IBM is one of the companies jockeying for position in the nascent field of quantum computing and the company thinks it has a roadmap to develop a universal quantum computer with up to 1,000 qubits. Such a chip would be capable of more complex workloads than any existing chip. Furthermore, IBM claims that even this 1,000-qubit CPU, codenamed Quantum Condor, is just the beginning, with long-term plans for a “million-plus” qubit chip at some unspecified point in the future.
To build the Quantum Condor chip and those to come after it, IBM is investing in a dilution refrigerator larger than any previously built to handle the cooling. Dilution refrigerators are capable of cooling devices down to 2 milliKelvin by using helium-3 and helium-4 isotopes. There are two kinds of dilution refrigerator — wet and dry — and it looks as though IBM is building the dry kind based on its work with companies like Blufors, which specializes in cryogen-free dilution refrigerators.
Just this month, IBM released its 65-qubit Quantum Hummingbird system, which offers capabilities like 8:1 readout multiplexing, meaning eight qubit signals are combined to reduce wiring complexity and improve scaling. The company claims that it’ll release Quantum Eagle next year, with up to 127 qubits and innovations including through-silicon vias (TSVs) and multi-level wiring to support fanning out “classical” control signals while keeping the qubits in a separate, protected layer. Quantum Eagle will implement a new topology IBM calls “heavy hexagon,” and it includes a form of built-in error protection that IBM believes will be essential as the number of quantum qubits in a system scale upwards. You can read more about the heavy hexagon concept at the preceding link — IBM released the information on this just two weeks ago, so the data is still current.
After Eagle comes the 433-qubit IBM Osprey, aimed for 2022. Osprey’s advances are not well-detailed, but the IBM brief states: “More efficient and denser controls and cryogenic infrastructure will ensure that scaling up our processors doesn’t sacrifice the performance of our individual qubits, introduce further sources of noise, or take up too large a footprint.” The implication here is that Osprey will scale up the advances found in Eagle, but will focus on improvements that support this scale-up rather than implementing additional and radical changes to the core architecture.
That brings us to 2023 and the release of the IBM Quantum Condor, a 1,123 qubit device named after a huge scavenger bird that urinates on itself as a way to cool down. If IBM plans to introduce some kind of radical cooling solution with the 1,123-qubit system, the company hasn’t announced it yet, so we’re assuming the name was chosen for sounding cool rather than to imply adopting the condor’s chosen method for thermal regulation.
“We think of Condor as an inflection point, a milestone that marks our ability to implement error correction and scale up our devices, while simultaneously complex enough to explore potential Quantum Advantages—problems that we can solve more efficiently on a quantum computer than on the world’s best supercomputers,” IBM writes.
The nod to Quantum Advantage is an acknowledgment that modern quantum computers, despite being theoretically capable of executing certain computations decades to millennia faster than a classical computer, are still in their infancy. Google announced last year that it had achieved Quantum Advantage (also known as Quantum Supremacy), but we have yet to achieve it in the sorts of problems that most interest scientists. Quantum computers are in the process of becoming complex enough for useful work, and that’s the kind of long-term goal IBM is hoping to achieve with its various Quantum birds.
IBM concludes its announcement by noting it has built a 10-foot tall, six-foot-wide “Goldeneye” dilution refrigerator intended to one day house a million-qubit machine. The company writes, “Ultimately, we envision a future where quantum interconnects link dilution refrigerators each holding a million qubits like the intranet links supercomputing processors, creating a massively parallel quantum computer capable of changing the world.”
Rather like a mainframe, one might say. At IBM, the future may look a heck of a lot like the distant past, at least where centralized computing is concerned. IBM is just one company pushing into quantum — Google, Intel, and D-Wave all compete in this space, though D-Wave’s machine is a quantum annealer, not a universal quantum computer. If IBM can keep its timeline, we should be able to see what quantum computers can practically achieve by the mid-2020s.
- Intel to Speed Up Quantum Computers With Cryogenic Chip
- How to Make Sense of Google’s Quantum Supremacy Claim
- IBM Preps 53-Qubit Quantum Computer for Launch in October