Next flight predictions for Starship, New Glenn, Vulcan and Ariane 6
Launch cadence learning curves predict early March and April launches for Starship, Ariane, and Vulcan, September for New Glenn
Predicted dates for the next two launches:
To date Starship has launched 7 times, Vulcan Centaur twice, and Ariane 6 and New Glenn each once. Ariane 6 is officially scheduled for launch on February 28, four days ahead of the predicted date. Vulcan has a launch listed for May 2025 with a Dream Chaser payload, a month later than predicted. New Glenn at the moment is listed as launching in June 2025, three months ahead of predicted, but that date was announced in September of 2024 and there was a booster anomaly during Flight 1.
As the saying goes, “all models are wrong, and some are useful.” It will be interesting to watch the extent to which Wright’s Law can predict the launch cadence of these newest rockets over time.
Methodology
The Starship learning curve has been analyzed for completed flights 1-7, with a current learning rate of 52% and an initial time cost to launch, K, of 427 days. Ariane 6, New Glenn, and Vulcan Centaur each do not yet have enough launches to calculate a learning curve. For them the industry average learning curve values LR = 32% and K=351 days were used (see my last post for the calculation of those parameters).
Forward looking launch dates can be predicted using Wright’s Law to calculate the days between launch for each incremental flight into the future:
where y = the days between launch, x = the number of cumulative launches (1, 2, 3, …) , and n = the power law exponent. The value of n is used to calculate the learning rate LR, which is the % reduction in time between launches for every doubling of total launches.
For an upcoming predicted flight, the expected time cost of that launch (in days between launches, DBL) can be calculated and then added to the date of the previous actual launch. For each launch vehicle, the first two predicted dates are listed in the table above. With a very small number of launches, accuracy for even a few projected dates is of course quite limited, but the learning curve framework is a useful benchmark for analyzing the relative rates of improvement in launch cadence.
Erratum: The initial version of this article contained predictions incorrectly extrapolated from flight 1 of Vulcan Centaur instead of from flight 2, resulting in launch date predictions that were 30 days too early. The graph and table have been updated with the correct dates. Some text has also been added to clarify the method for predicting upcoming launch dates.