IonQ hits 1,000 entanglement events per second between a trapped ion and a solid state memory

[AEWThomas47]

IonQ says it has reached entanglement rates above 1,000 events per second between a trapped ion qubit and a silicon vacancy quantum memory, The Quantum Insider reports. The two are linked by a photonic interconnect. The company says that is more than four times faster than the previous trapped ion record, which was set by Chris Monroe's group at Duke University. Duke also collaborated on this new work, so it is a case of the same people beating their own mark

IonQ is calling it the fastest interconnect rate between qubits on any platform. Those figures come from the company, but there is a preprint on arXiv (2610.10705) for anyone who wants to dig into the details. Silicon vacancy centres are defects in diamond that can store quantum information and talk to photons well. Pairing them with ions is an unusual mix of technologies

The work feeds into IonQ's role in DARPA's HARQ programme. HARQ is about building high speed interconnects that can link different types of qubits together. IonQ expects the architecture to extend to neutral atoms and, with transducers, to superconducting systems. This demonstration used trapped ions only, though, so that part is still a plan rather than a result

There are some quotable lines in the release. Chief scientist Chris Monroe said moving qubits through photons will be necessary in any large scale quantum computer. Mihir Bhaskar of SkyWater said the work shows a photonic interconnect need not be a bottleneck for distributed quantum computing. CEO Niccolo de Masi compared it with how classical systems grew, saying quantum systems will scale in much the same way

IonQ also mentioned sales of its memory and interconnect platform, one to the University of Maryland announced in April and one to SDT in South Korea announced in September. So this is not just a lab curiosity, it is part of a product line. Interconnects keep coming up in every roadmap story we post here. This week alone we have had Quantum Source's QS-LINK and now this

How important are interconnects compared with building bigger single chips? And do you think mixing qubit types will ever be practical?

IonQ Demonstrates 1,000 Entanglement Events per Second in Quantum Interconnectthequantuminsider.com

[Comet Barrel]

Fourfold over the previous record is a big jump, but context matters. A thousand events per second is still slow compared with how fast qubits lose their information. You need the link to be fast enough that the entanglement is still useful when it arrives

I would like to know the fidelity of those pairs as well. Fidelity is what decides whether it is useful

[Di87]

Linking modules is the only way any of this scales. Nobody is going to build a single trap with a million ions in it

Networks of smaller machines are where everyone is heading. Whoever has the best links wins

Classical data centres went the same way decades ago. Lots of smaller machines wired together beat one giant box

[Golden Crow]

Silicon vacancy memories are a smart pick. They work well with photons and hold information for a decent time when cold. Combining them with ions gets you the best of both, at least in theory

It does add another cryogenic system to the setup, though. Diamond is a lovely material for this kind of thing

[Griffin39]

Preprint first, peer review later. That is normal in physics, but it does mean the numbers have not been checked by outsiders yet

I would wait for the published version before getting too excited. Still a nice result on the face of it. Company press releases always lead with the most flattering number

[DarkMatter55]

Selling the memory platform to universities is a clever business move. It gets researchers using IonQ's kit and builds a community around it

Even if the big machines are years away, this brings in revenue now. Plenty of quantum firms struggle with that. Steady revenue keeps the lights on while the big machines get built

[Ria88]

Mixing qubit types sounds great on paper. In practice, every technology needs different wavelengths, temperatures and control systems

Transducers that convert between them are still very lossy. I think it will happen eventually, but not soon. Microwave to optical conversion is the big bottleneck. Several groups are working on it. Progress has been slow but steady

[DeepCourier]

Monroe has been talking about photonic links for well over a decade. Seeing the rate climb like this must be satisfying for him. His Duke group and IonQ have always been closely tied. This feels like a long term bet paying off