Non mmi uses of high throughput RNGs

High throughput hardware random number generators are useful for cryptography, especially realtime encryption of high bandwidth data streams like video conferencing, broadcasts, satellite comms, as well as simulations, high frequency trading, gambling. What are other non-mmi uses of high speed random number generators? Are there practical applications for high throughput QRNGs that other RNGs like FPGA based ones are simply not suitable for?

While researching this topic I saw that hardware TRNGs made with FPGAs reportedly can get up to 10gbps throughput. Hoes does this randomness differ from the QRNGs we are using to experiment with MMI? What makes them unsuitable for MMI over QRNGs? Is it something to do with the entropy rate?

Are there any non-MMI practical applications for hardware QRNGs that FPGA RNGs simply are not usable for? I am trying to understand the use of hardware RNGs in the wider world, and what hardware is exactly used by big corporations that do things like encrypting data of millions of people in real time for example.

I was inspired by this project and the entropy league and papers such as this one where they use novel methods to provide entropy sources and do pretty cool stuff with it, and that is what got me thinking about what projects specifically geared towards high-throughput randomness would be practically useful for.

I remember reading Scotts analysis on different hardware RNGs, so I understand that there is quite a bit of depth that I do not understand here. But I think its interesting that a potentially useful project like drand only generates 1440 random 512-bit values per day, and I am wondering if there is a market demand for something similar but for practical uses of high throughput rngs.

Entropy is truly a fascinating topic and I cannot find enough in depth information available on all the different ways people harness it, so I am grateful for this corner of the internet.