He sold 116,129 shares on 23 and 26 October for about $4,025,703, leaving 10,167,399 held directly. The ordinary reading — a founder diversifying mechanically under a plan, carrying no information — is the correct one.

This is the exact shape that usually rewards a second look, and that is why it was tested. A founder and 10% owner, selling into a 188% one-year gain, with the stock within about 5% of its 52-week high of $36.97. Three things could have turned it into a finding:

• A plan adopted after the run. It was adopted 17 June 2026, before the bulk of the move, and is reported on every tranche. Fails.
• Acceleration against his own cadence. The earlier 2026 rounds were about $8.6m and $5.6m; this one is $4.03m. The trend is down, not up. Fails.
• A residual small enough to read as an exit. He retains 10,167,399 shares — roughly 88× what he sold in this filing. Fails.

The execution settles it. The lots are 63,275 / 2,885 on 23 October and 1,100 / 47,666 / 1,203 on 26 October — the fingerprint of stacked limit orders filling at stepped prices, not of decisions being taken. There is nothing here but a plan running.

Recorded because the alternative discipline — treating every founder sale near a high as a signal — is the single most likely way this report would be wrong over a year.