Quantum news, weighed against the evidence.

Claim vs. evidence

Headlines report the biggest number. The Ledger reports what was shown.

Every claim gets a tag for how strong the evidence is: measured, estimated, simulated, announced or conditional. Four examples from one week:

• Claim1,000 entangled pairs per second over an ion–diamond linkIonQ
  →
  EvidenceMeasured

  1,032 pairs/s at 87.9% Bell-state fidelity, dropping to 612/s once verification readout is included. One ion linked to one memory.

• Claim~99% CNOT gate from 37 exchangesHRL / IBM
  →
  EvidenceEstimate

  Not demonstrated. What was measured: a 0.032% effective error per exchange between two spins.

• Claim10–40× cheaper magic statesQuEra / Harvard
  →
  EvidenceSimulation / design

  Lower simulated resource costs for a factory design. It is a design, not hardware.

• ClaimUp to $300M in DARPA QBI Stage C fundingIonQ
  →
  EvidenceConditional

  Selected for Stage C. Funding beyond the initial obligation is conditional, and the work runs through 2029.

Every Monday

Three things in every issue

• 1.

  What changed, company by company

  A short card for each quantum company in the news: what actually changed this week, how far along it is, and why it matters.

• 2.

  A ledger of claims, tagged by evidence

  The week’s headline claims in one table, each marked measured, estimated, simulated, announced or conditional.

• 3.

  What to watch, with the sources

  What to revisit in the weeks ahead, plus links to the papers, filings and announcements behind the issue.

Issue preview

What lands in your inbox

An excerpt from a sample issue covering the week of October 7–14, 2026.

Inbox

The Notebook Ledger

Issue #01 (sample) · Week of October 7–14, 2026

Gains that have to survive integration

The week in one paragraph

Much of this week’s quantum news was about what surrounds the processor rather than the processor itself: financing, manufacturing, testing, and the classical machinery that error correction depends on. On the research side, the recurring lesson was that better components and a cheaper complete computer are not automatically the same thing.

Company card

$IONQWhat changed

• Agreed to acquire Elisity for up to $1.3B in stock, including up to $150M tied to performance. Expected to close in Q4 2026, subject to approvals.Announced
• Selected for DARPA QBI Stage C, worth up to $300M; funding beyond the initial obligation is conditional.Conditional
• Published the experiment behind its quantum link: 1,032 pairs/s at 87.9% Bell-state fidelity.Measured

Why it matters: the open question is whether management can turn stock-funded acquisitions into value per share.

Watch next: Elisity close and integration · Stage C test results · higher fidelity and throughput on the link.

Claim-vs-evidence ledger

Six of the issue’s thirteen rows

Claim	Company	Evidence	Status this week
18–23% fewer errors per round	IBM	Measured	Supported, no trials discarded
Elisity acquisition	IonQ	Announced	Pending close, Q4 2026
Spectrum-DF prototype designed and tested	Infleqtion	Company claim	Funding for validation phase
Faster machines, fewer chips via GlobalFoundries	Xanadu	Announced	No measured gains yet
~¼ Toffoli count for QROM	Xanadu	Resource estimate	Not a runtime gain
Real-time GPU decoding	Alice & Bob	Proof of concept	Synthetic data, not real-time

Watch list

1. Elisity deal close and any detail on technical synergies
2. Complete-gate fidelity on HRL silicon-spin hardware
3. Infleqtion field validation of Spectrum-DF

The full issue continues with cards for $IBM, $INFQ, $QNT, $XNDU and $PSQL, one-line changes by technology (including sensing and clocks) and consolidated sources.
Get the next issue

Shortened from the sample issue. Figures are as reported in the week’s posts. Nothing here is a recommendation to buy or sell any security.

Science first

Built to separate real progress from press-release momentum

For investors and curious readers who follow quantum companies and want to know which results are real without reading every paper themselves.

See the work first

Judge it by what is already published

The Ledger draws on the same research this site publishes in the open. Read some before you subscribe.

Latest: The cost of connecting (research, Oct 3–9) · Paying for the next machines (news, Oct 3–9) · IonQ Has Built a Bigger Company

Questions

Before you subscribe