
allocator, buildor, trador, unit-01, nfa, imho, dyor, low iq chief intern @temple_ny, strategery @orderx, ex @bnyglobal, @lumidawealth
Double Scalar Multiplication
Joined February 2022
- Tweets5.3K
- Following441
- Followers3.4K
- Likes14.9K
the whole point of tokenization is to free trapped capital.
mobility.
collateral moves to any counterparty, at any hour, in seconds. post less margin overall and meet margin calls without selling positions or prefunding accts. today only about 15% of eligible collateral is used.
liquidity.
idle cash and fund holdings can be posted, lent or deployed instead of sitting still. earn yield on balances that earn nothing today, and do 40 to 60% more repo on the same balance sheet.
certainty.
trades settle instantly and in full. release the capital and backoffice cost held against settlement fails, counterparty risk and recon, an estimated $15 to 20B a yr.
2
138
John Nash Cutrone, CFA retweeted
Shield is acquiring Echodyne to expand our Mission Systems business with software-defined radars, engineering expertise and manufacturing capacity.
The $5.3B transaction is supported by a $2.5B raise, including Series H funding at a $28.8B valuation, and will accelerate our efforts to meet growing demand for multi-mission capabilities.
Echodyne’s industry-leading technology is used in hundreds of counter-drone and security systems worldwide, where it helps track drones, secure airspace, and protect people and infrastructure.
Full release: bit.ly/4iYGLWc
Funding includes a structured equity investment from @Blackstone and a Series H co-led by @disruptivetech and Innovation X Capital, with participation from @adventintl, the Strategic Investment Group of JPMorganChase's Security and Resiliency Initiative, @SnowpointVC, Apandion, Center15 Capital, and others.
Autonomy for the world. The greatest victory requires no war.
3 sigma move for nasdaq this morning
fear mounting that dems will go after data centers, hurting chip and neo cloud businesses
leading to risk off across nearly all asset classes and vix in high 20s
happening now bc state governor races concluded throughout yesterday afternoon, especially tx
1
3
142
if abbott wins re-election in a couple months, the amount of data centers in texas would double from 300 to 600
$QQQ will be up or down 10% based on what happens
29
yep
if abbott wins re-election in a couple months, the amount of data centers in texas would double from 300 to 600
$QQQ will be up or down 10% based on what happens
5
275
there will be fannie mae for the ai buildout
warsh and bessent have said we need to win the ai race
us govt will own tech companies and data centers, similar to japan
they'll need stablecoins to absorb tbills and tokenization to free trapped capital
1
97
hyperliquid + drw + cme
my take aways from odd lots interviews with jeff, don, and terry
1. all 3 say 24/7 trading is coming but disagree on why it's not here yet
jeff: tech and uptime
terry: banks are closed on weekends, can't prefund a $76t equity market
don: MMs don't want to quote saturdays
2. the perp fight
terry: perps are swaps and only for leverage/speculation
don: perps are futures, leverage is a design choice
jeff: a perp just needs an index price and hard oracle limits
terry and don agree gpu futures should list. don says coreweave bonds near 15% show why compute owners need a hedge
3. the weekend collateral problem is simple. a clearinghouse calls margin when prices move. on a saturday the fed wire is off and banks aren't sending dollars
so if a market gaps on sunday you can't post cash until monday. the clearinghouse eats that risk or makes everyone prefund extra margin all week
prefunding a $76T market just to trade two extra days ties up a lot of capital for little volume
the fix is collateral that moves on weekends. tokenized treasuries, stablecoins, onchain repo
whoever lets a firm post treasuries at 2am sunday and get margin credit right away removes the main reason markets close
I asked @CMEGroup CEO Terry Duffy if, in his view, there are literally any laws anymore when it comes to financial regulation.
122
compute markets feel forced
you don't need perps on gpu hours to hedge or speculate on compute
an h100 hr can be nearly all explained by interest rates and commodity inputs
depreciation + cost of capital = ~57% of h100 run cost. power is ~12%
coreweave's ig gpu-backed loan priced at sofr+225. 2 years ago the same collateral paid sofr+600 to 1300
$197B of gpu financing since 2023, 92% of it in the last two years
take borrow rates from 5.6% to 10% and modeled neocloud margins go from ~15% to ~5%
every gpu hour is a duration trade. hedge it with onchain treasuries
a data center uses 27T of copper per mw, 40-60T if it's built for ai training. half of tin is used for solder. silver, tungsten, nickel, molybdenum, etc...
far easier to hedge compute with onchain commodities and treasuries.
1
88
“The most important number in GPU pricing right now isn't $/GPU-hour.
It's the gap between what a hyperscaler pays to borrow (~90bp over Treasuries) and
what a neocloud pays (~1,100bp).”
I've been critical in the past of AI compute indexes and the willingness to price compute based on the spot rental price, but now that the AI reset is sinking into the repricing of AI buildout debt its time to realize that the cost of capital is what determines the price of compute.
The most important number in GPU pricing right now isn't $/GPU-hour. It's the gap between what a hyperscaler pays to borrow (~90bp over Treasuries) and what a neocloud pays (~1,100bp).
The cost of compute is now set by the cost of capital, and the spot rental rate no longer tells you much.
The price of an H100 isn't one number. The same chip in the same rack costs a different amount depending on whose balance sheet it sits on. A ~1,000bp funding gap is a structural cost advantage that no amount of operating skill closes.
For example, take one GPU-hour apart. Most of it isn't power or space. It's paying back the chip, the interest on the loan that bought it, and the return the equity expects. A $40–55k all-in GPU, 75% debt-funded, at ~7,000 utilized hours a year is a financing instrument that happens to do matrix multiplication.
As the +500bp on neocloud debt (10%→15%) adds $0.20–0.30 to the break-even cost of every GPU-hour, which is then added on to the rental price.
That's 8–10% of the H100 index, from a credit spread alone, driven almost entirely by the cost of capital.
That means that the compute spot index is a derivative of the AI infra debt credit curve and the squeeze on neocloud margins.
2
119


















