A lot of this number juggling seems to go on with so-called high speed films. Even Kodak's now defunct 10,000 ISO (Hah!) surveillance film didn't have a true speed of much more than 800 ISO.(posted 8622 days ago)
If you look at Fuji's published curves for Neopan 400 and NP1600, where's the difference?. The toe rises at exactly the same Lux/second exposure point.
It's reckoned that it takes a minimum of 3 photons to expose an individual AgX crystal, and then there's every chance that the displaced electron will wander around and pop back into the dislocation a short time afterwards. This must set a physical limit to film speed.