←
Compute1,642 posts
Compute avatar

Compute

@JoshInvestsAI

NYC 🗽| Concentrated Asymmetry Investor | $IREN, $NUAI, $CIFR, $IBRX, $SOFI, $ASST, $AMPX, $BMNR

◷ Joined May 2020▧ 90 Photos and videos
Compute avatar
Compute@JoshInvestsAI·Apr 19

$NUAI Thesis I know you’ve probably been seeing $NUAI across your timeline recently, and was wondering what the deal with the company is. When I first saw this micro-cap stock on my timeline, I was extremely hesitant. But, after looking deeper I started to draw parallels to companies like $IREN and $APLD before the markets inevitably priced in what they could be worth. For those of you who have been in $IREN you understand that a 10x means speculating on an asset before its true value becomes obvious to the market, and seeing a vision for a company that plans to transition to something different. I believe $NUAI is the next 10x opportunity at $4.60, and here’s why. Introduction $NUAI’s flagship site is called TCDC, or Texas Critical Data Centers, that is 1.4GW Gross or ~1GW Critical IT load, in Ector County near Odessa, TX. I believe the market is completely missing the scale of what is happening at the TCDC site. If you haven’t been paying attention closely it’s easy for the recent announcements to fly under the radar. Since the beginning of this year $NUAI has put out announcements every couple weeks forging partnerships with energy generation partners, hiring significant industry leaders, and working meaningfully to prep their site for their hyperscaler client. Following the Breadcrumbs At the beginning of the year $NUAI closed the acquisition of the remaining 50% stake of TCDC from SharonAI, their original JV partner on the project. This was to have total control over the land to build new partnerships, and likely at the hyperscaler clients request to bring a new JV partner to the table. In late February $NUAI announced a 450 MW Behind-the-Meter Generation Plan at TCDC with Thunderhead Energy and TURBINE-X Energy to generate power for the project. They have a SPV/venture where they have partnered with a private equity firm to fund the capital required to source, install, and run 450MW of gas turbines. They will generate an ROI as they sell power to NUAIs TCDC tenant. This equipment is normally extremely difficult to source, and $NUAI was able to grab these 2-3yr long lead time items in an extremely supply constrained environment, showing the connections of the upper management. The notable thing about this partnership is that they don’t have to put any capital upfront for this equipment, allowing them to start generating revenues while loaning the equipment without huge dilution (~70m). Then, in mid-March $NUAI hired Ted Warner, ex head of the Energy, Power, and Digital Infrastructure at Northland Capital Markers, where he successfully structured and managed more than $7 billion in financing solutions specifically for large-scale data center developments. This was the signal that caught my eye and made me significantly up my position in the company. Something notable is his PSU’s, or Performance Stock Units, which reward him significantly for “Entering into a binding commercial agreement with a hyperscaler for a minimum of 200 megawatts”. Judging by his history I put him as an A+ hire, and you can see more about his notable achievements here. x.com/litigious_dulce/status… Now, to get into the most recent developments that completely change this company from just a speculative random micro-cap to a more credible multibagger infrastructure play. On April 1st, $NUAI secured an LOI with Stream Data Centers. Originally, $NUAI was slated to co-develop the TCDC site with other partners (first Sharon AI, then Primary Digital Infrastructure). However after back and forth for months the unnamed hyperscaler tenant likely mandated that their own preferred execution partner handle the physical construction and operation. Stream is one of the top data center This project with Stream will be done in a GP/LP fashion, where instead of issuing billions in new stock to pay for construction, the deal utilizes a heavily levered GP/LP (General Partner / Limited Partner) joint venture structure, operating at roughly 80% Loan-to-Cost (LTC). The roles exist as such: The Originator ($NUAI): $NUAI acts as the local sponsor. They bought the TCDC 438-acre site, secured the initial power footprint, navigated local Texas politics, and laid the development groundwork. Originally I had thought that this agreement would mean passing up the GP role fully to steam and forfeiting their GP revenue streams as the originator. After contacting IR they said “we expect the final structure to reflect the contributions from each (NUAI as a project originator + local relationships, Stream as the developer)” The Operator (Stream): Stream Data Centers steps in as the development manager and operator. They bring the engineering blueprints, the construction expertise, and the direct, trusted relationships with the hyperscaler. The hyperscaler obviously has a preference for this developer and has likely worked with them before. The Institutional Capital (The LP): An unnamed institutional investor (99%+ confidence being Apollo, given their majority ownership of Stream) acts as the Limited Partner, writing the massive equity checks and leading the project financing. Apollo Global Management is consistently ranked among the top, most influential, and pre-eminent firms in private equity globally. It is commonly considered part of the "Big 4" of the PE industry and a mark of validation for this project that shouldn’t be looked past. What this structure does is it protects $NUAI shareholders from the dilution that is so costly to shareholders in a company like $IREN, which long term (2-3 years) has potential to grow into a 100B+ market cap giant, but will need to dilute massive equity to get the cashflow flywheel going (as shown by 6B ATM). A 1.4 GW campus costs upwards of $12 Billion to build, and $NUAI cannot fund that on its own balance sheet. I expect this LOI to be facilitated into a binding agreement in the next 2-6 weeks. The second piece of news that is the most convincing is the $290M credit facility from Macquarie. Just to put it out, in case it isn’t obvious already, Apollo and Macquarie don’t blindly gamble on some random micro-cap without doing extensive underwriting and due diligence on the parties involved. They deemed $NUAI worthy enough at a $250 million market cap to receive a $290 million multi-tranche facility that shifts the risk profile of the entire TCDC project. The structure goes as follows: a $20M committed Term Loan A-1 to kick off development, followed by $30M and $40M tranches, with a massive $200M Delayed Draw Term Loan waiting for execution milestones. That’s already impressive in itself, and likely came from Ted Warner’s existing relationship with Macquarie, but an even stronger validation is the equity kicker. Macquarie took a direct $5 million equity stake at a 20% premium to market share price, taking their entry at exactly $5.00 per share. They also have a tranche of warrants that will have an exercise price of $5.00. The warrants will be issued across the first $50 million drawn on the Facility. When one of the largest infrastructure lenders in the world is taking an equity stake at a 20% premium to the market’s price, it should tell you everything you need to know about the asymmetry of this setup. You have to ask yourself, would Macquarie offer a loan of this size to a random microcap without doing extensive due diligence and underwriting of an advanced hyperscaler LOI or term sheet? Luckily, we got the answer to that hidden in the SEC filing of the Macquarie term loan agreement without a formal press release from $NUAI. $NUAI currently has an LOI in place with a hyperscaler tenant as of March 24th, 2026. I believe the reason it was not put in the form of a press release was at the hyperscalers request, and you can read more about it below from @kamikazzzi1981 x.com/kamikazzzi1981/status/… Now that all of the pieces of the puzzle are starting to come together that a hyperscaler deal is imminent, what should the company actually be worth? Stock Price Projections Firstly, I'd like to get out of the way that if you believe that $NUAI will secure a hyperscaler deal in the first place the announcement alone will probably put the stock around 2x higher than the current prices, or around $9-$10/share. This is by giving Stream, one of the most reputable data center builders hand selected by a hyperscaler, around a 60% chance of execution from the math below. If you want to, from that point, you can decide whether you want to sell, or if you believe they can execute the numbers start to get pretty wild. From Phase 1 alone, with super conservative assumptions (more likely to be ~$1.50M EBITDA/MW) Phase 1: 200 MW x $1.35M EBITDA/MW = $270M project EBITDA. At 50% ownership, that is $135M to NUAI. At a 14x-16x EBITDA multiple, that suggests $1.9B-$2.2B of value, or $14-$16/share on 135.5M fully diluted shares. Northland’s analysis used a 19x EBITDA multiple in their analysis, 14x-16x is extremely conservative. At a 19x multiple you get a share price ~19$/share If they can fully execute on Phase 1, the full TCDC 1.4GW campus can be modeled out as done by @ThePrudentWhale x.com/ThePrudentWhale/status… • $34.51 Price Target in a 100% success case. • $25.91 Probability-Weighted PT (applying a 25% execution risk haircut). And here is super conservative bit of the model: this entirely excludes Behind-The-Meter power generation economics for phase 2+, GP stream revenues, 7 GW New Mexico development pipeline, and the reinvestment of cash flows into the futures phases of the projects for more equity.

Post attachmentPost attachment

Updated the $NUAI model. Added an EBITDA multiple based valuation alongside the NOI / cap rate valuation. Both screenshots below reflect base case assumptions. View v19 here: docs.google.com/spreadsheets…

Post attachmentPost attachment
mon avatar
⟳ Compute retweeted
mon@moninvestor·2h

$NUAI is up 70% over the past 5 days while $IREN is down 20%. $RIOT is up 5% and now has a bigger market cap than $IREN. $NBIS, $WULF and $CIFR are all green today. Something is not adding up. What is the market missing? NUAI is now 2% away from becoming my biggest position.

Post attachmentPost attachment
New Era Energy & Digital, Inc. (Nasdaq: NUAI) avatar
⟳ Compute retweeted
New Era Energy & Digital, Inc. (Nasdaq: NUAI)@NUAI_IR·Nov 16

$NUAI President & CFO Ted Warner took the stage at @cantorfitzgerld Convergence to discuss the evolving AI infrastructure market and how New Era’s competitive power position provides a foundation for building a scalable platform. Securing power is only part of the equation. How soon that power can be delivered, how projects are financed, and the development models used to bring capacity online are becoming increasingly important considerations for investors. With 207 MW of firm, contracted power expected to become available in Q3 2027, New Era sees a compelling opportunity to pair its near-term power position with a modular, phased development strategy designed to accelerate deployment, reduce upfront capital requirements, and provide greater flexibility in how capacity is delivered. This is all supported by a world class team that’s experienced in building hyperscaler data center infrastructure.

Post attachment
₿itcoin ₿utcher 🥩 🐑 🐷 avatar
⟳ Compute retweeted
₿itcoin ₿utcher 🥩 🐑 🐷@bitcoinbutcher1·Nov 15

$NUAI Phase 1 Signed is Worth $10 Today Inspired by my friend @SmallCapSnipa publishing the Northland sell side analysis, I found it appropriate to keep it simple and stick to the same format. Warning: this analysis is for Phase 1 only, and does not include Phase 2 or Phase 3 at TCDC that will at a minimum power 1.2 GW or 6x Phase 1. Now let's talk through some assumptions before you read the output I produced attached. 1) Rent - $riot executed for $130/kw/Month with $ant but my base case is $140 due to $NUAI being responsible for development and construction of the modular data centers versus just the powered pads and a later LOI date in September as power became harder to find in 2027 2) PUE - Northland is far too conservative as the $riot deal is 1.1. The initial LOI asked for 756 MW IT but landed on 910 MW IT because $ant opted for a simpler modular build and will be able to adjust power down during peak temperatures but will pay for the max capacity. My base case is 1.35 for now. 3) NOI / Ownership - 99% NNN and 95% for Vistra 5% share. 4) Multiple - Northland uses at 10 cap or 10x multiple but I find that cash return too high and believe 11x is more appropriate as a 9.1% cap rate is almost 4% higher than the 5.35% average of the 10 year and 20 year treasury. That's a very wide credit spread and the multiple should likely be 11.5x or 12.0x. 5) Build Cost - Same $6m assumption with 80% debt assumed. 6) DS0 - Used fully diluted shares of 160m from earnings to be conservative 7) Discount - Northland uses 11% for 2 years for a 2028 delivery but that's far too conservative as I believe delivery will be Q4 2027 given the modular build. That said, I upped the discount to 20% for one year for time value of money and execution risk. Ted Warner just telegraphed Phase 1 delivery and most likely Phase 2 at the Cantor panel to mitigate regulatory risk upon approval of the company's air permit by starting construction as soon as possible. I've never been more bullish on a setup because management is keeping it super simple which will result in accelerated institutional adoption as the deal structure can be digested in 5 minutes. See $riot performance following their deal versus peers. Load up. Final call. NFA DYOR

Post attachmentPost attachment
WOLF avatar
⟳ Compute retweeted
WOLF@WOLF_Financial·Nov 8

ANTHROPIC IS BUYING ITS OWN POWER PLANTS Anthropic agreed to pay $3.4 billion for 1.7 gigawatts of gas-fired generators for its AI data centers, per the FT. It's the largest known single purchase of power-generation equipment by an AI lab. • Anthropic took over Caterpillar $CAT generator orders from San Antonio-based OnePWR Solutions • It didn't buy the company itself • Anthropic said this week it plans to spend more than $895 billion in the years ahead, mostly on computing power and infrastructure AI labs are now skipping the middlemen to lock down scarce power equipment.

Post attachment
Lovedrop avatar
⟳ Compute retweeted
Lovedrop@LovedropV·Nov 7

$NUAI LOTS of noise this week around Texas governor-elect Gina Hinojosa and what she means for data centers. @level39 laid out why her direct power over TCEQ, PUCT and ERCOT is more limited than the headlines suggest. Worth the read on his page. So let's flip the question. Not what she can block, but how $NUAI stacks up against what she's actually asking for. We scored TCDC against her six "Texans First" priorities using filings, decks and management's own words. 2 strong alignments. 1 aligned by design. 3 still open. Close enough that TCDC could become the model for how data centers get built in Texas. Thread below with @fnitefiend and @optionology.

Post attachment
Mike Alfred avatar
⟳ Compute retweeted
Mike Alfred@mikealfred·Nov 6

I bought the stock for the first time this morning.

Mike Novogratz avatar
⟳ Compute retweeted
Mike Novogratz@novogratz·Nov 5

The blue wave has weighed on companies like ours. I get it. But let’s not confuse politics with long-term fundamentals. Finance is moving onchain and demand for AI is exploding. These are generational shifts that will play out across administrations and market cycles. We’ve spent years building Galaxy to win in both. That includes Texas, where we delivered Helios Phase I on schedule and 800 MW is already leased under long-term contract. We’ve worked with legislators and local communities there for years, and we’ll keep doing that with the new administration. We share the same goals: reliable power, protected water and strong communities. I see a massive disconnect between where $GLXY trades and what I believe this company is worth. That’s why I’m buying more stock, starting today. I’m betting on us, big time.

matthew sigel, recovering CFA avatar
⟳ Compute retweeted
matthew sigel, recovering CFA@matthew_sigel·Nov 4

My humble suggestion to the management teams in the sector with the weakest balance sheets: hire an advisor and formally discuss selling your company to someone who can better afford to build out your infrastructure. Be a hero. Set the valuation floor for everyone. Your legacy is secure. Find your next chapter.

Bitcoin miners have been crushed, and November is picking up where October left off. Rarely do we see this much valuation support across an entire sector of our digital assets universe. The market is pricing in a lot of bad news in our opinion. Our latest research explains why we think it's overdone. H/T Griffin MacMaster ⬇️

Post attachment
JustaKeelGuy avatar
⟳ Compute retweeted
JustaKeelGuy@justakeelguy·Nov 4

Deal economics in AI infrastructure are improving for the landlord. Two recent Frontier AI Lab comps show the shift: WULF (Anthropic, 401 MW): 20-year term, $19B contracted ($2.37M revenue per MW/year). More traditional campus build — landlord delivers a higher level of finished infrastructure. CapEx $10–12M/MW. 85% NOI margin. 18% yield on cost. RIOT (Corsicana, 910 MW): 15-year triple-net powered land lease, $28B contracted ($2.0M revenue per MW/year). Landlord delivers land, power infrastructure, and powered pads. Tenant brings the GPUs (same as WULF) and pays nearly all operating costs. CapEx only $6M/MW. 99% NOI margin. 33% unlevered yield on cost. Same class of tenant. Similar revenue per MW. The difference is how much of the build the landlord has to fund. The newer structure is more capital-efficient and higher-margin. This is the trend: as power becomes the scarce asset, landlords who control it can negotiate lighter CapEx and higher-margin structures instead of carrying full facility builds. $KEEL already has the power secured and is still in active negotiations. No pressure to sign the first deal that shows up — they have the cash to wait for the best structure and terms. While others race to lock something in, $KEEL can let the comps keep improving and take the higher-quality version of the same deal. The first signature will re-rate the entire company, and they get to pick the one that maximizes it.

Rittenhouse Research avatar
⟳ Compute retweeted
Rittenhouse Research@RHouseResearch·Nov 4

The Riot $RIOT - Anthropic "powered land lease" deal at Riot's Corsicana site might have just set the new high watermark for AI/HPC lease economics. Unlike the build-to-suit structures seen in the vast majority of prior transactions, RIOT is simply renting Anthropic access to its powered land, and isn't required to build the data center itself: "The lease is structured as a triple-net powered land lease, under which Riot delivers the land, the power infrastructure, and the powered pads, while continuing to own and maintain the high- and medium-voltage infrastructure that will serve them." - per RIOT 8-K RIOT's expected capital requirement is just $6MM/MW, reflecting $2MM/MW for site work and electrical infrastructure, and $4MM/MW in reimbursements to Anthropic (e.g. help fund Anthropic's construction capex). So in exchange for simply procuring and assembling the site's electrical infrastructure at an expected cost of $1.8 billion.. plus funding $3.6 billion of tenant reimbursements.. so an "all-in" cost of ~$5.5 billion.. RIOT receives $28 billion in rent over a 15-year period, or roughly $1.8 billion per year on average.. RIOT's expected year 1 (e.g. 2027-2028) NOI per MW of $1.6MM implies a 27% unlevered yield on cost. RIOT's expected average annual NOI per MW of $2MM/MW implies a 33% unlevered yield on cost. Previous leases have seen landlords collect $2MM/MW in rent.. but those landlords had to incur $12MM/MW+ in capex costs. RIOT is getting $2 million/MW for just $6MM/MW in capex. Insane.

Post attachmentPost attachmentPost attachmentPost attachment
P Equity Research 📰 avatar
⟳ Compute retweeted
P Equity Research 📰@pequityresearch·Nov 3

CEO of Modal, Erik Bernhardsson, GPU lead times are 3-4 months currently, and demand has grown 30-40% a month for a year. $NVDA $AMD

Modal CEO reveals current GPU lead times are 3-4 months, and says demand has grown 30 to 40% a month for a year, so he buys GPUs at 2 to 3x current capacity Erik Bernhardsson (CEO): “The GPU lead times right now is about three to four months. So we basically do a straight line or exponential line extrapolation, and we look at an aggregate across all our customers. We have a very diversified set of customers." "You can actually look at the demand in aggregate pretty well and predict based on monthly growth rates, which have been about 30, 40% month-to-month for the last 12 months, and you just project that 3 to 4 months into the future, you end up with like 2 to 3x. You go out and buy that capacity and make sure to ramp up.”

0:29
matthew sigel, recovering CFA avatar
⟳ Compute retweeted
matthew sigel, recovering CFA@matthew_sigel·Nov 3

Today’s data center deal makes powered shells look cheap--> DCRU sold stakes in 3 North American data centers to its sponsor DLR at 5.2-6.6% cap rates, or ~$21-26M per MW >PSPs are building for $10-12M per MW, so stabilized assets are clearing at roughly 2x build cost >The assets sold were weak: Toronto has 3 years of lease left, LA under 1, Virginia was vacant until a new lease starts Dec 1, and they were all built pre-2000. In contrast, PSP leases are on new builds and run 10-20 years with escalators >We estimate PSPs trade at ~9x 2028 EV/EBITDA vs ~18.5x for DLR/EQIX >Risk: the sponsor was the buyer and other bids came in lower

Post attachment
Nick Khami avatar
⟳ Compute retweeted
Nick Khami@skeptrune·Nov 3

a single GPU now makes more than the federal minimum wage of $7.25/hr fucked up to say, but isn't that a sign of compute being underpriced? one B300 produces much more value than a minimum wage human ☹️

you should scramble to get compute because I have a feeling it's about to get ugly

Post attachment
Tyler Page avatar
⟳ Compute retweeted
Tyler Page@rftylerpage·Nov 2

I would love to. The short-term market is disconnected from how well positioned we are. @bitcoinbutcher1 lets connect to find a time

Phil Rosen avatar
⟳ Compute retweeted
Phil Rosen@philrosenn·Oct 27

We've entered the next stage of the AI trade and it's all about debt financing. I sat down with @GoldmanSachs CIO Alexandra Wilson-Elizondo to unpack AI debt and bubble fears, why the best hedge against AI is more AI, and where she's investing beyond the US, from Korea and Japan to European financials. TIMESTAMPS: 0:00 - 4th stage of the AI trade 2:09 - AI debt indigestion 3:51 - Is AI a bubble? 5:47 - GPU obsolescence risk 9:23 - Credit warning signs 10:34 - Best hedge against AI is more AI 11:32 - Pharma and utilities 13:42 - S&P 500 concentration 16:30 - European financials 18:15 - Emerging markets and Korea 20:54 - Japan 24:37 - Longevity trade 29:31 - Back to fundamentals 31:10 - Follow Alexandra

31:41