Every figure in this work is computed from a model rather than quoted from another write-up, and known weak points and open questions are listed at the end, in order of how much they would move the answer. If you find an error, send it — corrections are credited by name.
Working analysis · Updated November 7, 2026
Solana Validator Economics
What actually pays, what it costs, and where the stake comes from — after Alpenglow, the SIMD-0525 slot-time ramp, the SIMD-0550 disinflation vote, and the November 6 SFDP rewrite. Breakeven after Alpenglow is 65,078 SOL at the $30,000 hosting case and 56,160 at a measured $21,636 contract — at every slot speed, because the Validator Admission Ticket scales from 1.6 to 0.8 SOL per epoch as slots shorten, confirmed in the shipped Agave code. Before activation on October 21, per-slot voting at 200ms ran at ~113,000–121,000. A thirty-day October settlement record measures revenue below the model and puts the measured post-Alpenglow figure at ~70,000–76,000. Includes the commission asymmetry, a reconciliation against the disinflation proposal’s own breakeven model, the storage and indexing side priced against a public per-byte floor, and settlement records spanning all four slot-time stages. SGP-0002 passed on August 28 at 67.00%; SGP-0003 failed. New in this revision: the Foundation’s November 6 delegation rules — 100 validators, a 50,000 SOL cap, twelve-month terms — under which the program carries at most a third of a breakeven book, and only for a year.
Field notes · August 18, 2026
When the Instrument Lies
An agave validator crash-looped on missing XDP capabilities. The fix took twenty minutes; proving it had worked took most of a day, because all three tools that could verify AF_XDP report “off” on a healthy node. Ten days later a health check that had silently done nothing for weeks turned out to differ from the documented string by one capital letter. Why absence of signal is not signal of absence — and why a monitoring check that has never fired is not evidence of health.