IonQ says its new quantum chip cuts the cost per qubit 300× by moving to standard chip factories

IonQ, the public quantum-computing company that trades on the NYSE under the ticker IONQ, says it has finished its first quantum chips on a standard commercial foundry line, a manufacturing shift IonQ is now betting its 2027 and 2028 roadmaps on. The chip, called Superion 256, is IonQ's sixth-generation quantum processor, and the company is framing the launch less as a physics advance than as a fabrication milestone: quantum hardware, IonQ argues, can now ride the same rails as ordinary silicon.

A qubit, the basic unit of information in a quantum computer, is harder to manufacture than a transistor. Until now, most of IonQ's trapped-ion qubits had to be steered by racks of free-space lasers positioned around the chip, a setup that scales poorly and ties each processor to a custom optical bench. Superion 256 replaces that laser rig with on-chip electronic control. The technology, which IonQ calls Electronic Qubit Control (EQC), embeds microwave and radio-frequency circuitry directly into the semiconductor substrate so each qubit can be addressed without external optics. The approach comes from Oxford Ionics, the British ion-trap specialist IonQ acquired earlier this year.

The manufacturing side comes from a different deal: SkyWater Technology, a U.S. pure-play foundry that fabricates chip designs for outside customers rather than running its own products. IonQ says six rapid tapeout cycles at SkyWater during the first half of 2026 compressed individual design cycles from nine months to two, and produced 12 times more wafer lots than the boutique cleanroom partnerships IonQ used previously. First prototype Superion 256 units are already trapping ions at IonQ's U.S. facilities, and the company opened commercial pre-orders for the system in the first quarter of 2026.

IonQ's central economic claim is a 300× reduction in cost per qubit across the roadmap, driven by removing the free-space optical stack and using standard CMOS fabrication. The figure is IonQ's own estimate, no third-party benchmark has been published, and it spans a multi-year roadmap rather than a shipped price. The company also says a Superion 256 unit can fit into a standard data-center server rack, which would let quantum processors plug into the same physical chassis customers already use for conventional accelerators. Production systems are slated for both remote access through IonQ's cloud and on-premises server-rack delivery.

The bigger hardware target sits one generation out. Superion 10K keeps the Superion 256 chip footprint as its baseline and layers cryogenic CMOS logic, classical control circuits designed to operate at extremely low temperatures, directly onto the quantum chip. That on-chip classical layer handles multi-qubit routing, executes classical and quantum instructions in concert, and runs the schedules for quantum error correction, including the quantum low-density parity-check (qLDPC) code blocks that underpin IonQ's fault-tolerance plan. The architecture, branded Walking Cat, is the company's stated path to a logically protected quantum computer.

The schedule is now the story. IonQ is targeting 2027 for commercial Superion 256 deliveries, with laboratory-level fault tolerance on the Superion 10K in 2027 and volume manufacturing in 2028. Each of those dates is a company milestone, not a verified outcome, and ion-trap roadmaps have historically slipped. Reporting so far is single-source, drawn from Quantum Computing Report's aggregation of IonQ's announcement, and leaves the foundry-cadence and cost-per-qubit claims unverified by independent measurement. The 300× number and the design-cycle compression will be tested when the first commercial Superion 256 systems ship in 2027.