Nuclear Missile Warning Satellites: The US Just Ended Its 56-Year Space Vigil for Nuclear Launches — and the Replacement Is Already Watching

For 56 years, a chain of satellites in geostationary orbit kept watch for the one thing no government ever wanted to confirm: nuclear fire leaving the ground. They never blinked. They never slept. They circled 35,786 kilometres above the Earth and stared at every continent simultaneously. This week, the US military officially ended the nuclear missile warning satellite program that started in 1970.

Not because the threat has gone away. Because something better took its place — and an even newer system is already on the launchpad.

What do nuclear missile warning satellites actually do?

They hunt heat. A ballistic missile engine produces an enormous infrared plume the moment it ignites — visible from space within seconds of launch. These satellites, fixed in geostationary orbit, stare down at their coverage zones continuously. When that signature appears, automated systems alert command centres. Decision-makers get minutes, not hours, to respond.

That window matters more than almost anything in nuclear strategy. A submarine-launched missile fired from certain positions in the North Atlantic can reach Moscow in under 15 minutes. A Russian ICBM from Siberia reaches Washington in roughly the same time. The difference between 12 minutes of warning and 7 minutes of warning is the difference between a considered decision and a reflex. These satellites existed to buy those minutes.

How did the Defense Support Program begin?

The first DSP satellite launched in November 1970 — three years before the first mobile phone call, a generation before the internet. The Cold War was at its most dangerous, Soviet nuclear doctrine was evolving fast, and the Pentagon concluded that watching from orbit was the only truly reliable early warning method. Ground-based radar could be jammed, spoofed, or destroyed. Space could not.

What replaced DSP — and how much better is SBIRS?

The Space-Based Infrared System has been the primary missile warning network since 2011, and the improvement over DSP is almost incomprehensible. DSP used scanning sensors — sweeping back and forth across its coverage zone like a searchlight. SBIRS uses staring sensors: cameras fixed permanently on a region, capturing every infrared event with no moving parts. It can detect a launch and simultaneously estimate trajectory, missile type, and likely target area — all in the seconds after ignition.

SBIRS also uses highly elliptical orbits (HEO) — specifically to improve coverage of the northern polar regions, where Russian and Chinese ICBMs would fly toward North America. That is not a coincidence. It is geometry applied to nuclear deterrence.

What comes next — the Next Generation OPIR program?

The US military has confirmed a next-generation system is in development: the Next Generation Overhead Persistent Infrared program (Next Gen OPIR). Details are classified. What public disclosures confirm is that it's specifically designed to handle a threat DSP was never built to track: hypersonic glide vehicles.

A ballistic missile follows a predictable arc — launch, coast through space, re-enter. SBIRS handles that. A hypersonic glide vehicle is different. It skims the upper atmosphere at Mach 20+, manoeuvres unpredictably, and never climbs high enough for easy GEO tracking. The new system must see more of the electromagnetic spectrum, process data faster, and track objects that are deliberately designed to be invisible to its predecessor.

Has this system ever come close to triggering a nuclear war by accident?

Twice, memorably. In 1979, a training tape simulating a Soviet nuclear first strike was loaded into a live US detection computer. For several minutes, NORAD genuinely believed the country was under attack. Bombers taxied. The error was caught in time.

In 1983, a Soviet early-warning satellite — their equivalent of DSP — incorrectly detected five incoming US ICBMs. The duty officer, Lieutenant Colonel Stanislav Petrov, decided on instinct it was a false alarm. He had no confirming data. He was right. His decision not to report it up the chain of command may have prevented nuclear war. He later said he thought: five missiles doesn't make sense. If you want to destroy a country, you don't send five.

Why does this matter in 2026 specifically?

The world has more nuclear-capable states than it did in 1970. North Korea has successfully tested ICBMs capable of reaching the US mainland. China is expanding its arsenal at a pace not seen since the 1960s. Russia has fielded hypersonic weapons explicitly designed to defeat existing detection architecture. The threat has not shrunk. It has multiplied and changed shape.

Retiring DSP while SBIRS is operational and Next Gen OPIR is on schedule is the correct strategic move. Maintaining 50-year-old satellites costs money, diverts maintenance resources, and creates single points of failure in a system that cannot afford any. The US Space Force — which now operates these capabilities — confirmed the program end follows a planned, phased transition. The vigil is not over. The hardware is just newer.

Can you actually see these satellites from the ground?

Some of them. Geostationary satellites don't move relative to the ground — they hover over the same equatorial point at all times. That makes them locatable with binoculars if you know exactly where to look. Amateur astronomers have spotted SBIRS satellites. What they are watching, however, is entirely invisible to you.

The SkyLens live tracker shows satellites currently orbiting Earth across all orbital classes. GEO satellites — the class these military birds operate in — appear as stationary points far above the dense cloud of LEO traffic. Right now, as you read this, something in that ring of geostationary orbit is watching for heat. Learn more about orbital mechanics and what each altitude band means for the systems operating there.

For those interested in what military aerospace sensors have captured that remains unexplained — including cases where surveillance assets recorded objects that no official program has yet accounted for — the PURSUE declassified archive holds 375 records across five public releases, spanning 2013 to 2026.