Quantum computers keep hitting the same wall. It is not only a shortage of qubits. It is getting separate chips to talk to each other fast enough to act like one machine. IonQ reports it has pushed on that wall with a faster photonic interconnect.
Working with Chris Monroe’s group at Duke University, IonQ researchers linked a trapped-ion qubit to a diamond-based quantum memory. They report generating entanglement between the two more than 1,000 times every second, which IonQ says is more than four times faster than the previous trapped-ion record. That record was held by Monroe’s Duke group. The team posted a preprint on October 7, and The Quantum Insider covered it two days later. Monroe is also IonQ’s co-founder and chief scientist, so the record holder and the new team overlap.
Why It Matters to You
This is not a chip landing on your desk next year. Picture a nurse we’ll call Devon, an illustrative example, not a real person, at a small hospital that waits on drug discovery relying on heavy molecular simulation. Linked quantum chips might one day help with that kind of work, but nobody has shown a networked quantum computer solving a real chemistry problem. For now this is one engineering step, and IonQ has begun selling the technology: the University of Maryland made the first commercial purchase of the memory and interconnect platform, announced in April, and South Korea’s SDT signed for a second system in September.
How the link works
A single trapped barium ion holds a qubit. It emits a photon that carries quantum information, and that photon is converted in wavelength, reflected off a silicon-vacancy center, and detected. The silicon-vacancy center is a tiny defect inside a diamond cavity that works like a memory for light. Detecting the photon signals that the two qubits are entangled, meaning they share a quantum state.
The full technical paper reports an average rate of 1.03 kHz and a fidelity of 87.9 percent, with all conversion losses included. That is about one entangled pair per millisecond, which the authors say matches the cycle time of planned trapped-ion processors. A link slower than the processor it serves becomes a bottleneck.
One route to bigger quantum computers is to build smaller, cleaner chips and link them with light, the way data centers link racks of servers. IonQ CEO Niccolo de Masi said in the company’s release that “quantum systems will scale in much the same way.” Monroe said moving qubits through photons “will be necessary in any large-scale quantum computer.” This is a different problem from IonQ’s earlier quantum error decoder, tested in simulation, because it is about getting separate modules to cooperate at all. IonQ also says it works with DARPA’s HARQ program, which seeks fast quantum interconnects that work with several qubit types.
What the numbers leave out
The figures come from IonQ’s press release and a company-written preprint that has not been peer reviewed, and no outside lab has repeated the result. The 4x comparison is to one earlier trapped-ion link. IonQ also calls this the fastest interconnect rate “of any platform,” but only the trapped-ion comparison is cited, so the broader claim rests on IonQ’s word. Fidelity is 87.9 percent, while the Innsbruck group reported 95.9 percent over 1.2 kilometers in a different setup, per PostQuantum, so speed here comes with lower accuracy than some other links. The test used trapped ions only; work with neutral atoms or superconducting qubits is expected, not shown. It was one link between one ion and one memory, not a networked computer running a program, and PostQuantum estimates modular systems will need thousands of high-fidelity pairs per second. The gap between a lab record and that target is unresolved. Fidelity is the number to watch next, since speed without accuracy just moves errors around faster.
| Link | Reported rate | Reported fidelity | Source |
|---|---|---|---|
| IonQ and Duke: trapped ion to diamond memory (2026) | 1.03 kHz | 87.9% | arXiv preprint |
| Duke: two trapped ions (2024 PRL) | About 250 pairs per second | Not reported | PostQuantum |
| Innsbruck: multiplexed link over 1.2 km | 4.59x speedup (base rate not given) | 95.9% (plus or minus 1.5) | PostQuantum |
These links use different setups, so the table compares reported figures, not like-for-like performance.
Two questions
Is it really four times faster than before?
The preprint says it is four times the rate of the fastest photonic link between two trapped ions, and IonQ’s release makes the same comparison. The earlier record came from Monroe’s own Duke group. The paper is an arXiv preprint submitted October 7, 2026, and has not been peer reviewed.
Do networked quantum computers exist now?
Not yet. IonQ entangled one trapped barium ion with a silicon-vacancy center in a diamond cavity about 1,030 times per second, at 87.9 percent fidelity. That was a single link in a lab setup, and compatibility with other qubit types is expected but not demonstrated.
Sources
- IonQ: IonQ Demonstrates World-First Quantum Memory-Enhanced Interconnect (Oct 09, 2026)
- arXiv: Quantum Networking at the Speed of Quantum Computation (Oct 07, 2026)
- The Quantum Insider: IonQ Demonstrates 1,000 Entanglement Events per Second (Oct 09, 2026)
- PostQuantum: IonQ Photonic Interconnect and Networked Quantum (Apr 20, 2026)
- IonQ: Photonic Interconnect Milestone, Networked Quantum Systems (Apr 14, 2026)
