Io-Aurelius

FFIF. This is a pretty specific question and specific questions tend to be unlikely to happen. But high demand + specific candidates already being proposed make it more likely.

Demand across a broad set of use-cases --> more unique devices. 

- Reasons for such demand: desire to personalize, large market and many sub-niches, varied and important operating environments... 

Personal electronics and environmental sensors are clear favorites. More weight on personal electronics because I have a prior against environmental sensors (e.g. for measuring sunlight) commercially benefitting from biological components.

Different from @DimaKlenchin at least in part because diabetes trackers fall into personal electronics in my head and not environmental sensors. 

Massive error bars on forecast. 

I'm following crowd out of humility, but I don't think DeepSeek v3 massively lowers Nvidia's importance; training the initial foundation model for deep seek (600b parameters!) required massive amounts of specialized hardware even if much of the intelligence was able to be condensed into a smaller model.

tm

.

larger gap

Confirmed previous forecast

.