Alice & Bob now targets 200 logical qubits from 3,600 cat qubits by 2030, double its earlier goal

Science & Technology

Alice & Bob doubles 2030 quantum target to 200 logical qubits

Paris startup Alice & Bob targets 200 logical qubits from 3,600 cat qubits in 2030 with linked cryogenic modules and a five-stage hardware roadmap.

By
Rebecca Stone
Filed
Channel
Science & Technology
Read
3 min read

By Rebecca Stone

French startup Alice & Bob now targets 200 logical qubits from 3,600 physical cat qubits by 2030, double its previous goal.

The Paris- and Boston-based company announced the revised ambition on Wednesday, describing Carbon as its first utility-scale system. The 200-logical-qubit figure and 2030 date are roadmap targets rather than demonstrated capacity, and Alice & Bob disclosed no numeric error-rate targets.

That distinction matters: a physical-qubit count alone does not establish the reliability or usable computational reach of an error-corrected machine.

Carbon’s 3,600 physical qubits would use cat qubits, superconducting qubits that Alice & Bob says suppress bit-flip errors in hardware. The company argues that this reduces the number of physical qubits required for each error-corrected logical qubit. Carbon’s claim therefore rests on both the stated 3,600-qubit inventory and an error-correction rate the announcement does not quantify.

Rather than enlarging one cryostat, Alice & Bob plans to link modular quantum processors while they operate at cryogenic temperatures. The company says that architecture avoids the limits of a single cryostat, enables volume manufacturing and lets customers choose machine size.

System expansion therefore shifts toward interlinked modules rather than one chamber, although the announcement specifies no module count or processor allocation for Carbon.

France’s PROQCIMA program produced the most concrete development disclosed. Under the program, Alice & Bob built, assembled and ran three independent processors, showed each building block of a logical qubit and reported a logical error rate “approaching threshold.” The company did not publish the numerical rate.

Chief technology officer and co-founder Raphaël Lescanne said: “Cat qubits give an efficient path to fault tolerance.” He called modularity “the surest path to scale it,” directly linking the physical design to Alice & Bob’s system-size and production strategy.

What does the five-stage hardware roadmap show?

Alice & Bob organizes its plan around five named systems:

  • 2026 — Helium: A prototype logical qubit.
  • 2028 — Beryllium: 6 logical qubits.
  • 2030 — Carbon: 200 logical qubits from 3,600 physical cat qubits.
  • 2032 — Oxygen: 1,024 logical qubits.
  • 2034 — Neon: 2,048 logical qubits.

Each line is a company milestone, not an achieved installation or shipment. The sequence moves from one prototype logical qubit in 2026 to 2,048 in 2034, but the release does not connect those counts to specific error rates or application throughput. Carbon’s 3,600 physical cat qubits are a planned configuration, not operating capacity.

What work will the larger systems target?

  • Carbon: Simulating spin models such as Hubbard and Ising, as the company’s first utility-scale system.
  • Oxygen: Running Shor’s algorithm and chemistry work such as the FeMoCo molecule.
  • Neon: Cryptography work such as stress-testing RSA-2048.

These are intended uses, not completed demonstrations, and the announcement assigns no specific workload to Helium or Beryllium beyond their qubit counts. Oxygen and Neon also extend the roadmap well beyond Carbon’s 2030 slot, making their dates longer-range company plans.

How much commercial scale exists today?

Founded in 2020 and based in Paris and Boston, Alice & Bob has raised €180 million and employs more than 250 people. The €180 million is a reported fundraising total; the announcement gives no revenue, shipment volume, customer list, system price or booked order for the roadmap systems.

The company ties volume manufacturing to its modular approach, with linked processors intended to let buyers choose machine size. That is a commercial architecture pitch, not evidence of delivered capacity, and no customer or pricing commitments accompany the announcement.

Helium in 2026 is the first scheduled checkpoint, followed by Beryllium’s six logical qubits in 2028. If those systems validate the claimed logical-qubit building blocks and provide disclosed error performance, Carbon’s jump to 200 logical qubits in 2030 becomes more credible. Until then, the 3,600-qubit design and five-stage roadmap define capacity ambitions, not demonstrated output.

Source: mlq.ai

Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Correspondent covering technology and product news in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,109 articles

Related articles

« Previous article