Leonardo Corporation's whitepaper promises 24/7 electricity from Zero-Point Energy. We read all 20 pages — the specs, the ChatGPT sustainability table, the Renault Twizy on a go-kart track — so you can weigh the evidence yourself.
Andrea Rossi has been "almost ready" since 1997. His latest document — a polished, 20-page whitepaper from Leonardo Corporation, revised as recently as December 2025 — claims the wait is finally over. The NGU Power Cell, a white plastic disc about the size of a hockey puck, allegedly harvests electricity directly from the quantum vacuum. No fuel. No emissions. 24/7, forever. Thirty-five dollars per 10-watt cell.
If true, this is the most important invention in human history — bigger than fire, bigger than the transistor. If not, it joins a long and distinguished line of LENR-adjacent announcements that generated significantly more enthusiasm than watts.
We read every page. Here's what we found.
The whitepaper is refreshingly concrete about the hardware — up to a point. From the outside, the NGU Power Cell is described precisely. What happens inside is proprietary, though the document does mention "a miniature xenon vacuum tube and a powerful AI controller." ◎Self-Claimed ?Not Independently Verified
| Configuration | Power | Cells | Notes |
|---|---|---|---|
| Single NGU Cell | 10W | 1 | Base unit |
| Basic Generator | 100W | 10 | Parallel connection |
| Demo Unit | 3,000W | 300 | Used in Latina 2024 demo |
| Power Plant (small) | 1,000,000W | 100,000 | Fits in a 20-foot container |
| Power Plant (large) | 2,000,000W | 200,000 | Fits in a 40-foot container |
The theoretical foundation is Zero-Point Energy (ZPE) — the lowest possible energy state of a quantum mechanical system. ZPE is real: it has measurable effects (the Casimir effect, Lamb shift) and is a well-documented phenomenon in quantum field theory. ✓Documented
What mainstream physics does not accept is that ZPE can be extracted as usable work at useful scales. The whitepaper itself acknowledges this directly: "The theory behind E-Cat Power is still not generally accepted by science." ⚠Disputed by Physics
Rossi's theoretical model is laid out in his 2019 paper, E-Cat SK and Long-Range Particle Interactions, which draws on the work of Tesla, Dirac, Puthoff, and others to propose an electron-based mechanism for extracting electromagnetic energy from ZPE. The paper has accumulated more than 170,000 reads. ✓Documented It has zero peer-reviewed independent replications. ?Not Independently Verified
The whitepaper situates E-Cat Power within the real energy transition challenge — citing IEA data. The numbers are accurate. Only 29% of global electricity came from renewable sources in 2022. Coal and gas still supply 58%.
| Source⇅ | Global Share 2022⇅ | Category⇅ |
|---|---|---|
| Coal | 35.4% | Fossil |
| Gas | 22.7% | Fossil |
| Hydropower | 14.9% | Renewable |
| Nuclear | 9.2% | Carbon-free |
| Wind Power | 7.2% | Renewable |
| Solar Power | 4.5% | Renewable |
| Biomass / Other | 2.7% | Renewable |
| Other | 3.4% | — |
| E-Cat Power | 0% | Claimed renewable |
Source: IEA World Energy Outlook 2022. Click column headers to sort.
Rather than awaiting independent scientific evaluation, Leonardo used GPT-4o to generate a 1–5 rating of E-Cat Power vs. other renewables. The input was Leonardo's own unverified performance data. The whitepaper includes a disclaimer noting the ratings are "subjective and still provisional." Amber rows indicate self-reported data. Click headers to sort.
| Energy Source⇅ | Environmental Impact⇅ | Resource Availability⇅ | Energy Efficiency⇅ |
|---|---|---|---|
| Solar Power | 4/5 | 5/5 | 3/5 |
| Wind Power | 4/5 | 4/5 | 4/5 |
| Hydropower | 3/5 | 4/5 | 4/5 |
| Biomass | 3/5 | 3/5 | 3/5 |
| E-Cat PowerSELF-SCORED ✦ | 5/5 | 5/5 | 5/5 |
✦ Ratings for E-Cat Power were generated by ChatGPT GPT-4o, using unverified data provided by Leonardo Corporation as input. The model was not skeptical.
The first public demonstration used two Renault Twizy 80 electric vehicles — the lightest, slowest, and most modest EV sold in Europe. One carried a 3kW NGU Power Generator. Both drove simultaneously on a 0.6 km racetrack in Latina, Italy.
The standard vehicle (EV2) ran until the battery died. The E-Cat vehicle (EV1) drove for six hours, ending with a higher battery charge than it started with.
| Vehicle | Start SoC | End Time | Distance | End SoC | SoC Δ |
|---|---|---|---|---|---|
| EV1 — E-Cat (3kW NGU) | 68% | 380 min | 201 km | 83% | +15% |
| EV2 — Standard Twizy | 98% | 140 min | 74 km | 0% | −98% |
Leonardo uses the standard LCOE (Levelized Cost of Electricity) metric. At $35/cell, a 3 kW residential system requires ~300 cells = $10,500 in cells alone. The whitepaper prices the full initial system at $3,700/kW = $11,100 for 3 kW, suggesting tight integration costs. Toggle between scenarios below.
| Factor | residential value |
|---|---|
| Installed Power | 3 kW |
| Power Efficiency | >97% |
| Initial System Cost | $3,700 / kW |
| Cell Replacement (at 10 yr) | $1,750 / kW |
| Installation | $30 / kW |
| Site Building | $50 / kW |
| Maintenance | $30 / kW / yr |
| Grid Network (optional) | $60 / kW / yr |
| Uptime Efficiency | 99% |
| Yearly Production | 25 MWh |
| Calculated Lifetime | 20 years |
| Lifetime Production | 0.5 GWh |
| E-Cat Power LCOE | < $0.10 / kWh |
Source: Leonardo Corporation whitepaper, Chapter 5. Note: VAT, taxes, and green energy subsidies excluded. Grid export income not included (would further reduce LCOE). Independent verification pending.
Since March 2023, Leonardo Corporation has accepted pre-orders at ecathenewfire.com. Pre-orders are non-binding and require no payment. Production begins when the equivalent of 10 million 10W NGU Power Cells are pre-ordered.
Twelve million cells at $35 each would represent $350 million in theoretical demand — though since no money changes hands, the milestone is structurally a petition rather than a commercial commitment. Leonardo has not disclosed current pre-order figures.
The whitepaper notes that the manufacturing system is "already prepared," including in-house production of key components and a strategic outsourcing network. The company is waiting only for demand to be demonstrated before flipping the switch.
The E-Cat Power whitepaper is, by the standards of the genre, a good document. It is clearly written, internally consistent, and — to its credit — forthright about its own limitations. It does not hide the fact that the science is disputed, the technology unverified, and the business model contingent on a threshold that has not been reached. These admissions are buried in disclaimers, but they are there.
What the whitepaper cannot offer is the one thing that would settle all questions: a functioning unit in independent hands. The inner design is proprietary after 27 years. The demonstration used the company's own equipment, inspected by event-appointed specialists, on a private track, with no published raw data. The sustainability scores were generated by a language model that was not designed to be adversarial.
The LENR community has seen extraordinary claims before — some of which turned out to have real anomalies worth investigating, most of which did not. The E-Cat's journey from 1997 cold-fusion replication attempt to 2025 Zero-Point Energy commercial pre-order has been consistent in two respects: the energy output has always been impressive, and the independent verification has always been imminent.
If this device does what it claims, it is the most consequential technology in human history. That is precisely why the bar for evidence should be proportionally high — and why we will be watching, carefully, for independent data.