Learning to let go should be learned before learning to get.
Life should be touched not strangled.
- Ray Bradbury (writer)
The intellect can be a great danger to seeing things the way they are. It rushes to rationalise and in this rush leaks/drip-feeds made up reasons for the way things are that you - “the thinker” interpret as real.
The moment you “think” about the way things are, you stop seeing the most primary and essential expression of how they ACTUALLY are.
Somewhere in the refined delicate balance between the pure seeing (awareness) and reasoning about this seeing lies the path to infinite intelligence.
PL retweeted
You’re 400 million times closer in size to the observable universe than to the Planck length.
Let that sink in.
PL retweeted
We’re committing a portion of Anthropic’s shares to the Anthropic Public Benefit Fund, a new wholly owned entity dedicated entirely to the public benefit.
It will start off with 5% of our shares. For every 50% increase in the company’s value over a set baseline, we can contribute up to an extra 2%, subject to our board’s discretion.
In total, this means that roughly 20% of Anthropic’s shares are committed to the public good. Our co-founders have already pledged to donate 80% of their equity, and around half our employees are donating a portion of their equity as part of a donation matching program we’ve run since our founding.
In the post below, we share more details about our new fund.
anthropic.com/news/public-be…
the most thorough deep dive on $CRWV. updated today by @RHouseResearch
"Blackwell-era contract.. TLDR: 20% unlevered IRRs with assumptions including: - 5 yr. contract term - $27MM/MW in capex - $12MM/MW in rev. (2.25 yr. rev. payback) - Utility ($0.075 c/kWh, 85% utilization, 1.4 PUE), co-location ("/MW w/ 3% annual escalator), and cust. support expenses in aggregate avg. ~24% of rev. (resulting in ~76% contract-level EBITDA margins) - Zero prepayment"
Published some thoughts on substack (no paywall) around CoreWeave's unit economics, margins, use of debt, and what I think the real risks to the business are
open.substack.com/pub/ritten…
10% is the yield on the junior tranche alone not "cost of capital" lol
the 60bn deal is -
$42B senior debt (70%): Yield not disclosed. Benefits from Broadcom's residual-value support.
$18B junior debt (30%): Subordinated, higher-risk financing. A part of this is -
$9B Blackstone allocation (15%): Offered a 10% coupon.
This tranch carries highest risk so priced higher.
The junior tranche is NOT unsecured either: it is backed by chip-lease cash flows and exposed to Anthropic's credit risk. Its defining disadvantage is subordination and the absence of Broadcom's guarantee.
For comparison in June 2026, Project Big Sky debt raise that Anthropic did was -
The FT's original June 9 report on Project Big Sky gives this breakdown:
June tranche Amount Rate
A1 senior bank notes$6BTreasuries +100 bps
A2 senior asset-backed notes$24B5.75% yield
Junior debt$4.4B8.5% interest rate
Since June, lets look at movements in some relevant debt metrics for this in general:
Using the exact dates of the June transaction and the latest available October observation:
ICE BofA BB Index June 9 October 15 Change
Effective yield 5.96% 6.84% net change: +88 bps
Option-adjusted credit spread 168 bps 197 bps net change: +29 bps
So: taking June junior tranche figure
8.50%+0.88%=9.38%
and the current Blackstone one is priced at 10%.
There is hardly anything implied in this pricing or deal about Anthropic's "deteriorating" or otherwise credit risk. Seeing as they are raising via debt at market rates, there's little to nothing implied in this deal about them "hedging" or "not hedging" any IPO timing or otherwise.
In fact, according to the FT's October report, the $18 billion junior tranche was expected to be marketed after Anthropic's IPO, when lenders would have access to fuller financial disclosures. The debt syndication is being prepared before the IPO; that does not mean the full $60 billion has already been funded. Anthropic announced its expanded Broadcom compute partnership in April, and Broadcom, Apollo and Blackstone launched a $35B infrastructure financing programme in June.
Why would Anthropic necessarily wait for the IPO to finance committed computing infrastructure? Debt and IPO equity serve different purposes. Financing identifiable infrastructure assets with debt preserves equity capital and avoids unnecessary dilution.
Neither the pricing nor the financing sequence establishes that Anthropic's creditworthiness has deteriorated or that management doubts the IPO will proceed.
Interesting to see FUD getting more and more widespread and entrenched. The mood swings of this market :)
first time in ages, probably since tariffs panic (which actually had a fair basis) that I see bearish sentiment so entrenched and the amount of FUD on AI related fundamentals and capex and debt out of line with at least what objective measures of the same are showing at the moment.
also when I say fundamentals i mean actual earnings, and re debt - actually looking at a balance sheet of entities in question, which are being subjected to FUD (eg $CRWV) and assessing likelihood of default etc. This is not to say $CRWV is a great equity long lol. It's not... but the FUD over their risk of default is materially overblown.
Regardless, market is speaking and price is the primary input into any meaningfully profitable trading signal. All the unprofitable ones use fundamentals alone ;)
Sometimes market anticipates broadbased general tops in industries even with a lead of a year+. It is definitely possible in this instance, and should not be underestimated.
However, based on current evidence of underlying metrics that track spend, AI usage metrics growth etc I am leaning towards this being a shorter violent correction rather than the beginning of a broad based bear market likely to last year+.
Average stock is already obliterated and with breadth metrics so poor already, a bear market in $QQQ would have to be super narrow, which could other things being equal imply it could be more violent and fast. $QQQ seems likely to mount some kind of selloff attempt at or near 200day MA and vs the "Leopold" low, so how it trades there becomes of interest to the l-t trend.
Otherwise, if this were to reverse higher more abrutply, that would actually put the market into an attractive swing long environment into year end say Nov-Dec. Earnings season upon us, with "consensus" skewed bearish - we've been here before - typically in the final innings of a correction.
Taking it a day at a time, nothing to do here, waiting for 5* mean reversion longs at this stage. We had an episode on Thursday with some 4*+ intraday mean reversion long setups, and probably mid tier for 3-5 day hold, which ended up failing Friday.
Alot of questions in trading, especially on position management, where all the complexity in decision making typically lives - have no categorical "right answer".
The answer depends on whether you are optimising for the average or the outlier part of the R returns distribution,
what your projected time horizon is - this is a function of how the edge you are betting best distributes -> intraday, intra-week, multi week.
These questions have objective optimal answers. It's just time consuming, thorough and burdensome to do the looking. Categorical stances such as "day trading vs swing", "trailing vs selling into strength" seem like noob sounding all-or-nothings, which probably won't have a meaningful positive bearing on actual trading performance.
don't forget the frequently omitted/neglect HIDDEN state in feature research.
a feature can be a signal enhancer (+EV) or signal detractor (-EV) OR its not_applicable or not_relevant (Ie outside of context pertaining to the feature you are researching). This is a critical omissio, you can get features that improve the overall signal so its better to have them than not. but when you look deep under the surface, they've amplified edge on predicting your AVERAGE setup outcome, while they may actually smudge or down-out your top-of-the-range setup outcome.
Firstly, there is information that is relevant to the "information subspace" within which you are operating.
so the first node/split is on RELEVANT/NOT_RELEVANT (ie in_context vs out_of_context)
and only then conditional on that
whether it has a bearing, ie:
has positive correlation with your defined objective or heuristic aka colloquially " is true"
OR
has negative correlation with your defined objective ..
If you just throw in stuff thats not actually context-relevant, you get alot of smudge in your model signal..
So really the decomposition of what's at play here is as follows (borrowing concepts from Fuzzy Logic theories):
1) Availability: is an observable you've defined actually computable for the feature in question
2) Relevance: how relevant (in_context) this feature is to your objective function/heuristic (can be 0-1 dial)
3) what is its value, how do you normalise it or make it tractable to your model signal.
All (1) (2) and (3) need good answers.
Otherwise you end up with something that may work, but it will leave alot on the table.
Great analysis on $CRWV’s actual cost of capital.
Doesn’t seem sensible to be overly l-t bullish the stock due to the size of debt and impact on equity it may have as time goes on, but there’s a difference between that and “co is going busto, look at the CDS” take :)
Details matter yet again
Should be pretty evident CoreWeave's cost of capital is declining over any timeframe greater than a few months.. DDTLs (the ABS, project-finance style facilities backed by specific clusters and customer contracts) comprise the majority of CoreWeave's debt.
Pricing on the DDTLs has tightened a lot over the years. From S+9.62% for DDTL 1.0 issued in 2023.. to as low as S+2.25% for DDTL 4.0 done in March 2026. Note how the older, higher-priced DDTLs 1.0 and 2.0 are declining as the facilities are fully self-amortizing.. And new, lower-priced DDTLs are drawn to fund the most recent capex.
Will be interesting to see where the next DDTL prices. Not going to argue it will be tighter than the S+2.25% (DDTL 4.0) and S+4.50% (5.0) prints from earlier this year. But probably worth noting that DDTL 5.5 was done at S+5.50% (not shown below as it was completed post quarter-end) in July 2026 when AI financing markets were similarly volatile.. And DDTL 5.5 has a 5 year term yet funds capex associated with non-investment grade customers on 3 year contracts, making it a much more aggressive structure than anything prior. In any event, where DDTL 6.0 lands seems much more important than where its (illiquid) unsecured senior notes are trading.
$CRWV
PL retweeted
Lol. So Google's existing models all suck. But because Google employees allegedly claim that not their next model.. But the one after that.. (which will be released in what, 2028?).. "feels exactly" like Anthropic's current model.. They're going to "catch the frontier"?
Your best friend will be a quantifiable edge.
Motivational quotes and “mindset” and quotes of the day on discipline will start to / continue to feel like a fickle mistress.
Hindsight analysing samples of N=1 outcomes and retrofitting overly specific criteria without testing whether they hold for bigger distributions of outcomes will continue to be the main pitfall.
When all the layers of the superficial are shredded, what’s left is:
the cumulative value of { your trading universe, your selectivity criteria, your signals, your position management strategies for executing those signals, your research faculty in polishing your edges }
