Google’s estimate is about 1,800 Starship launches—but it is a conditional result from a launch-cost model, not a count of launches required before any space data center can begin operating. Google’s 2025 analysis estimates that reaching launch prices below about $200 per kilogram by roughly 2035 would require 370,000 tonnes of additional mass launched to orbit. At a nominal 200-tonne payload per flight, that is roughly 1,800 launches over a decade. Google’s first orbital-computing prototype has already launched, so the figure concerns a modeled path to much cheaper access to orbit, not a prerequisite for the initial experiment.
What Google’s 1,800-launch estimate means
The number comes from Google Research’s 2025 analysis of how launch costs might affect its orbital-computing proposal. In the paper’s scenario, launch prices fall below about $200/kg by roughly 2035. The model estimates that reaching that price would require about 370,000 tonnes of additional cumulative mass launched to orbit.
As an Amazon Associate I earn from qualifying purchases.
Divide that mass by a nominal 200-tonne Starship payload and the result is 1,850 flights, commonly rounded to about 1,800. Spread across ten years, that is approximately 180 launches a year. Both the payload and cadence are assumptions in the estimate; they are not a demonstrated Starship operating record or a published schedule for Google’s project.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallMost importantly, Google describes the analysis as a high-level evaluation of how launch costs could affect the proposal’s viability, not a comprehensive economic-feasibility study. The estimate is therefore not a prediction that Google itself will book 1,800 launches, nor a minimum launch count required to put its first satellite in orbit.
#1 Best Overall
- HOBBY MODEL KIT – Unassembled model packed in an envelope with easy to follow instructions. Ideal for ages 14 and up.
- NO GLUE OR SOLDER NEEDED – Parts can be easily clipped from the metal sheets. Tweezers are the recommended tool for bending and twisting the connection tabs.
- VOYAGER – 1.5 Sheet Model with a moderate difficulty level. Assembled Size: 1.38 x 1.77 x 6.70 inches.
- FROM STEEL SHEETS TO 3D – Pop out the pieces and connect using tabs and holes. Includes illustrated instructions.
- HIGHLY DETAILED ETCHED MODEL – Display your 3D model once completed - collect and build them all.
Why a space data-center prototype is already in orbit
According to a 2026 TechCrunch report, Google’s first orbital-compute prototype was built by Planet Labs and launched on a SpaceX rocket from California on October 1, 2026. Its job is to test whether a Google Tensor Processing Unit (TPU) can compute in space while receiving about one kilowatt of continuous power, handling heat and running models. The planned test runs the TPU in 15-minute bursts.
That prototype is an experiment, not a full orbital data center. A later demonstration is expected to use two purpose-built compute satellites connected by laser communications. Google’s longer-term concept envisions a network of 81 satellites processing workloads in parallel. The launch-cost estimate addresses the scale and economics of a possible future system, not the launch of the initial test satellite.
Rank #2
- Model Kit
- May Require Paints and Glues to Assemble
- Accurate Scale Model
- Detailed Instructions Provided
- Decals/Transfers Included
How Google’s launch-cost model gets to 2035
The estimate rests on an extrapolation of a roughly 20% annual learning rate from SpaceX’s historical launch-cost reductions. Applying that assumed trend leads the paper to model launch costs below $200/kg by about 2035, alongside the 370,000-tonne cumulative-mass requirement. A learning rate is a modeling assumption about how costs may decline over time; it does not guarantee that result.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteThe paper also models Starship 4 costs to SpaceX under different component-reuse assumptions. These are calculated outputs based on public specifications and extrapolations, not observed Starship operating costs:
Rank #3
- HOBBY MODEL KIT – Unassembled model packed in an envelope with easy to follow instructions. Ideal for ages 14 and up
- NO GLUE OR SOLDER NEEDED – Parts can be easily clipped from the metal sheets. Tweezers are the recommended tool for bending and twisting the connection tabs
- APOLLO CSM – 3.5 Sheet Model with a challenging difficulty level. Assembled Size: 5.07 L x 2.28 W x 3.45 H inches.
- FROM STEEL SHEETS TO 3D – Pop out the pieces and connect using tabs and holes. Includes illustrated instructions
- HIGHLY DETAILED ETCHED MODEL – Display your 3D model once completed - collect and build them all
| Modeled Starship 4 case | Estimated cost to SpaceX | What the case assumes |
|---|---|---|
| No component reuse | About $460/kg | Components are not reused. |
| 100-times component reuse | About $15/kg | Components are reused 100 times. |
| 15% refurbishment-cost sensitivity case | About $38/kg | The model includes refurbishment costs at 15%. |
These modeled vehicle costs and the paper’s sub-$200/kg target are different quantities in the analysis. The former are scenario outputs for SpaceX’s costs; the latter is the launch-price threshold used in Google’s cumulative-mass calculation. Neither figure establishes the price a customer would actually pay for a Starship launch.
What the paper’s power-cost comparison does—and does not—show
For the satellite designs examined, the paper estimates launched-power costs of about $810–$7,500 per kW-year at a $200/kg launch price. It compares that modeled range with reported U.S. spending of $570–$3,000 per kW-year for power at terrestrial data centers capable of machine-learning workloads.
Rank #4
- HOBBY MODEL KIT – Unassembled model packed in an envelope with easy to follow instructions. Ideal for ages 14 and up.
- NO GLUE OR SOLDER NEEDED – Parts can be easily clipped from the metal sheets. Tweezers are the recommended tool for bending and twisting the connection tabs.
- JAMES WEBB SPACE TELESCOPE - 2.75 Sheet Model with a moderate difficulty level. Assembled Size: 4.13 L x 2.75 W x 2.75 H inches. 1:221 Scale. 62 Pieces
- FROM STEEL SHEETS TO 3D – Pop out the pieces and connect using tabs and holes. Includes illustrated instructions.
- HIGHLY DETAILED ETCHED MODEL – Display your 3D model once completed - collect and build them all.
| Power-cost measure | Reported or modeled range | Scope |
|---|---|---|
| Orbital launched-power cost | $810–$7,500 per kW-year | Google Research’s modeled satellite designs at a $200/kg launch price. |
| Terrestrial data-center power spending | $570–$3,000 per kW-year | Reported U.S. spending for ML-capable data-center power, as cited in the paper. |
This is a comparison of power-related costs, not total data-center costs. The paper excludes infrastructure and building costs, and excludes chips because it treats chip costs as arising in both configurations. The ranges overlap, but that alone cannot establish that an orbital facility would be cheaper overall; the result depends on the modeled satellite design and launch-price assumption, among other costs not captured in this comparison.
Recommended Free Tools
The engineering questions a launch-cost model cannot settle
Cheaper access to orbit would address one barrier, not demonstrate that orbital AI infrastructure can perform reliably or economically. Google’s prototype is intended to test several challenges together:
Best Value
- HOBBY MODEL KIT – Unassembled model packed in an envelope with easy to follow instructions. Ideal for ages 14 and up.
- NO GLUE OR SOLDER NEEDED – Parts can be easily clipped from the metal sheets. Tweezers are the recommended tool for bending and twisting the connection tabs.
- HUBBLE TELESCOPE – 1 Sheet Model with a moderate difficulty level. Assembled Size: 3.00 x 2.00 x 2.50 inches.
- FROM STEEL SHEETS TO 3D – Pop out the pieces and connect using tabs and holes. Includes illustrated instructions.
- HIGHLY DETAILED ETCHED MODEL – Display your 3D model once completed - collect and build them all.
- Power and thermal management: The test must show that the spacecraft can supply the planned continuous power and remove heat from computing hardware in orbit.
- Radiation reliability: Google’s reported ground testing found a very low error probability for typical inference operations, but that is not the same as proving reliability during long, large-scale training runs in space. Google executive Travis Beals said that a multi-thousand-chip training run lasting months would be more problematic.
- Bandwidth and latency: Workloads spread across multiple TPUs or satellites need fast communication between processors. Beals noted that bandwidth and latency matter greatly for multi-rack workloads, and said Google is considering workloads that may exist in five years as well as those used today.
- Scale and operations: A short, burst-based test on one satellite does not by itself establish the performance, availability or operating costs of a multi-satellite network.
As Beals put it, “We’ve done testing on the ground, but you know, there’s no test that’s completely as good as the real thing.” That is why the first mission is a technology demonstration rather than evidence that the proposed network is ready for routine AI workloads.
Is Google’s space-data-center plan economically realistic?
The paper makes a conditional case that falling launch prices could improve the economics of orbital computing. It does not prove the entire proposal is economically viable. Its own comparison is limited to power costs and omits infrastructure and building expenses; the engineering questions of cooling, radiation tolerance, communications and reliable operation also remain material.
A useful way to judge future claims is to keep distinct measures separate: actual launch cadence and payload performance; customer prices per kilogram; modeled dollars per kW-year; and demonstrated reliability, thermal management, bandwidth and latency. A favorable value in one category cannot stand in for evidence in the others. On the evidence presented, roughly 1,800 launches is best read as the scale implied by one cost-reduction scenario—not a settled forecast for Google’s project or proof that orbital data centers will beat terrestrial ones.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
