Confirmed 19 Nov 2026 · Compute A one-minute read
Google booked a dedicated heavy-lift rocket to put AI data centre racks in orbit.
Google and Blue Origin announced the New Glenn launch on November 19, targeted for the first quarter of 2029. Caveat: no price, no satellite count.
Orbital compute was a research demonstration with one test chip.
It has a rocket, an orbit and a date 3 years out.
Why it matters
In October a single Google chip flew as a rideshare. A dedicated heavy-lift launch is a different commitment: it books a rocket, an orbit and a quarter. Each satellite is said to carry roughly one data centre rack. Google chose Bezos, not Musk, for the first big one.
Established
- Google's own post names Blue Origin and the New Glenn rocket.
- The target is the first quarter of 2029, into sun-synchronous orbit.
- Google calls it the first heavy-lift launch for Project Suncatcher.
Not established
- What it costs. Neither company disclosed the contract value.
- How many satellites fly. Google gave no number.
- That orbital compute works at scale. One test chip has flown.
The five-minute readThe record, the timeline, the weight and 4 sources
The record
Google Research published the announcement on November 19, 2026, in a post by Travis Beals, senior director for Project Suncatcher. Its words: "We're partnering with Blue Origin on the first heavy-lift launch for Project Suncatcher", and "Targeted for Q1 2029, Blue Origin's heavy-lift New Glenn rocket will deploy our satellites", each "with roughly the capacity of a data center rack", into sun-synchronous orbit. The post calls the project "a long-term moonshot to scale AI infrastructure in space" and says "This dedicated launch with Blue Origin represents a crucial milestone on our roadmap", adding that it "builds on our ongoing work with other satellite and launch partners". Both companies announced on the same day. Not disclosed by either: the contract value, the number of satellites, the mass or power of each satellite, the orbital altitude, the launch site, and whether the booking is firm or carries options. The prior step is on the record: in October 2026 a Google tensor processing unit flew as a rideshare on a Planet mission launched by SpaceX, and Google reported running a model on it in orbit and getting an answer, which Edition 067 carded on October 30. Blue Origin's own statement is not quoted in the trade coverage we read, and we found no Blue Origin release carrying terms.
How it got here
- 2025
Google announces Project Suncatcher, framed as research into scaling AI infrastructure in space on solar power.
- Oct 2026
A Google tensor processing unit flies as a rideshare on a Planet mission launched by SpaceX. Google runs a model on it in orbit and gets an answer. Edition 067.
- Oct 30
Google contracts 702 megawatts of carbon-capture fuel cells for a $17 billion Wyoming site, one of several moves to secure terrestrial power. Edition 076.
- Nov 19
Google and Blue Origin announce a dedicated heavy-lift New Glenn launch for Suncatcher satellites into sun-synchronous orbit, targeted for the first quarter of 2029. No price and no satellite count.
Blue Origin, which now has a named customer on New Glenn's manifest; SpaceX, which launched the test and did not get the contract; the Federal Communications Commission and the bodies that must license an orbital compute constellation; Nvidia, since Google flies its own silicon; and Amazon, whose founder's rocket company is paid by a cloud rival.
What to watch
- Would strengthen itGoogle or another hyperscaler orders a second dedicated launch for orbital compute, or files for spectrum or launch licences naming more than 10 compute satellites, by December 31, 2027.
- Would weaken itThe Q1 2029 target slips by more than 12 months, the booking is reported as cancelled, or Google's own posts stop describing Suncatcher as on a roadmap, by June 30, 2028.
- Would strengthen itGoogle awards a further Suncatcher launch to a provider other than SpaceX, or SpaceX publicly disputes the choice, by December 31, 2027.
Why it weighs this much
Cost is medium and partly inferred: a dedicated heavy-lift launch is an order of magnitude dearer than the October rideshare, but no price was disclosed, so we will not pretend to the figure. Commitment is medium: a joint announcement naming the rocket, the orbit and the quarter is a real booking, and Google calls it dedicated, yet launch contracts routinely slip and we do not know whether this one is firm or optioned. Reversibility is moderate, because a 2029 launch can be rescheduled or cancelled with deposits lost and nothing else. Persistence is years by construction: the target is 3 years out, and the programme it serves is described by Google as a long-term moonshot.
The words
- Heavy-lift launch
- A rocket class able to carry roughly 20 tonnes or more to orbit. A dedicated one carries a single customer's payload, which costs far more than sharing.
- Sun-synchronous orbit
- A polar orbit that keeps a fixed angle to the Sun, so a satellite sees near-constant sunlight. For solar-powered computing that means continuous power.
- Rideshare
- Buying space on someone else's launch. It is the cheap way to fly one experiment, and the reason the October test was not a commitment.
Sources
- BGoogle Research blog by Travis Beals: Blue Origin and Project Suncatcher heavy-lift launch partnership, Nov 19, 2026blog.google
- SVia Satellite: Google orders heavy-lift launch for Project Suncatcher from Blue Origin, Nov 19, 2026satellitetoday.com
- SSatNews: Google selects Blue Origin's New Glenn for 2029 AI satellite launch, Nov 19, 2026satnews.com
- SSeeking Alpha: Google taps Blue Origin's New Glenn for Project Suncatcher, Nov 19, 2026seekingalpha.com
Your call
Which explanation do you believe?
A call is final. Other people's calls show after yours.