> Your logic is backwards.
I think we're essentially saying nearly the same thing and agree it eventually leads to a similar outcome. The difference is in part saying it in different ways and in part addressing the flow and feed backs at different points. Ultimately there are three key players in an ecosystem flow that supplies end-user demand:
[NVidia (NV)] <-> [HyperScalers (HS)] <-> [Frontier Labs (FL)]
TSMC is upstream of NV but they're diversified enough to be sheltered, and end user demand is the ultimate gating factor. Most people agree that in the long-term (10+ yr) AI compute services will be a very big viable business. What you and I understand which some don't, is that the Bubble Burst scenario doesn't require growth in AI demand to ever drop. A big crash can be triggered just by the growth rate in cash flow from AI end-user demand tapering from super-linear down to strongly linear.
Much like 2008, the lenders supplying the build-out got too speculative. The fact that 10 years from now all today's AI investments and loans will look like incredibly cheap, outstanding investments doesn't matter if over the growth rate tapers from super-linear to just linear growth for 6 months anytime over the next 12 to 48 months. That's enough to trigger the Bubble Burst spiral due to leveraged and circular investments. The risk is entirely near-term timing and relative over-valuation, not absolute value in the long-run. I think some are assuming we're saying growth or value will decrease in absolute terms. I'm certainly not. I think the growth rate in actual cash flow from AI end-user demand will not only NEVER go negative, it will continue to increase every month for the next decade - and there will still be a big crash - simply because too many speculators made leveraged financial bets which require near-term growth rates to remain consistently super-linear.
> Frontier labs are losing money and are most likely to go bankrupt and be absorded by bigger companies.
Anything's possible but I don't see how that's likely. It really depends on who is left 'holding the bag' on committed contracts and callable loans. While all the Frontier Labs have essentially taken what nets to equity-secured loans from NV, AFAICT it's not collateralized in a way that allows NV (or HS) to force FLs into bankruptcy. Creditors with capital were willing to take FL equity for cash, making them investors, not creditors. After their IPOs, that will change. FL sales will continue to steadily grow but not at rates required to make their equity worth what investors thought it would be worth in the time frame they expected/needed.
> Hyperscalers ... already control the frontier labs.
Through what legal or financial mechanisms is this control? If end-user demand continues to grow super-linearly, FLs will need more compute from HS than they get. In that case, HS will continue to have pricing power due to scarcity increasing. If future end-user demand growth rates taper from super-linear to just strong linear, FLs will need less compute than they've projected and HS pricing power will erode. This is the feed back point most likely to trigger the Burst Spiral (more specifically, watch things like the spot market for 12 month future contracts for B300 computer/hours).
Hyperscalers ... are ultimately profitable
Unlike FLs, HS are collateralized to their bankers, their GPU and their power suppliers in ways which CAN force bankruptcy. If the growth rate in demand tapers, even a little, prices will fall. HS have bought, and committed to buy, more GPUs, power and real estate sooner than they are going to need it from bankers who can force them into bankruptcy. The HS have secured their loans by pledging their GPUs, committed power contracts and real estate/physical plants as collateral.
When cash flow is high and the growth rate of cash flow keeps increasing, lenders will loan $2 to $5 for every dollar of 'sellable asset' the loan is secured by. The HS are already extending their GPU depreciation time frames to SIX years which inflates the forward book value of their GPUs - so they can borrow more at cheaper rates. When the growth rate of cash flow starts to taper from super-linear to just very good, lenders start to demand more collateral per dollar (or capital costs go up). Both are already happening. NV and FL will be damaged but HS are most likely to get screwed because NV and FL have investors (who just lose lots of money) but HS have cash creditors who can force bankruptcy and will panic when they realize they're holding billions in loans secured by evaporating collateral value - much like "Collateralized Debt Obligation (CDO)" contracts triggered the 2008 financial collapse. The irony is those tens-of-thousands of residential houses backing the 2008 CDOs were still livable homes and almost all are worth more today than they were at the peak of the bubble. No one thinks the mountains of GPUs (and future GPU deliveries) HS overpaid for will be worth more than ~50% in 3 years.