IonQ (IONQ -0.57%) and Rigetti Computing (RGTI -1.99%) are taking entirely different technological routes in their efforts to build useful quantum computers.
IonQ uses trapped-ion technology, where individual charged atoms held in place by precise electromagnetic fields serve as qubits, the basic units of information in any quantum computer. Rigetti Computing uses superconducting qubits, which are tiny electrical circuits cooled to extremely low temperatures.
Image source: Getty Images.
IonQ has demonstrated more than 99.99% fidelity, or accuracy, for 2-qubit gates (operations performed on 2 qubits at the same time). Rigetti Computing has reported roughly 99.1% median 2-qubit fidelity on its 108-qubit Cepheus system. Hence, its typical 2-qubit operation succeeds about 99.1% of the time.
While IonQ may appear to be the obvious technology winner, these figures are not directly comparable. IonQ's 99.99% result came from testing a particular 2-qubit operation, rather than measuring the overall performance of a large commercial quantum computer. A more comparable measure comes from IonQ's Forte systems, which have reported a median 2-qubit fidelity of 99.3% to 99.5%. Hence, although IonQ's qubits still appear to perform operations more accurately than Rigetti Computing's, the gap in accuracy between them is much smaller.
IonQ's quantum computing technology
That narrower accuracy gap could prove crucial when quantum calculations become more complex. Using the more comparable figures above, 99.5% fidelity implies an error rate of roughly 0.5%, compared with about 0.9% at 99.1%. Hence, Rigetti Computing's error rate is roughly 1.8 times as high in this comparison.
NYSE: IONQ
Key Data Points
IonQ also has an advantage in how its qubits communicate. The company's trapped-ion systems offer all-to-all connectivity, meaning any qubit can directly interact with another qubit in the system.
Quantum computers with more limited connectivity may need additional operations, known as SWAP gates, to move quantum information between qubits that cannot directly interact. These extra operations can introduce more errors.
IonQ's connectivity can therefore reduce the number of operations needed for some quantum calculations. This can partly offset IonQ's slower quantum gates, which are the basic operations performed on qubits. In one published comparison, IonQ's Forte systems took about 950 microseconds for a 2-qubit gate, while Rigetti Computing's 108-qubit Cepheus system performed such gates in about 60 nanoseconds.
IonQ also needs to prove that the advantages of trapped-ion technology persist as its quantum computers become larger. The company's next-generation Superion 256 platform is designed around 256 qubits. IonQ has already manufactured 256-qubit Superion processor chips at SkyWater, the semiconductor foundry it acquired in July. The company has also successfully trapped ions on prototype versions of the system. However, these systems are not yet being delivered to customers; commercial deliveries are expected to begin in 2027.
In short, the Superion 256 is an important milestone, but the real test for IonQ will be whether its hardware can maintain high accuracy and connectivity as the number of qubits increases.
IonQ has also made progress in quantum error correction, which combines multiple physical qubits to create more reliable logical qubits. In one experiment, the company used 18 physical qubits to encode 4 logical qubits. Some of those logical qubits preserved quantum information for about as long as, or slightly longer than, the individual physical qubits. Researchers call this breakeven performance. This is still far from a fault-tolerant quantum computer, but it shows that IonQ has begun demonstrating its error-correction approach on actual hardware rather than only through simulations.
DARPA -- the U.S. Defense Advanced Research Projects Agency -- has also advanced IonQ to Stage B of its Quantum Benchmarking Initiative. At this stage, the agency is evaluating the company's development plan, technical risks, and the prototypes needed to reduce those risks.
Rigetti Computing's quantum computing technology
Rigetti's much faster quantum computing operations could give it a major advantage if it can improve accuracy while building larger systems.
NASDAQ: RGTI
Key Data Points
The company has already made progress on the scale front. Its 108-qubit Cepheus system combines 12 interconnected 9-qubit chiplets into a 108-qubit processor. Chiplets are smaller processor blocks that can be connected to build larger systems. Rigetti Computing says this approach can help it increase the qubit counts of its quantum computers while avoiding the manufacturing complexities of building increasingly large chips.
However, Rigetti Computing has reported a median 2-qubit fidelity of 99.8% on its 9-qubit system and 99.6% on its 36-qubit system. The figure fell to roughly 99.1% on the 108-qubit Cepheus system. So while it has demonstrated that it can build larger quantum computers, maintaining accuracy as qubit count increases remains a challenge.
Management has set a goal of developing a system with roughly 1,000 qubits, 99.9% 2-qubit fidelity, and 2-qubit operations that take less than 50 nanoseconds, and given itself a timeline of approximately three years. If Rigetti Computing can build a single quantum computer that achieves all of those milestones, its combination of speed, accuracy, and scale could significantly narrow the gap between its tech and IonQ's.
Which company has the technology edge?
I think IonQ has the technology edge today. The company has moved beyond building more physical qubits to demonstrating error correction on actual hardware.
However, its lead is not secure. The Superion 256 is still in the prototype stage. IonQ still needs to prove that its accuracy and connectivity can hold up as its systems become much larger. Rigetti Computing could narrow the gap if it improves accuracy while preserving its existing advantages in speed and scalability.
