Sabitlenmiş Tweet
TeraWulf WULF Investment Thesis (PDF in link below, no paywall)
rittenhouseresearch.substack.com/p/terawulf-inv…
$WULF
English
Rittenhouse Research
2.4K posts
Rittenhouse Research
@RHouseResearch
AI infrastructure, data centers, neoclouds, software, fintech.
Miami, FL Katılım Mayıs 2025
641 Takip Edilen14K Takipçiler
This argument falls flat on its face when you consider that Anthropic is already, and currently operating at positive free cash flow (despite their intent to invest as much as possible).
CapexAndChill@CapexAndChill
In Ackman's latest interview he divides the AI market, stating that it is not a bubble at the foundational layer, but highly risky at the software/model layer. He sees "near infinite demand" for compute. He believes infrastructure providers such as Microsoft, Amazon, Meta, and even SpaceX will earn massive returns on their data center investments. He is highly skeptical of pure model layer businesses like OpenAI and Anthropic. He notes they face massive capital burn rates and an existential threat from open-source models such as Chinese model and SpaceX that are driving the cost of AI access effectively to zero. A business model reliant on charging for proprietary models will collapse if open-source alternatives are essentially free. $AMZN $META $MSFT $SPCX
English
@OnodaCapital $SPCE is down so bad since de-SPAC you can’t even make out the 200% pump it had this year before the SpaceX IPO.
English
Man on CNBC with current SPAC says he regrets going on CNBC with old SPAC
Wall Street Rollup@WallStRollup
Chamath Palihapitiya says he regrets coming on CNBC and talking about SPACs in 2020 & 2021
English
@rennyzucker You really think anyone on here is reading a security’s legal agreement, let alone an ADR conversion doc?
English
Idk I think if you ever find yourself reading ADR conversion docs you need to put it down and back away very slowly because something unintended has happened and all hell is about to break loose
You’re trading paper and you do not have a claim
Some paper is better than others
English
@muddywatersre @chamath The 2026 version doesn’t need to say it because he’s already rich from the 2020-21 version saying it, lol.
English
Nah @chamath you’ll always be SPACJesus to me.
Wall Street Rollup@WallStRollup
Chamath Palihapitiya says he regrets coming on CNBC and talking about SPACs in 2020 & 2021
English
OpenAI is worth a lot less today than it was before $META released muse spark 1.1
If Meta is successful, wouldn't be surprised if OpenAI crumbles and never makes it to IPO...
English
@aleabitoreddit Nebius’s customer base is more concentrated than CoreWeave’s, lol
English
$NBIS signs $1B+ compute agreement with Reflection AI, for GB300 access through 2029.
Reflection also signed a multi-billion dollar agreement with $SPCX earlier.
Interesting to say the least, seeing Nebius drop -5% off the news today.
Also... counterparty to get this done kinda reminds me of OpenAI, where they might not have the funds to actually execute on these LTAs yet compared to $META or $MSFT.
But generally positive long term developments, customer diversification was one of the core strengths of Nebius.
English
@HedgeyeComm I mean you start off by saying “if OpenAI crumbles” then play semantics “I said ChatGPT not OpenAI!”.
And no, you work at Hedgeye, so you’re not doing anything right.
English
@RHouseResearch lol if that’s all you got I’m doing something right
English
@HedgeyeComm Smarter than the guy whose boss screamed about "Quad 4" for years during a raging bull-market, for sure.
English
@RHouseResearch I said ChatGPT, not OAI ARR.
You seem super smart though…
I’m sure you understand.
English
@HedgeyeComm OAI ARR went from $20B to ~$40B+ in the first 5-6 months of the year.. And is likely accelerating as we speak..
Gemini is definitely the real issue for OpenAI though. Certainly not uh, Anthropic.
English
@RHouseResearch Pretty clear that Gemini put a lid on ChatGPTs growth trajectory...
English
@anissagardizy8 Lol, "racing to offload stakes"...
If a financial sponsor running a sale process, for a portfolio company they've held for 5+ years, in order to return capital to their LPs constitutes "racing to offload their stake", then sure..
English
SCOOP: Data center firms are working to sell majority equity stakes worth tens of billions of dollars this summer.
- The largest deal could see DataBank sell for as much as $25B
- The big question is whether there's enough capital for all of these transactions
“There are not that many people who can write those checks,” Ravi Purohit of Paul Weiss told me.
full story in @WSJ: wsj.com/finance/invest…
English
life insurance and cruiselines should be biggest ai beneficiaries?
biggest tam/use case of AI ROI is improving drug discovery -> lower mortalities
life insurers collect premiums longer + fewer payouts. decades of proprietary data.
elderly spend rest of time on cruises
English
The X team should create a feature where anyone talking about personal finance on here also has their portfolio performance numbers flashing next to their name.
English
Yeah, which equates to around a 3 year unlevered FCF payback period, which is what my model outputs (based on a 2.25 year revenue payback period assumption).
In their S-1, they disclosed expected 2.5 year cash payback periods for 2024 based on contracts in place at that time. My payback period is longer given the underlying contract periods on new deals have extended (4 year avg in 2024 to 5 year avg in 2025).
English
@OnodaCapital @RHouseResearch to get a 20% IRR with 0 TV over 5 years doesn’t that imply a cash flow yield of something like 33% vs the initial investment?
English
Pretty remarkable that an investment bank employs someone as their "Head of Technology Research" when said person does not understand the definition of a take-or-pay contract, lol.
Could CoreWeave and Nebius's backlog growth slow if Meta no longer enters into new cloud contracts with them? Sure (though I consider this outcome unlikely).
Can Meta cancel their existing contracts (e.g. those included in backlog) with CoreWeave and Nebius on a whim? No..
$META $CRWV $NBIS
English
Yup (added cumulative column.. noting $1.7B of undiscounted FCFe / $6B TCV = 28% conversion). I'd note that I ran this analysis at 75% LTC, 8% interest, which was before they gave additional color on their latest DDTL (allows them to fund IG-contracts at 90%+ LTC and 6% interest).. So would expect the conversion of TCV into FCFe to require less equity capital and convert at a higher rate (all else equal) going forward.
They also put out a deck in March showing ~26% of TCV converting into FCFe (10% return of equity capital contribution, 16% residual), in line with my modeling..
s205.q4cdn.com/133937190/file…
English
@RHouseResearch @evrgn11112231 Am I right that your model implies $6B of TCV creates ~$1B of undiscounted FCFe? ~$1.6B of Levered FCF - ~$600M Capex. Round nums.
In other words, undiscounted equity value is ~1/6th of TCV?
English
Nope. Bloomberg Surveillance. Here is the literal quote:
"It's very bad for Nebius and CoreWeave because Meta has gone from a buyer of compute, more than half of Nebius's backlog is Meta. More than a third of CoreWeave's backlog is Meta. To a seller of compute.
So that means even if Meta follows through on these agreements, which, they may or may not, they could try to negotiate out of them, or they just don't renew them. Either way that backlog is worth a lot less than it was a couple of days ago assuming Meta follows through."
Again, he is directly suggesting the existing take-or-pay contracts are at risk. Which, of course, they are not.
This guy is a complete jackass. There are plenty of valid bear cases against Nebius and CoreWeave, but suggesting that executed contracts included within backlog can be cancelled for shits and giggles is absurd.
youtube.com/live/X4JvJTPtr…
YouTube
English
If it's the Yahoo video from July 1, he said the contracts with NBIS and CRWV are secure, but if META is long more compute than it needs, then the incremental demand META produces for CRWV and NBIS is not strong. He posits that will hurt NBIS and CRWV if investors price in slower growth.
finance.yahoo.com/video/why-meta…
English
Indeed. Wrote this a few months back (modeling out a hypothetical 100MW cluster). I'd imagine the unlevered IRRs on their hyperscaler contracts (e.g. Meta and MSFT) are a bit lower because those deals are levered close to 100% LTC.
IMO the key takeaway is that there is a decent return on the 5-year contract, but the real upside / bull-case occurs if the GPUs are re-contracted out after 5 years. The debt used to finance them is fully paid off by year 5, so principal and interest, the largest cash outflows during the contract period, go to zero thereafter.
x.com/RHouseResearch…
Rittenhouse Research@RHouseResearch
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/rittenhous…
English
@RHouseResearch 20% irr if you mark the assets to zero in year 5?
English
@RandolphDuke7 He was suggesting Meta might tell NBIS they aren't paying either the $12B take-or-pay piece or the $15B backstop (which I agree will likely be sold elsewhere given spot rates > long-term contracted rates)
English
@RHouseResearch There is a $12B take-or-pay piece and a $15B backstop. I would think if NBIS could obtain higher per-GPU prices than provided for in the Meta backstop, NBIS would sell that capacity elsewhere. No idea what the analyst said, though.
English
@RandolphDuke7 I went back and watched the live interview where he suggested Meta could renegotiate away the pre-existing, executed contracts.. Lol.
English
@RHouseResearch To be charitable, perhaps they were putting a multiple on the backlog and a lower one today. I wouldn't do it that way, but perhaps that is what they were doing.
English
@evrgn11112231 5 years. They make ~20%+ unlevered IRR on the contract so any renewal is upside.
English
@RHouseResearch What’s the contract length and what happens if Meta doesn’t renew at expiry?
English