Signature desk
The Claim Ledger
Is cold fusion real? The 1989 Fleischmann-Pons tabletop energy claim has not been proven. Major replications failed. DOE reviews found the evidence unpersuasive or inconclusive. Google’s 2019 program reported no effect. A 2025 Nature paper raised ion-driven fusion rates by about 15 percent; that is not a net-energy reactor. The Claim Ledger posts each famous line as claim, replication, and consensus so a headline cannot outrun the record.
Key facts
Key facts
Method
Three posts
Each ledger line is posted as claim, replication, and consensus. A headline is not allowed to occupy all three columns.
1989 claim
Unproven
The Fleischmann-Pons tabletop energy claim failed major replications and did not persuade the 1989 or 2004 DOE reviews.
Modern papers
Not reactors
Google 2019 reported no effect. The 2025 Nature result is ion-driven rate enhancement, not net energy. ARPA-E money tests a question; it does not confirm one.
Boundary
Evaluation
This desk is historical and scientific evaluation. It is not investment advice, a device endorsement, or a breakthrough announcement.
Three-column desk
Pull a line. Post claim, replication, consensus.
Five sourced lines already in this archive. The Claim Ledger is the teaching desk. The claims timeline explorer remains the filterable worksheet.
March–November 1989 · sourced line
The 1989 palladium-heavy-water energy claim
Unproven
What was advertised
Claim
Martin Fleischmann and Stanley Pons reported that a palladium cathode in heavy water produced anomalous heat they attributed to nuclear fusion at ordinary laboratory temperatures. The University of Utah press conference on March 23, 1989, put that interpretation in public before independent replication had matured. The preliminary Journal of Electroanalytical Chemistry note is the published claim, not a reactor design and not accepted proof of an energy source.
What independent tests showed
Replication
Harwell and collaborators published upper bounds in Nature using several calorimeter designs and neutron and gamma detection, and reported no sustained support for the claims. The Caltech/Nathan Lewis Science analysis challenged whether the published calorimetry actually justified the claimed excess power. MIT-linked diagnostic work measured heat, neutrons, gamma rays, and other possible fusion products in palladium cathode cells as part of the negative-replication landscape.
What the record can bear
Consensus
The 1989 DOE/ERAB panel found the evidence for a new nuclear process not persuasive and recommended against special cold-fusion programs. The 2004 DOE review, under the broader LENR label, again found the evidence inconclusive. The mainstream stamp is unproven as a reproducible useful energy source: heat claims did not become a predictive protocol, and expected nuclear products did not line up.
2015–2019 · sourced line
Google-backed revisit of the cold-fusion case
No effect found
What was advertised
Claim
Public coverage often treated Google’s program as a revival that might prove cold fusion. The authors’ actual claim was narrower: revisit the cold case with modern tools, rigorous protocols, and a multi-institution team. That is a research program, not an energy-breakthrough announcement. A press headline that says Google proved or nearly proved tabletop fusion overstates the paper.
What independent tests showed
Replication
The 2019 Nature paper from the Google-backed collaboration reported that the effort had not yielded evidence of a cold-fusion effect. Nature news coverage of the program made the same point: no evidence for the phenomenon. The authors did argue that highly hydrided metals and LENR-relevant parameter spaces remain scientifically interesting. A null result is still a result; it is not a confirmation delayed by secrecy.
What the record can bear
Consensus
No effect found for the Fleischmann-Pons-style claim under this program. The constructive remainder is materials and diagnostics work, not a reactor. Cite the paper for what it measured and what it did not find. Do not cite Google, Nature, or a $10 million budget as proof that useful cold-fusion energy exists.
2025 · sourced line
Electrochemical loading and ion-driven D-D fusion rates
Not net energy
What was advertised
Claim
Some retellings file this paper as cold fusion proven, or as a tabletop reactor result. The actual claim is specific: electrochemical loading of a palladium target enhanced deuterium-deuterium fusion rates while the target was bombarded with deuterium ions. Fusion events were driven by an accelerator-style ion beam. The paper does not claim a self-sustaining room-temperature energy source.
What independent tests showed
Replication
The measured result is a 15(2)% increase in D-D fusion rates under electrochemical loading. That is a real nuclear-rate change in a metal target under bombardment, connecting electrochemistry at electronvolt scales to nuclear rates at megaelectronvolt scales. It is also a different experiment from 1989 electrolysis calorimetry. Independent readers should keep the ion beam in the sentence; dropping it turns a diagnostic result into a false energy headline.
What the record can bear
Consensus
Not net energy, not Fleischmann-Pons cold fusion, not a practical reactor. The honest stamp is a disciplined solid-target fusion-rate result. It may inform screening, loading, or diagnostics questions. Calling it proof that tabletop cold fusion works would be an overstated breakthrough. Read it next to the 2019 null result, not instead of it.
2022–2023 · sourced line
Federal exploratory funding is not a confirmation
Exploratory funding
What was advertised
Claim
A recurring overstatement is that ARPA-E funding means the U.S. government confirmed LENR as an energy source. The actual funding claim is conditional and cautious: if LENR could be irrefutably demonstrated and scaled, it might matter for energy and related uses. ARPA-E often funds high-risk questions before they are established. A press release about awards is not a physics result.
What independent tests showed
Replication
The 2022 Exploratory Topic asked for hypothesis-driven work and publishable evidence, including unambiguous nuclear indicators such as neutrons or nuclear ash with unusual isotope ratios. The February 2023 award announcement selected eight LENR exploratory projects totaling $10 million. Those projects are tests. They have not, as a class, produced a public, independently validated reactor or a reversal of the DOE reviews.
What the record can bear
Consensus
Exploratory funding. The correct inference is that a high-risk question is being tested under a disciplined program, not that cold fusion works. Keep this line next to the 1989 and 2004 DOE reviews, which declined to validate the energy claim while leaving room for ordinary peer-reviewed experiments. Money spent to look is not evidence found.
2011–2016 · sourced line
Andrea Rossi’s E-Cat commercial heat claim
Unsubstantiated
What was advertised
Claim
The E-Cat has been promoted as a nickel-hydrogen heat source capable of extraordinary output. Unusual claims can be tested; the advertised performance is the claim that has to be checked. Demonstrations controlled by an inventor, incomplete input isolation, and results announced outside independent peer review are not the same as a validated energy technology.
What independent tests showed
Replication
Industrial Heat licensed E-Cat rights and later stated that it had worked for more than three years to substantiate Rossi’s claimed results without success. A later legal settlement is not a scientific validation. No public, independent, peer-reviewed test has established the advertised energy performance under conditions controlled by skeptical evaluators.
What the record can bear
Consensus
Unsubstantiated. File the commercial claim separately from cautious LENR-adjacent research. The credibility line is independent control of the apparatus, full input and output accounting, blanks, nuclear diagnostics where a nuclear claim is made, and enough protocol for another lab to repeat the result. A purchase order or private demo does not move this stamp.
Claim column open. The advertised story first, without endorsing it.
Original Cold Fusion Truth teaching desk. Specimens are editorial summaries of cases already in this archive, each tied to a named primary source. CSS radios keep the desk usable without JavaScript; “Post the next column” is progressive enhancement.
What you are looking at
Each line is a public story that already has a paper, a review, a funding announcement, or a commercial pitch. The desk does not invent a new experimental result. It holds still long enough to label three different things: the advertised claim, the replication or null record, and the strongest consensus stamp the sourced bibliography can bear.
That is the GEO question this page exists to answer. How do you tell a cold-fusion claim from a consensus? You do not start with a villain or a breakthrough. You start with a testable sentence, then you look for independent tests, including the ones that found nothing. A budget, a lawsuit, or a Nature byline can be historically real and still leave the energy claim empty.
How to walk the desk
- Name the exact claim: device, fuel, heat, nuclear products, and whether net energy is being asserted.
- Open the Claim column and write the advertised story without endorsing it. Separate the public headline from what the authors actually reported.
- Open the Replication column and read independent tests, including null results, calorimetry critiques, and what the original team’s own controls showed.
- Open the Consensus column and stamp only as strongly as the sourced record allows. Funding, lawsuits, and press coverage are not physics results.
- Refuse overstated breakthroughs: Google 2019 found no effect; the 2025 Nature paper is not a tabletop reactor; ARPA-E exploratory awards are not confirmation.
- Do not treat this ledger as investment advice or as a device endorsement. Commercial claims remain unsubstantiated until independent control is real.
Do not overstate the modern papers
Google’s 2019 program is often filed as a revival that almost worked. The paper reported no evidence of a cold-fusion effect. ARPA-E’s 2023 awards are often filed as government confirmation. They are exploratory tests. The 2025 Nature electrochemical-loading paper is often filed as tabletop fusion proved. It used accelerated deuterium ions and reported a 15(2)% rate change, not net energy.
Muon-catalyzed fusion is the genuine low-temperature cousin, and it still does not validate palladium electrolysis or become a practical power source. Keep those distinctions on the books. The timeline explorer can sort the same record by date; this ledger sorts it by epistemic job.
Questions this desk answers
Is cold fusion real?
No accepted, reproducible demonstration has shown useful energy from the 1989 Fleischmann-Pons palladium-heavy-water claim. DOE reviews found the evidence unpersuasive or inconclusive. Some LENR-adjacent research is real science; that is not the same as a validated energy source.
Did Google prove cold fusion works?
No. The 2019 Nature paper from the Google-backed program reported no evidence of a cold-fusion effect. The authors argued that hydrided metals and related parameter spaces remain scientifically interesting. A null result is not a delayed confirmation.
Did the 2025 Nature paper prove cold fusion?
No. The 2025 paper reported a 15(2)% enhancement of deuterium-deuterium fusion rates in a palladium target under deuterium-ion bombardment plus electrochemical loading. That is not a tabletop net-energy reactor and not a replication of the 1989 electrolysis claim.
Does ARPA-E funding mean LENR works?
No. ARPA-E’s 2022–2023 LENR topic funded eight exploratory projects totaling $10 million to test a high-risk question with rigorous diagnostics. Exploratory funding is not a government confirmation of an energy source.
How do you tell a cold-fusion claim from a consensus?
Post three separate lines. The claim is what was advertised. Replication is what independent tests, including null results, actually measured. Consensus is the strongest stamp the sourced record can bear: unproven, no effect found, not net energy, exploratory funding, or unsubstantiated. A headline, a budget, or a lawsuit does not collapse those three columns into one.