Kurzweil Scorecard: Nanotech Promised the Cheap Watt. China Delivered It Instead.

In 2005, Ray Kurzweil bet that the end of fossil fuels would arrive through a very specific door: K. Eric Drexler’s molecular nanotechnology, working alongside quantum-dot solar cells that would be “as cheap as newspaper and as tough as asphalt” (The Singularity Is Near, ch. “Response to Critics”). Twenty-one years later, solar modules really are about as cheap as newspaper — wholesale prices touched $0.07 to $0.09 per watt in late 2024 — and renewables have, for the first time since 1919, generated more electricity than coal. But the watt didn’t get cheap because Drexler’s atom-by-atom assemblers arrived. It got cheap because Chinese factories scaled crystalline silicon and tandem perovskite to industrial absurdity. The prediction came true. The mechanism didn’t.

The predictions

Kurzweil tucked these four claims into Chapter Nine of The Singularity Is Near, his response to critics who doubted molecular nanotechnology would ever do anything useful. They form a chain: a 2004 physics result (carrier multiplication in lead-selenide nanocrystals) was supposed to enable a new generation of nanoengineered solar cells, which would in turn cause the end of fossil fuels. He also flagged a 2004 NASA working interest in beaming solar power down from orbit. Two are testable historical claims; two are forecasts now ripe for a verdict.

Where we actually are

The 2004 carrier-multiplication paper. This one is easy. Schaller and Klimov’s “High Efficiency Carrier Multiplication in PbSe Nanocrystals: Implications for Solar Energy” sits in our literature index at work_id 6891965428, posted on the arXiv as cond-mat/0404368 in April 2004 and published shortly after in Physical Review Letters. Kurzweil reported it accurately. What he could not know was that the result would be contested for the next four years. A 2008 follow-up titled — and this is the actual title — “In Spite of Recent Doubts Carrier Multiplication Does Occur in PbSe Nanocrystals” (304 citations, DOI 10.1021/nl0807225) settled the dispute. By 2015, a Nature Communications paper on PbSe nanorod solar cells reported external quantum efficiencies above 100%, peaking at 122% for the smallest-bandgap rods. The physics worked. The cells worked. They just never escaped the lab.

The cheap-and-tough solar cell. Kurzweil’s restatement in The Singularity Is Nearer drops the asphalt metaphor but keeps the substance: “Putting tiny structures called nanotubes and nanowires inside solar cells can steadily improve their ability to absorb photons… placing nanocrystals (including quantum dots) inside cells can increase the amount of electricity generated per photon of sunlight absorbed.” He also writes a section header that would have read as fantasy in 2005: “Renewable Energy Is Approaching Complete Replacement of Fossil Fuels.”

On price, this prediction has been spectacularly verified. Module spot prices in China hit $0.07 to $0.09 per watt across 2024 and the first half of 2025. A copy of The New York Times — a literal newspaper — costs roughly $3, weighs around 200 grams, and covers a surface area near a square meter. A 400-watt module covers a similar area, weighs ten times more, but at $0.08/W sells wholesale for $32. Per square meter, a newspaper and a solar module are now in the same order of magnitude on cost. Kurzweil’s metaphor was prescient.

But the path he predicted was wrong. The cost collapse did not come from quantum-dot architectures or molecular self-assembly. It came from monocrystalline silicon scaled to a degree the 2005 industry would have considered grotesque, plus a new top layer — perovskite — bonded to silicon in tandem cells. LONGi reported a certified 34.85% efficiency for a perovskite-silicon tandem in April 2025; JinkoSolar and Aiko both hit 34.76% later in the year. Oxford PV’s commercial product is shipping at 24.5% on a 26.8% record. Our patent index shows perovskite-tandem filings rising from two in 2020 to eleven in 2024 and ten in 2025, with six already through the office in 2026 alone. Read the claims of US 12,598,856 (April 2026, “Mechanically strong connections for perovskite-silicon tandem solar cells”) and the invention is unmistakably industrial: patterning trenches into the perovskite subcell, filling them with insulator supports, bonding modules with conductive adhesive over those supports. This is precision manufacturing, not Drexlerian assembly.

The “as tough as asphalt” half is unevenly delivered. Most installed solar in 2026 is still rigid glass-and-aluminum. But flexible perovskite is moving: US 12,484,368 (November 2025) describes a scalable fabrication method for a flexible perovskite solar cell using a tin-oxide and phenyl-trimethyl-ammonium-chloride colloidal electron transport layer annealed at 100°C or less — a low-temperature process suited to plastic substrates. That is a real step toward Kurzweil’s “rolls, films, coatings,” though it isn’t asphalt and isn’t deployed at scale.

The end of fossil fuels. Ember’s 2026 Global Electricity Review settled the question for 2025: renewables generated 33.8% of global electricity, edging past coal at 33.0% — the first time since 1919 that coal was not the single largest source. Solar generation grew 30% year-over-year, adding 636 terawatt-hours, and met 75% of the year’s electricity demand growth. Cumulative installed solar passed 2.2 TW at end of 2024. Fossil-fuel electricity generation fell, in absolute terms, by 0.2% — a small number, but the first structural (not recession-driven) decline.

That is the direction Kurzweil predicted. The pace is the problem. In the United States, fossil fuels still produced about 60% of utility-scale electricity in 2023, and renewables sat at 24.2% in 2024. Globally, oil and gas demand in transport and industry remain stubbornly high. “End” implies an endpoint. The 2020s have given us an inflection, not an end.

NASA and space-based solar. Kurzweil’s claim that NASA was “actively considering space-based solar power as an energy solution by 2004” is verifiable historical record. What’s interesting is how unfrozen this field has become. Caltech’s SSPD-1 demonstrator launched in January 2023 and, in March of that year, executed the first wireless transmission of power between satellite components in low Earth orbit; the team also directed a beam toward a receiver on the Caltech rooftop. Our patent index shows the underlying art maturing in real time: US 11,251,658 (February 2022) claims a “solar powered microwave beaming system” with transparent microwave patch antennas collocated with solar cells on a flexible substrate, each cell powering an individual microwave transmitter for distributed beam-forming. US 12,470,090 (November 2025) and US 12,525,827 (January 2026) extend the architecture, with the latter describing a satellite-mounted photovoltaic array driving laser diodes via direct, unconditioned electrical connection — power-down-from-orbit via optical rather than microwave beam. China has published a roadmap to a 10-kilowatt demonstrator in 2028, a 1-megawatt geostationary station by 2030, and 2 GW commercial scale by 2050.

The scorecard

What Kurzweil missed (and what he nailed)

The throughline of this batch is a forecasting pattern worth naming: Kurzweil was right about the destination and wrong about the vehicle. He correctly identified solar as the technology that would erode fossil fuels. He correctly anticipated that exotic physics — carrier multiplication, quantum dots, nanocrystals — would be active research areas. He correctly anticipated that wholesale module cost would fall to newspaper-territory. What he missed was that the cheap watt would arrive through a different door entirely: massive Chinese state-directed industrial scale-up of unsexy crystalline silicon, with perovskite arriving as a tandem-cell topcoat in the late 2020s rather than as a quantum-dot revolution in the 2010s. The cells inside the modules on US roofs in 2026 are not Drexler’s. They are Czochralski’s, refined for sixty years.

This points to a systematic bias in 2005-era forecasting from the nanotech community: an assumption that step-change efficiency gains would emerge from atomically precise structures, and that those structures would be made by molecular machines. In practice, efficiency has crept upward through pedestrian materials engineering and the path to abundance has run through manufacturing learning curves, not assembly-line bots. The carrier-multiplication papers Kurzweil cited turned out to be true science with no commercial heir. The end of fossil fuels is happening anyway — slowly, through a route Kurzweil’s 2005 framework didn’t predict and his 2024 update partly retrofits.

Two things worth watching. Perovskite-silicon tandems at 34%+ research efficiency are now moving through Oxford PV, JinkoSolar, and LONGi factories. The Shockley-Queisser limit on single-junction silicon (~33%) is being routinely breached in production-scale tandem geometry; the next decade will measure how fast that flows from records into roofs. Separately, the space-solar story moved from “actively considering” to “physically transmitted” between Kurzweil’s two books. If China hits its 2030 target of 1 MW in geostationary orbit, this batch gets rescored.

Method note

The patent counts and the patent text in this post come from a 9.3-million-document U.S. patent corpus searched by full-text query and read in detail for the most relevant inventions. The carrier-multiplication and multiple-exciton-generation literature was searched and ranked by citation count across a 357-million-paper academic corpus. Solar deployment, module-price, and renewables-vs-coal figures come from the International Energy Agency, the IEA Photovoltaic Power Systems Programme, Ember’s 2026 Global Electricity Review, OPIS, InfoLink, and Wood Mackenzie. Kurzweil’s restated views are quoted from The Singularity Is Nearer (2024). Original predictions are paraphrased from The Singularity Is Near (2005).

Sources:
– Ember, Global Electricity Review 2026
– IEA, Renewables 2024 — Electricity
– IEA-PVPS Snapshot 2025
– pv magazine: JinkoSolar 34.76% perovskite-silicon tandem
– LONGi: new world record at SNEC 2025
– Caltech: Space Solar Power Demonstrator wirelessly transmits power
– InfoLink PV spot price index
– pv magazine: China module prices 2026
– Nature Communications: Multiple-exciton generation in PbSe nanorod solar cells
– China space-based solar roadmap (Sustainability Magazine)