I got a lot of flack when I said this before, but I continue to think that model providers will end up being a lot like airlines: Critical for the global economy but highly commoditized, with huge capex requirements and low (and often negative) margins.
About 8 months ago, I warned that “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering.” This take was controversial at the time; now look how many people are saying it.
We're about to run back blockspace fees.
1. Early leaders (Ethereum, Anthropic) drive BIG fees
2. Price elasticity breaks (gas, LLM budgets)
3. Cheaper 80/20 (blockspace, models) pop up
4. Revenue collapses, Jevon's paradox cited
5. Jevon's paradox wins LT, but there's a gap
Citadel Securities just put institutional weight behind what the AI bulls won't say out loud.
In a new macro note titled "Tokenomics," Citadel makes the argument plainly: even the most powerful technology on earth still has to pass through the boring discipline of cost curves, capacity limits, and marginal returns.
The evidence is piling up:
– Amazon removed its token usage leaderboard
– Microsoft cancelled Claude Code subscriptions
– Multiple companies reporting unexpectedly massive token bills
Their conclusion is the part that matters.
Adoption is no longer about what AI can do in principle. It's becoming about the price and scarcity of the inputs needed to run it at scale. Compute. Power. Cooling. Memory bandwidth. Inference budgets. All real, all binding constraints.
And here's the kicker from the chart.
The Silicon Data LLM Token Expenditure Index, a benchmark for how much the market is actually spending on AI tokens, has started rolling over. Citadel reads it as a shift toward cheaper models. Companies substituting away from expensive frontier AI toward "good enough" alternatives.
That's economics 101 doing what it always does. When the price of something rises, people use less of it, or find a cheaper version.
Citadel sees a bifurcation forming. Frontier AI concentrated among a few firms with the balance sheets to absorb the cost. Everyone else quietly downgrading to simpler, cheaper models.
This is the part of every technology revolution the early narrative ignores.
The technology being real was never the question.
The question was always whether the economics could carry the valuations.
When one of the most sophisticated trading firms on earth starts writing about AI in the language of cost curves and rationing instead of limitless demand, the conversation has quietly changed.
The hype was about what AI could do.
The reckoning is about what it costs.
$squire | @squire_bot
this is the path i see.
5th wave target 20-30m.
from there, we likely see an A-B-C corrective move, consolidate / base out and restart an impulsive 5 wave structure.
let's play it out, level by level.
Two Solana-native networks, one marketplace.
UsePod now supports @JatevoId as a BYOK provider. Relay your backed capacity, resell inference, and get paid in USDC.
https://t.co/VnT0l94pWP
Piers has researched and wrote one of the most compelling reports on Space I've ever read 🚀🛰️
His report comprehensively walks through the decreasing cost to send a KG of material into space and at each level what this unlocks for the space economy
Piers also got me hooked on Three Body Problem so this is ironically up the space alley
- $1,500/kg is where we are now. This era gave us megaconstellations and earth observation. Starlink runs 75% of all maneuverable satellites in orbit and its network capacity already equals roughly 20% of the planet's real time internet traffic.
- $500/kg, expected 2027 to 2030, is when the physical economy of space starts. Orbital data centers, commercial space stations, in space servicing. At $1,500/kg the orbital business is information.
- $200/kg, 2030 to 2033, is when space starts producing. In space manufacturing crosses breakeven for drugs and materials Earth physically cannot make. The TAM for space manufactured goods sold back to Earth crosses $10B annually for the first time.
- $50/kg, 2033 to 2040, unlocks the lunar economy. The wild stat here: by 2035 propellant produced on the lunar surface could be cheaper than shipping it from Earth. That would be the first time in history any commodity made more sense to produce off world.
- $10/kg, 2040s, is the Mars supply chain, lunar mass drivers, and 45 minute London to Sydney flights. Speculative on timing but the direction feels inevitable.
Lock in @elonmusk we're going to Mars
I think we’ve reached the point where normal people can’t really determine whether new models are better than previous ones. Like Fable doesn’t seem that much better to me, but every 150 IQ person I know is like “wow the singularity came sooner than I thought”.
Model routing is great, yes. Users get accustomed to an interface but forget to tune their model to their needs. A smart model routing system makes sense then. But what about when a user really just wants the model they specify?
https://t.co/BdqUXt3LAG is a 2-sided marketplace for model pricing. Users with inference credits can come in and resell their inference onto a liquid market. AI users can change 1-line in their stack and start getting 40% off their billing for all of the frontier and open-weight models they're already using.
No compromise on quality, just a liquid price market for inference that results in real savings.