yes that’s exactly the plan and the reason these companies are paying profound premiums back to nVidia. If they don’t pay up nVidia will just announce another spark roll out and scare the crap out of them. They know as soon as newer computer hits the consumer market they are dead in the water.

You sort of can with solutions like MeshLLM

I actually explain exactly the plan here. I wrote an article explaining it. They intend to use several channels to socialize losses. The big ones are the front door. Literal bail outs and subsidies on multiple levels. The back door is passing it on to passive investors. At the beginning of next year we will see a near total economic collapse as this bubble bursts. Once it does passive investors will be wiped out. A whole generation of retirement savings and pension plans wiped out while simultaneously the fed will bail out the major players. Here is the article https://themildtake.com/articles/2026-06-05-priced-for-rescue/. Most people picked up on this when SpaceX went public. What they didn’t mention was even if your in the S&P that only shields you from a single bad player and a minor one at that. Microsoft, oracle, google all these big hyperscalers and service providers renting that compute are all going to take a dive.

up all RAM they ensured that you can’t run much more than lobotomized models. this is what I mean they’ve chosen to target a different consumer and rightfully so in that they are paying a premium for that product but they are also being subsidized by the same vendor so it’s not clear that they are paying more due to the bizarre circular funding pattern.

Ok so this is interesting. I am not sure this number is accurate I haven’t fully vetted it. I do this same type of forecasting. Morgan Stanley said for current capex commitments by 2027 4th quarter they needed to be 700 billion in profit. The problem is to meet that number they would need full market saturation and another 3 trillion in capex for build out. That means the entire accessible market would be using AI. They have another problem that more than half the world cannot access AI in its current form. The tooling these companies are using are wasteful and don’t meet the bandwidth needs of mobile first countries. So until the industry adopts things like https://codecai.net/ half the world’s markets are inaccessible. Finally they are under attack from all sides even if they had the money they may not have the ability to build out the AI infrastructure because that build out is becoming untenable for most communities. Even more the open source models and low cost AI are already in someways encroaching quickly on frontier systems. If you look at it right now the only reason you don’t have home AI as a commonplace solution is because the manufacturer has chosen not to sell it to you.

I’ve been doing a project https://themildtake.com/articles/2026-07-04-a-declaration-of-independence/ ever since Independence Day and I have been shocked by how insane the open-source tooling out there is now. NextCloud and Collabara alone solves 90% of company’s needs. ERPNEXT covers a huge swath in mid-range needs. I combine it with a mailcow server and Authentik for SSO and I am not sure how much else most businesses need. The whole stack is zero cost.

Because both major GPU manufacturers are controlled by the same family. The primary model repo is controlled by that same family now. Both major pipelines are controlled by that same family. ONNX and the diffusion pipelines are maintained by Microsoft. The alternatives are controlled by hyperscalers (with the exception of civitai I guess). All frontier model pipelines are directly controlled by a single family in China. So its very much possible that you could be banned from accessing them yes. We really don’t need a pirate instance we just need AN instance at this point.

I had a problem with DNS over TLS with PIA this sounds very similar. The conflict was DNS encryption type. I use Gluetun infront of PIA and the DNS resolver defaults to DNS-over-TLS on TCP 853. Through PIA’s tunnel, port 853 is blocked or doesn’t work reliably. SO I switched DNS to DNS-over-HTTPS on port 443 (I used Cloud9), which PIA allows and Gluetun can actually use. Not sure if this is you issue at all but if you have a common ARR stack with a vpn through gluetun this very possibly could be

The AI itself doesn’t need self-awareness to kill us all. It just needs an idiot with too much money and even less self-awareness controlling it. We are already headed towards disaster in the US. https://themildtake.com/articles/2026-06-05-priced-for-rescue/