Deep-Tech Press Release Credibility: 8 Cases

A credible deep-tech press release shows what happened, who took part and what was measured, and it labels future plans as plans. This Virgo PR audit of eight 2026 announcements found that seven contained a hard number and six reported a completed event, but those results describe this small sample only, not the industry or the odds of earning media coverage.

Research update: November 4, 2026. Analysis by Virgo PR. Part of a series that began with how funding announcements earned coverage.

Limitations, up front: this is a one-time, eight-case qualitative content audit of first-party announcements, selected as a convenience sample to illustrate different evidence structures. It is not a random or representative sample, not an annual index, and not a study of editorial pickup or of whether the claims are true. Source claims were coded as written, and no technical claim was independently replicated.

Summary: what did the audit find?

• Seven of the eight announcements contained at least one concrete numeric fact.

• Six reported a completed technical or operational event, and six named a customer or collaboration counterparty.

• Only one quoted an independent outside expert who was not part of the announced work.

• The clearest dividing line was between demonstrated results and proposed future plans, which scored 2 and 1 out of 5.

Contents

• What this audit covers

• How the eight announcements scored

• What the sample results show

• Completed results versus plans

• Why independent quotes are rare

• How to make evidence visible

• What this study does not show

• Frequently asked questions

What does this deep-tech press release audit cover?

Deep-tech companies often announce technologies that take specialist knowledge to assess. For a journalist, the central challenge is not whether a claim sounds impressive. It is whether the announcement supplies enough information to establish what actually happened.

Virgo PR reviewed eight first-party announcements published in 2026 across quantum computing, photonics, defense autonomy and semiconductor infrastructure. Each was scored against five yes-or-no criteria, worth one point each:

• A. Named external partner: a customer or collaborator is identified.

• B. Quantitative fact: at least one technical, financial or operational number is given.

• C. Independent expert: an outside expert is quoted who is not a co-announcing partner.

• D. Completed event: the release reports a result or event, not solely a plan.

• E. Implementation detail: concrete details or a near-term timeline are provided.

A score is a descriptive completeness count, not an objective credibility rating. A low score does not mean an announcement is untrue, and a high score does not mean it is correct.

How did the eight announcements score?

The table lists each announcement, its date, the evidence it offered and its descriptive score. Each case links to the original announcement.

AnnouncementDateSubject and evidenceScoreStatus

QuantinuumOct 19, 2026Quantum error correction: 99.9995% logical fidelity; third-party scientist commentary4/5Completed Shield AI and HoneywellOct 13, 2026Autonomous aircraft: two-aircraft mission; six-week training timeline4/5Completed Shield AI, Sedaro and NOVIAug 24, 2026Satellite autonomy: on-orbit experiments; six-satellite simulated constellation4/5Completed Photonic and TELUSFeb 13, 2026Quantum networking: 30 km of installed commercial fiber4/5Completed Shield AI and KrakenSep 28, 2026Maritime autonomy: two vessels completed a coordinated mission4/5Completed Shield AI and L3HarrisJul 16, 2026Electronic warfare: completed live flight test with autonomous rerouting4/5Completed Quantinuum and OracleAug 11, 2026Cloud quantum access: named deployment partner; deployment planned2/5Forward-looking Photonic Project VANGUARDSep 14, 2026Semiconductor facility: proposed project of up to CAD $500 million, not committed construction1/5Forward-looking

Six announcements scored 4 out of 5, one scored 2 and one scored 1. The pattern is striking, but with eight hand-picked cases it is an illustration of how evidence is structured, not a measurement of how often each pattern occurs.

What do the sample results show?

Counting each criterion across the eight announcements gives the following sample results.

CriterionAnnouncements meeting it

Named external counterparty6 of 8 Numeric fact7 of 8 Independent expert quote1 of 8 Completed milestone6 of 8 Implementation detail or timeline7 of 8

The sample leans toward technically specific releases, because that is how it was selected. It cannot establish that the same proportions apply across deep-tech PR generally.

What separated completed results from future plans?

The strongest announcements state what happened, who participated and what was measured. Quantinuum's October 19 release is a clear example. It reported a logical fidelity of 99.9995% on its Helios computer, described as five errors per million operations, and named the comparison it was making. It also included commentary from Robin Blume-Kohout of Sandia National Laboratories, who the release says was not involved in the study.

The Photonic and TELUS announcement described quantum information transfer over 30 kilometers of installed commercial fiber, which is a specific distance under real-world conditions. The Shield AI announcements with Honeywell, Sedaro and NOVI, Kraken and L3Harris each reported a completed demonstration involving a named counterparty.

At the other end of the scale sit announcements about what a company intends to do. The Quantinuum and Oracle release named a deployment partner but described a deployment that was still planned. Photonic's Project VANGUARD described a proposal for a multi-tenant semiconductor manufacturing facility with an estimated cost of up to CAD $500 million, selected for inclusion in a government investment prospectus. A proposed project is not equivalent to a completed facility or a committed investment, and the release itself describes it as a proposal.

Future-oriented announcements can still be newsworthy. They simply need explicit labels so an editor can tell a result from an ambition. Even a strong release can mix the two: Quantinuum's own forward-looking statements section flags its Apollo fault-tolerant computer, expected by 2029, as a plan rather than an achievement.

Why are independent expert quotes so rare?

Only one of the eight announcements quoted an outside expert who was not part of the announced work. That is unsurprising. A company press release is a first-party document, and companies usually quote their own executives and their named partners.

The independent-expert criterion is also not required for legitimate journalism or for a legitimate announcement. A partner's endorsement is useful context, but it is not independent verification, and communications teams should avoid presenting it as such. Where an outside researcher can speak candidly about a result, as in the Quantinuum release, it gives an editor an additional reference point. Where none is available, the next best option is to point to the primary technical document.

How can communications teams make evidence visible?

The practical lesson is to make the evidence visible. This study does not claim that these practices guarantee press coverage. It identifies the factual scaffolding that helps an editor evaluate a technical announcement. Teams working in technology PR can use this checklist.

• Lead with the verifiable event. Open with a completed test, a signed award or a delivered product, not an adjective such as "breakthrough."

• Include the measurable result and the test conditions. Avoid comparisons that lack a clear baseline.

• Identify external counterparties when authorized. Distinguish a partner's endorsement from independent verification.

• Label future projects accurately. Say whether something is a proposal, a planned deployment, a memorandum or an executed contract.

• Provide a primary technical document. Link to the underlying evidence where possible, and do not present internal claims as independently established.

What does this study not show?

• It is not a media pickup analysis. Virgo PR did not measure whether any of these announcements earned coverage.

• It does not validate the science. The review evaluates the information disclosed in each release, not whether the underlying technology performs as claimed.

• It is not representative. The eight cases were chosen to illustrate differing evidence structures, so the proportions cannot be applied to the industry.

• A score is not a credibility verdict. The checklist counts completeness. It does not rate honesty, scientific validity or audience interest.

Method in full

Selection: eight accessible first-party announcements published in 2026 across quantum, photonics, defense autonomy and semiconductors. Scoring: a binary checklist of five criteria, one point each, applied to the announcement as written. The scores are descriptive completeness counts, not objective credibility ratings. Technical claims were not independently replicated. Before publication, Virgo PR re-opened the Quantinuum and Photonic Project VANGUARD announcements to confirm the figures and wording cited in this article.

Frequently asked questions

What makes a deep-tech press release credible?

Specific measurements, clear attribution, identified counterparties where possible, and an accurate separation of completed achievements from future plans.

Does a high score guarantee media coverage?

No. This checklist measures announcement completeness, not press pickup, scientific validity or audience interest.

How many releases were reviewed?

Eight deliberately selected first-party announcements published in 2026.

Was the underlying technology independently tested?

No. The analysis evaluates information disclosed in the releases, not independent replication of technical claims.