Kurzweil Scorecard: Flying Cars, Without the Microwings

In 2005, Kurzweil predicted the 2030s would belong to nanotech. Solar panels woven from quantum dots and carbon nanotubes. Fuel cells the size of a postage stamp. Cars and aircraft “rebuilt” from the molecule up by Drexler’s nanofactories. Energy and transportation, he wrote, would “effectively become information technologies.”

Twenty-one years later, the outcome is closer than the chapter looked when published. Electric flying taxis are months from FAA type certification. Drones deliver coffee to suburban driveways. Tandem solar cells convert more than a third of incident sunlight to electricity. But the mechanism is wrong almost everywhere it matters. Lithium-ion automotive cells, not nano-fuel cells. Composite layup, not nanofactories. Software, not microwings.

The predictions

Seven predictions from this batch, drawn from chapters titled “The Criticism from Lock-In,” “The Criticism from Malthus,” and “A Panoply of Criticisms” in The Singularity Is Near. The cluster is unusually testable because Kurzweil tied each one to the late 2020s or 2030s and to specific technological pathways.

In The Singularity Is Nearer (2024), Kurzweil restates the underlying thesis: “the costs of environmentally friendly renewable energy have been dropping exponentially as we apply increasingly sophisticated technologies to the design of the underlying materials and mechanisms.” He notes the doubling cadence of solar electricity since 1983 and projects that “about 4.8 doublings” would “put us at 2032 to meet all of our energy needs from solar alone.” The 2024 update is calmer about Drexler’s nanofactories. It treats them as a hopeful bet, not a baked-in milestone.

Where reality landed

Solar got cheap. The chemistry isn’t quantum dots.

Kurzweil wrote that “extremely inexpensive, lightweight, and efficient nanoengineered solar panels together with nano-fuel cells for storage and distribution will become feasible” (ch. “The Criticism from Malthus”). The cost half is met and then some. Cumulative installed solar PV stood at roughly 2.9 terawatts at the end of 2025 with about 647 GW added in 2025 alone. Module prices have fallen by a factor of more than 50 since 2005.

But the nanoengineering is happening in a chemistry he didn’t single out. The 2025 efficiency record for a two-terminal perovskite-silicon tandem cell is 34.85%, set by LONGi in April 2025 and certified by NREL — past the Shockley-Queisser limit for any single-junction silicon device and within reach of the 43% theoretical ceiling for two-junction tandems. “Flexible perovskite/silicon tandem solar cells with 33.6% efficiency” (Nature, 2025, doi:10.1038/s41586-025-09849-4) shows the same architecture working on bendable substrates.

Patents tell the same story. Perovskite-photovoltaic grants went from one in 2014 to 80 in 2025. Oxford Photovoltaics’ US 12,288,825, granted April 2025, claims a metal halide perovskite chemically bonded to an organic passivating agent — a coating that suppresses ionic migration along grain boundaries, the mechanism that has historically aged perovskite cells in days. CubicPV’s US 12,359,122, granted July 2025, describes ionic-liquid-treated perovskite devices for spray and slot-die coating. These are nanoscale-engineered materials, just not the carbon-nanotube/quantum-dot path Kurzweil named.

The nano-fuel-cell half of the prediction has not arrived. Lithium iron phosphate batteries took about 90% of new global grid-storage deployments in 2025; hydrogen and ammonia are still in pilot stage. The U.S. alone installed a record 57.6 GWh of battery storage in 2025.

Verdict: Ahead of schedule on outcome, wrong mechanism on chemistry.

The 30-trillion-watt math.

Kurzweil’s most-cited energy line is that “all projected global energy needs of 30 trillion watts in 2030 could be met with solar power by capturing only 0.03 percent of the sunlight hitting Earth” (ch. “The Criticism from Malthus”). The arithmetic checks. The Earth intercepts about 173,000 TW from the sun; 0.03% of that is roughly 52 TW, comfortably above the 30 TW figure.

What’s off is the demand baseline. Global primary energy consumption in 2024 ran about 600 exajoules a year — roughly 19 TW continuous — and is growing at about 2.2% annually, putting 2030 demand near 21–23 TW rather than 30. Solar’s contribution averaged over the year (after capacity-factor losses) is about 0.6 TW. To get to 100% solar would require a ~35x scale-up. The doubling math says it’s possible by the early 2030s; the grid-integration math says it’s not.

Verdict: On track for the headline, conservative on the demand baseline.

Flying cars, drones, and the lock-in critique.

Kurzweil wrote that “it will become feasible to build inexpensive small flying devices for direct package delivery and larger inexpensive vehicles using nanoengineered microwings to transport people” (ch. “The Criticism from Lock-In”). The outcome is arriving early. The chemistry is utterly conventional.

Joby Aviation cleared FAA Stage 4 type certification in March 2026 and is on track to fly paying passengers in late 2026 in partnership with Delta Air Lines, beginning at New York and Los Angeles airport connections. Joby’s batteries are NMC-811 lithium-ion pouch cells — 288 Wh/kg, the same automotive-grade chemistry that powers a 2025 Hyundai. Archer Aviation has more than 100 eVTOL-related grants. The recently issued US 12,404,020 (“VTOL aircraft using large, variable speed tilt rotors”) describes a battery-powered, 500-pound-payload aircraft with two to four rotors that tilt for forward flight and “flies in an autopilot or pilotless mode.” US 12,100,850 (“Systems and methods for improved battery assemblies for eVTOL aircraft”) claims a foam-and-sleeve battery block whose thermal-isolation walls keep one cell’s runaway from cascading into adjacent cells. No microwings. No molecular self-assembly. Standard battery-pack engineering, just at a level of redundancy automotive packs never demanded.

Drones got there even faster. Zipline reports more than 1.8 million commercial deliveries; Wing has crossed 350,000. Most are scaled-up cousins of $400 hobby quadcopters, certified for routine operations beyond visual line of sight.

Kurzweil’s lock-in argument, made against Jaron Lanier’s critique that infrastructure investment freezes out new architectures, also lands strangely. Lanier was wrong about the internet — telecom was upended by mobile and fiber. But Kurzweil’s transportation argument is being vindicated by an industry that has embraced the existing aviation regulatory frame, not bypassed it. Joby’s path runs through FAA Part 23 amendments, ALC airworthiness reviews, and Type Inspection Authorization — a tighter bind to legacy infrastructure than Kurzweil suggested would matter. The system bent. It did not break.

Verdict on the flying-car prediction: Ahead of schedule (commercial passenger service in 2026, before the predicted 2030s).
Verdict on the lock-in claim: Mostly right on outcomes, mostly wrong on mechanism.

Molecular nanotechnology, energy and transportation as information technology.

Kurzweil predicted “the full advent of molecular nanotechnology-based manufacturing will bring the law of accelerating returns to physical industries such as energy and transportation” (ch. “The Criticism from Lock-In”) — and that those industries would “effectively become information technologies” (ch. “A Panoply of Criticisms”).

The first half is unmet. There is no $2-per-kilogram molecular manufacturing process. Drexler’s nanofactories remain a research target, not a product line. The second half is partly true in a way Kurzweil didn’t quite predict: what’s making energy and transportation feel software-like is autonomy and AI control, not molecular printing. Joby files patents on flight-plan generation, pilot-conflict avoidance, robotic charging, and ride-sharing optimization. Archer files patents on aircraft function prioritization and crash-detection battery isolation. The bits running on top of the lithium pack are doing the work that nanoengineered hardware was supposed to do.

Verdict on molecular manufacturing: Behind schedule.
Verdict on “becomes information technology”: Wrong mechanism, partly right on outcome.

The scorecard

What Kurzweil missed (and what he nailed)

The pattern is consistent across the batch: Kurzweil was right about direction and cost curves, wrong about the path. His framework treated nanotechnology as the keystone — the lever that would pull energy and transportation onto Moore’s-law-style trajectories. Reality used five different levers (perovskite chemistry, lithium-ion automotive supply chains, AI flight control, software certification, modular grid storage) to roughly the same effect. The trajectory was right; the bridge he drew under it wasn’t.

The clearest miss is molecular nanotechnology itself. Nanofactories, diamondoid mechanosynthesis, and atom-by-atom assembly remain research programs without a commercial pathway in 2026. Where nano-engineered materials did show up — tandem solar cells, battery cathode coatings, carbon-fiber composites for eVTOL airframes — they are subcomponents of conventional manufacturing, not the manufacturing process itself.

The clearest win is direction. A 2005 critic could fairly have said: solar will never undercut natural gas. By 2025 it does, in most of the world. A 2005 critic could fairly have said: passenger flying machines will never compete with cars. They are entering revenue service this year. A forecaster who got the curves right and the mechanisms wrong is more useful than one who got both wrong, even if specific bets — quantum-dot photovoltaics, hydrogen fuel cells for personal cars — look misdirected in retrospect.

Method note

Patent counts and abstracts came from a corpus of 9.3 million U.S. grants and pre-grants, joined to assignee records. Highly cited papers came from a 357-million-record OpenAlex index. Web searches surfaced 2025–2026 industry data on solar capacity, eVTOL certification, drone delivery volumes, and battery chemistry. Verdicts reflect evidence as of April 26, 2026.