DevLazar
Note
087
Series
The Wire · Infra
Date
Nov 16, 2026
Read
8 min · 1,544 words
Rev
1
Drawn by
L. Stanković · Niš
Lab notes / The Wire

Vulcan's Return to Flight Ends Ten Minutes In, and Takes 40 Amazon Leo Satellites With It

ULA's Vulcan failed about 10 minutes into its November 15 return to flight, losing 40 Amazon Leo satellites. What it does to Amazon's FCC clock, the Pentagon's plans and a short list of rockets.

Lazar StankovićEngineer, DevLazar · Niš, Serbia

United Launch Alliance's Vulcan lifted off from Cape Canaveral at 12:40 p.m. EST (17:40 UTC) on Sunday, November 15 with 40 Amazon Leo satellites aboard, and by ULA's account "suffered an anomaly 10 minutes into the flight." The satellites were lost. Spaceflight Now calls it ULA's first total mission failure since the Boeing and Lockheed Martin joint venture was formed in 2006, and it came on the flight meant to prove Vulcan could work again after nine months on the ground. Syndicated coverage put it bluntly: Pentagon rocket provider hit by failure while SpaceX soars.

Ten minutes in, on the long burn

This was Vulcan's fifth flight and its first in nine months. The configuration was new: the VC6L flew six Northrop Grumman solid boosters for the first time, a 70-foot fairing, and a payload ULA put above 50,000 pounds, a company record. It also debuted a redesigned booster nozzle, and was the first flight of Blue Origin's BE-4 since a New Glenn exploded on its pad in May.

The boosters, the usual suspects, appeared to work normally and fell away about 1 minute 40 seconds after liftoff. The trouble came later. By Spaceflight Now's reading of the broadcast, the two BE-4s seemed to fire longer than planned, the first stage separated late, and the Centaur upper stage made unusual orientation adjustments early in a burn scheduled for 12 minutes 15 seconds. The webcast ended without saying whether that burn finished.

The weekend had warned everyone. Friday's attempt died with a power interruption at an Air Liquide nitrogen plant; Saturday's slipped on a network issue. A hold at T-minus 20 seconds on Sunday was cleared; whether it played a role is not known. ULA has promised to find the cause, and NASA Administrator Jared Isaacman offered help.

Why this was the comeback flight

Vulcan's earlier problems lived in its strap-ons. In October 2024 a nozzle came off one booster, traced to a manufacturing defect; on USSF-87, on February 12, 2026, another appeared to burn through less than 30 seconds in, though the payloads still reached orbit. The Space Force benched Vulcan for national security work "until this anomaly is solved," Col. Eric Zarybnisky said, per Gizmodo citing Aviation Week.

ULA's fix, by its account, addressed three credible causes it has not named with nozzle design and build changes and a full-scale booster firing. On Sunday the nozzles appear to have held. Something else did not.

One engine, two rockets, one nitrogen plant

The uncomfortable detail is the engine. The BE-4 also powers New Glenn, grounded since its May pad explosion, blamed on a BE-4 oxygen valve. CEO Mark Peller had said the engine and its main oxidizer valve are "common to the BE-4s we fly on Vulcan", and these engines carried Blue Origin's retrofit. Nobody has tied Sunday's failure to the BE-4; a long burn seen on a webcast is an observation, not a finding.

Still, look at Amazon's order book: 38 Vulcan launches, 24 New Glenn, 24 Ariane 6 and 13 Falcon 9. By our arithmetic, 62 of those 99 launches ride on one engine family, and neither rocket that uses it is flying today. Friday's nitrogen plant serves several launch providers too. Engineers know the pattern: two availability zones on one substation.

Amazon's clock, and how much slack the FCC left

Amazon had launched nearly 400 satellites across 14 missions since April 2025, and none since July. LV-01 would have lifted the total to 436, which leaves 396 by our subtraction. The FCC licence covers 3,232 first-generation satellites, half of them (1,616) originally due by July 30, 2026, and all by July 2029.

That midpoint already moved. In June the FCC waived it, saying the waiver serves the public interest by "promoting a second large satellite broadband constellation." The price is priority: satellites launched after the deadline temporarily lose spectral priority, restored once 1,616 are launched, or on October 30, 2027 if Amazon certifies that 1,616 have been built. Every lost batch makes the second route more likely.

Amazon Leo satellites: 396 launched before LV-01, 1,616 for the FCC midpoint, 3,232 for the full licence
Chart: Amazon Leo satellites, launched versus FCC requirements, count. Source: Spaceflight Now; Engadget; Mobile Internet Resource Center.

Satellites are not the constraint: Amazon finished its 1,000th satellite in early September and has hundreds flight-ready. Vulcan was to fly more than 1,500 of them over 38 missions, "close to half our first-generation constellation," in Melissa Wuerl's words. The next flight, LE-04, was targeted for no earlier than November 23 on an Ariane 64 with 36 satellites. Meanwhile Jonathan McDowell counts more than 11,000 Starlink satellites in orbit, and SpaceX spent the same weekend catching a Starship upper stage, as we reported on Sunday.

The Pentagon was already waiting

National security is where this costs ULA most, though we found no new Space Force statement. After a nine-month pause, the service was to weigh ULA's investigation results. In February ULA planned 18 to 22 launches in 2026, 16 to 18 on Vulcan, more than half for the Space Force, which had already moved the past three GPS III launches to Falcon 9. By our count Vulcan has flown twice this year. We found no reassignment tied to Sunday's loss.

What we could not verify

  • The cause. The BE-4 and Centaur details are Spaceflight Now's reading of the webcast, not ULA findings. If the fault is in Centaur alone, the shared-engine argument above weakens.
  • The FAA. We found no FAA statement or mishap investigation notice; who leads the inquiry is unconfirmed.
  • Value and insurance. Neither Amazon nor ULA has published either, so the financial hit is unknown.
  • Amazon's response. Amazon's launch-progress page carries a short statement calling ULA "a trusted partner" and saying Amazon believes ULA will "learn from this mission". We found nothing on schedule or fleet changes, so schedule effects here are inference.
  • Derived numbers. The 396 count, the 62-of-99 share and Vulcan's two 2026 flights are our arithmetic on cited figures.
  • The FCC terms come from Engadget and the Mobile Internet Resource Center; we did not open the order itself.
  • Starlink versus Leo. McDowell counts satellites in orbit; Amazon counts launches, so we do not chart them together.

The 30,000-foot read

Amazon controls its factory and has proved it: a thousand satellites built, hundreds waiting. It does not control the rockets, and this weekend showed how few there are and how much they share. The FCC softened the deadline; physics did not. Every month Amazon waits, a rival with its own rockets fills the sky on its own schedule.

For engineers the lesson is old. Redundancy is measured in independent failure modes, not in logos on a slide. Watch three things: whether ULA's root cause touches the BE-4, whether LE-04 flies on November 23, and whether Amazon buys more Falcon 9 launches from the company it is trying to beat. The last one would say more than any press release.

Sources: Spaceflight Now, November 15 (liftoff time, ULA anomaly statement, first total failure since 2006, fifth flight and first in nine months, booster, BE-4 and Centaur behaviour, T-20 hold, Isaacman offer, New Glenn grounding and BE-4 valve cause, 436 total, Amazon launch bookings); Spaceflight Now, November 13 (VC6L configuration, record payload, first redesigned nozzle and first BE-4 since May, nitrogen plant scrub and network slip, Cert-2 and USSF-87 anomalies, three causes, Peller BE-4 quote, none launched since July, Wuerl quote, nine-month pause, McDowell Starlink count); Amazon, LV-01 mission page (40 satellites, 1,000th satellite, flight-ready inventory, Wuerl quote, 38 missions and 1,500+ satellites, image); Amazon, launch progress updates (nearly 400 satellites across 14 missions, LE-04 date and payload, Amazon's statement after the failure); Engadget (June FCC waiver and its stated rationale, July 2029 full deadline); Mobile Internet Resource Center (3,232 and 1,616 figures, July 30, 2026 deadline, spectral priority terms, October 30, 2027 date); Gizmodo (Zarybnisky quote via Aviation Week, Space Force share of planned launches, GPS III moves to Falcon 9); Spaceflight Now, February 11 (18 to 22 planned 2026 launches, 16 to 18 on Vulcan); Spaceflight Now, February 12 (USSF-87 booster anomaly less than 30 seconds in); Orlando Sentinel via Hanford Sentinel (ULA's pledge to find the cause); ULA newsroom (Peller named president and CEO, August 17, 2026); The Next Web (Starship upper stage caught on Saturday); Tribune News Service via ArcaMax (headline, independent report of the failure about 10 minutes in); Yahoo Finance live coverage (corroborating headline on the Boeing-Lockheed rocket's mid-flight failure). CAVEATS: no FAA or Space Force page was opened, and the only ULA page opened is Peller's appointment release; ULA's statements reach us through Spaceflight Now and the Orlando Sentinel, Isaacman's post through Spaceflight Now; the flight timeline is single-sourced to Spaceflight Now; no FAA statement, insurance figure or satellite value was found; Amazon's only post-failure statement is the short one on its progress page, and we found no other; the 396 count, 62-of-99 share and two 2026 Vulcan flights are derived; Spaceflight Now's November 15 report describes the constellation as more than 3,300 satellites, while FCC coverage uses 3,232 for the licence, which we chart; Amazon's progress page, updated November 15, still lists LE-04 for no earlier than November 23; Amazon's mission page lists no photo credit, so the credit line is our attribution; the Space Foundation page cited in early leads could not be opened. Correction to early reports: the launch slipped from November 13 and flew on Sunday, November 15, not November 14.

Lazar Stanković
Engineer and writer, running a one-man shop in Niš, Serbia. I build systems that survive their launch and write down what the work teaches.
Related notes
Note 084 · The Wire · 7 min
Starship Flight 15: The Ship Comes Home From Orbit, the Booster Does Not
SpaceX caught a Starship upper stage returning from orbit on November 14, a first, after it deployed 26 Starlink V3 satellites. The booster was lost. What it means for bandwidth supply and the cost of orbit.
Note 079 · The Wire · 7 min
'Don't Ask Again' Remembers Your Password: Claude Code Approvals Are Leaking Secrets to GitHub
Apiiro found 1,379 secrets in public copies of Claude Code's local settings file, written there by the 'don't ask again' button. How it happens, who counted, what Anthropic has said, and how to clean up.
Note 077 · The Wire · 6 min
Microsoft 365's Outage Passes 15 Hours, and the Health Checks Are the Story
A Microsoft 365 outage that began November 12 ran past 15 hours. Microsoft cites servers failing health checks and has rolled back a network change, with no sustained fix yet. The lessons are about health checks and fallbacks.
← Note 086 · More Than Two Dozen AI 'National Champions', Reportedly: Washington's Export Label, Read Closely
The Wire · Weekly · Free
Notes like this one, every week.

AI infrastructure, shipping on a deadline, and the craft of building things that last. No noise.