are joined by co-founders Kannan, Menke, and Youngkyu Sung, all recent EQuS PhDs. Atlantic is fabricating its chips at a lab and manufacturing facility within the quantum computing center at Chalmers, leveraging the US and Sweden’s cooperation in quantum information science and technology. At MIT, Atlantic is using The Engine’s Cambridge lab space to build out its quantum computer infrastructure.
Despite the incredible capability of today’s classical computers, there remain problems they cannot efficiently tackle. For drug discovery, optimization problems, logistics puzzles, and financial models, quantum computing promises a transformative advancement—if a quantum computer of sufficient scale can be built. ”We have all these great applications, but at the end of the day you need an actual computer to run them on,” says Kannan. “We want to actually build that—and once that is done, everything else will fall into place.”