Quantum computing is a rapidly emerging technology that utilizes quantum mechanics to solve complex problems faster than classical computers. Researchers are working hard to develop quantum computers that can perform certain computations that are beyond the reach of classical silicon-based computers.
Some of the biggest players in the tech industry, such as Microsoft and Google, along with startups and nation-states, are all racing to develop and scale up quantum machines.
However, they are still facing challenges in making these machines reliable enough in the real world to beat conventional computers consistently. As quantum computing machines have grown in size and power, researchers have faced the challenge of dealing with data errors that arise due to the complexity of the technology.
IBM has recently unveiled the first quantum computer with more than 1,000 qubits – the equivalent of the digital bits in an ordinary computer. The company hopes the new quantum computing chip and machine will serve as building blocks of much larger systems a decade from now.
But, the company has also decided to shift its focus towards making its machines more error-resistant rather than larger.
IBM has been steadily increasing the number of qubits in its quantum-computing chips every year, following a road map that aims to double them annually. Its latest quantum computing processor, called Condor, has 1,121 superconducting qubits arranged in a honeycomb pattern. This chip follows on from their other record-setting, bird-named machines, including a 127-qubit Eagle processor in 2021 and a 433-qubit Osprey last year.
As part of its new tack, the company has also introduced a new chip, called the IBM Quantum Heron, that features 133 fixed-frequency qubits with a record-low error rate. Its newly built architecture offers up to five-fold improvement in error reduction. It is the first in a new series of utility-scale quantum processors with an architecture engineered over the past four years to deliver IBM’s highest performance metrics and lowest error rates of any IBM Quantum processor to date.
Error correction in quantum computing is a critical concept, as it helps to overcome the inherent noise and instability in quantum systems. However, researchers have stated that state-of-the-art error correction techniques require more than 1,000 physical qubits for each logical qubit that performs useful computation. This means that a quantum computer would need millions of physical qubits, making a useful machine very difficult to build.
However, a new error-correction technique called quantum low-density parity check (qLDPC) has recently attracted a lot of attention from physicists. This technique promises to cut that number by a factor of 10 or more, according to a preprint by IBM researchers. IBM is now focusing on building chips that can hold a few qLDPC-corrected qubits in just 400 or so physical qubits and then networking those chips together to form a larger quantum system.
At the annual IBM Quantum Summit in New York, the computer and artificial intelligence technology giant also unveiled IBM Quantum System Two, its first modular quantum computer and cornerstone of IBM’s quantum-centric supercomputing architecture. The first IBM Quantum System Two, located in Yorktown Heights, New York, has already begun operations with three IBM Heron processors and supporting control electronics.
With this critical foundation now in place, along with other breakthroughs in quantum hardware, theory, and software, the company is extending its IBM Quantum Development Roadmap to 2033 with new targets to significantly advance the quality of gate operations. This would enable larger and more complex quantum circuits to be run and help to realize the full potential of quantum computing at scale.
The company aims to reach 5,000 gates with Heron in 2024 and then introduce new generations of processors with higher quality and gate counts. By 2029, they expect to reach a milestone – executing 100 million gates over 200 qubits with its Starling processor that uses the innovative Gross code for error correction. This will be followed by Blue Jay, a system that can execute 1 billion gates across 2,000 qubits by 2033. This innovative roadmap will also demonstrate the technology that will enable the Gross code through l-, m-, and c-couplers, which will be demonstrated by Flamingo, Crossbill, and Kookaburra, respectively.