Between 1946 and 1961, the United States spent roughly $1 billion (about $11 billion today) on the Aircraft Nuclear Propulsion program — a serious, sustained effort to build a bomber that could stay airborne for weeks on a reactor the size of a refrigerator. The earlier NB-36H carried a reactor as a passenger; the Convair X-6 was meant to be flown by one. Two parallel engine programs raced to the finish line: General Electric's Direct Air Cycle (X-39 turbojets coupled to the HTRE-1, -2, and -3 reactors at the National Reactor Testing Station in Idaho) and Pratt & Whitney's Indirect Air Cycle, which used a liquid-metal loop to isolate the crew from the core.

GE's approach actually worked. On January 31, 1956, HTRE-3 ran a J47 turbojet entirely on reactor heat — the first jet engine in history powered by fission. The reactor used beryllium oxide moderator and ran at over 1,000°C, hot enough to deliver useful thrust through nickel-alloy heat exchangers. P&W's indirect cycle used molten sodium and later examined fluoride salts — the same chemistry that would later define Oak Ridge's MSRE.

So why did it die? On March 28, 1961, President Kennedy canceled ANP in a single paragraph. The stated reasons:

• Shielding mass. Protecting the crew required so much lead and polyethylene that payload collapsed. Some designs proposed leaving the shielding off and using elderly crews — seriously.
• The ICBM made it pointless. Atlas and Polaris could deliver warheads in 30 minutes. A bomber that loitered for two weeks was solving 1948's problem.
• Direct cycle was dirty. GE's open exhaust irradiated air molecules. Every test flight would have left a trail of activated nitrogen-16.
• Materials. No alloy reliably handled 1,000°C oxidizing flow for thousands of hours. Heat exchanger creep was the wall.

Why it matters now. Strip away the bomber mission and look at what ANP actually proved: compact, high-temperature reactors can drive aviation-grade turbomachinery. That capability is suddenly interesting again for three reasons.

First, the materials problem is solved. Modern SiC-SiC ceramic matrix composites, already flying in GE9X turbines at 1,300°C, would have been miraculous to Convair's metallurgists. Inconel 740H and ODS steels make P&W's indirect loop genuinely buildable.

Second, the reactor architecture is mature. P&W's molten-fluoride concept was the direct ancestor of today's MSR designs from Kairos, Terrestrial Energy, and Copenhagen Atomics. A 50 MW thermal MSR weighing under 10 tonnes is no longer fantasy — it's a funded engineering problem.

Third, the mission has returned. DARPA's DRACO program (Lockheed/BWXT, first flight slated 2027) is building a nuclear-thermal rocket using HALEU. The Air Force Research Laboratory funded a nuclear cruise-missile study in 2024. China's revealed work on a thorium-fueled cargo airship in 2023. And for ultra-long-endurance ISR drones — picture a 90-day loiter over the Arctic — chemical fuel simply cannot compete with fission energy density (roughly 2 million times that of jet fuel).

The crew shielding problem that killed manned ANP disappears when there is no crew. The dirty-exhaust problem disappears with the indirect cycle that P&W had already designed. What's left is an engineering project that Convair, Lockheed, and GE got within a heat exchanger of completing in 1961 — and which we now have the materials, controls, and unmanned platforms to finish.