TeraWulf WULF Investment Thesis (PDF in link below, no paywall) rittenhouseresearch.substack… $WULF

It’s called soccer now 🇺🇸

If you haven’t watched Landon Donovan’s goal vs Algeria at least 10 times this week, are you even a real American?

Jane Street traders when the market closes today

Bill Gurley’s Super Bowl tomorrow

Adam Selipsky, former CEO of AWS from 2021-2024 and now CEO of Helix Digital Infrastructure, pushing back against the premise that hyperscalers can self-build the entirety of their required data center capacity and don't need third-party developers. Selipsky alludes to the idea that having deep connections in places like Texas is a key value-add for developers. Helix Digital Infrastructure was launched today with backing from KKR, NVIDIA, and the Kuwait Investment Authority. "The clients are the hyperscalers.. there's a misnomer that they've got all these capabilities, (and) they've been doing this for a long time and so these projects just come together and the power is there and the connectivity is there and the data centers there, and.. they just do it all in-house themselves.. I'm here to tell you that is not true. And I know that from the inside of building a $100B year business (AWS), which relies fundamentally on this infrastructure. There's all the old jokes about two folks in a pickup truck in Texas. And they've got a great deal. But you know, they golf with the Governor's brother-in-law. So don't worry about the permits.. And these are real stories. This actually happens. They (hyperscalers) absolutely have a lot of capabilities to do a lot of these things in-house, (but) they absolutely need reliable partners who can bring them scaled projects." $AMZN $MSFT $GOOG $GLXY $CORZ $WULF $CIFR

Microsoft has been engaged in “price wars” for decades (Dynamics is cheaper than Salesforce or Oracle, Fabric is cheaper than Databricks and Snowflake, Teams is cheaper than Zoom, etc). And it’s still doing $300B of revenue at 45% operating margins.

Surely the Knicks aren’t about to blow a 2-0 lead with two straight losses at MSG..

The cost of capital for CoreWeave $CRWV, and data center developments tied to its creditworthiness, continues to decline.

The Broadcom, Apollo, and Blackstone "AI XPV Platform" announcement references a 20GW target by 2028, starting with 1GW across Anthropic's "Fluidstack-based" sites in mid-2026. Fluidstack's 1GW of sites (where leases are backstopped by Google) are being developed by: - TeraWulf $WULF Lake Mariner (378MW) and Abernathy JV (168MW) - Hut8 $HUT River Bend (245MW) - Cipher Digital $CIFR Barber Lake (207MW) Probably fair to say we can expect more Fluidstack leases with $WULF $HUT $CIFR if Broadcom, Apollo, and Blackstone get anywhere near their 20GW target..

If I had to guess the Crusoe data center "pause" in Wyoming is much more likely to be attributable to growing community backlash than a lack of demand for AI infrastructure. The WSJ had a story yesterday on how the Laramie County Board of Commissioners pulled the approval for a proposed 5,600 bed "man-camp" at the last-minute, despite the planning commission having previously voted unanimously to advance the project. Meta and Microsoft already have sites in the area, which suggests Crusoe's "Project Jade" campus will be leased to either Amazon or Google. Both are raising money hand over fist to invest in AI infrastructure (e.g. Google's $85B of equity last week).. The idea that either decided yesterday that they actually don't want or need more data center capacity is absurd. It seems far more reasonable that the pause allows Crusoe the tenant to address the backlash and work with local officials on a path forward. Jamming the project through without the 5,600 bed "man-camp" would drive local rents through the roof, pricing existing residents out of their homes, as Laramie County is only a population of ~100K. I think the hyperscalers are wise enough to understand that this is not palatable, even if they have the permits to build. $GOOG $AMZN

Crusoe has to be the most undervalued private company right now with their last round done at just $10B. Even the rumored $30B - $40B “pre-IPO” round seems way too low?

More on the skilled worker shortage.. Yesterday, Meta $META launched "America's Workforce Academy", the largest ever private-sector investment in skilled trades with a job guarantee. A data center developer's access to top general contractors, and thus ability to procure labor, is only going to be of increasing importance.. $GLXY $CORZ $WULF about.fb.com/news/2026/06/am…

CoreWeave $CRWV credit spreads have plummeted this year.

Google $GOOG selling $35B of their chips to their second largest competitor, only to then lease chips from SpaceX $SPCX for ~$1B/month is a real head scratcher.. If Google is as capacity constrained as their $85B equity raise and deal with SpaceX would have you believe, why did they commit to give even a single TPU to Anthropic..?

Cipher Digital's $CIFR 8-K filed today provided a peek under the hood with respect to the lease at their Stingray campus in West Texas.. In short, the economics and structure are extremely compelling: - Tenant: AWS - 70 MW (IT load), 100MW gross - Triple net structure (e.g. all operating expenses passed through to tenant, resulting in 100% NOI margins) - $10.5MM of expected CapEx/IT MW, with AWS agreeing to cover any construction cost overruns above $10.5MM/IT MW - 15% unlevered yield on cost ($112MM of first full year NOI divided by $735MM of total CapEx, or $10.5MM IT MW x 70MW)

A quick recap on Galaxy Digital's $GLXY Helios data center campus: - 1.63GW (gross) of fully ERCOT approved, front-of-the-meter power.. An incremental 1.8GW sits in ERCOT's approval queue, which if approved as expected, would scale the site to 3.4GW... "The largest single campus data center in America:"✅ - 800MW gross, 526MW IT load contracted to CoreWeave for $1B of annual revenue over a 15 year lease, with the first phase of the project on-time, on-budget, and now generating revenue (execution chops:✅) - 830MW gross, ~550MW IT load to be leased out "by the time Fourth of July comes around, or certainly by the summer ends" (tenant diversification: 🔜) - 1,500 construction workers currently on-site, which "will max out to 2,500" (labor procurement: ✅) - Located in Afton, TX, a remote town with a population of only ~200 people.. "Very business friendly.. a part of the country, and a part of Texas, where you can build data centers without a lot of community problems. We're in the middle of nowhere, we got lots of spare electricity out there, so were not driving electricity prices up for the local community... (no) noise pollution. Texas is one of the best places for data centers, (and) we've looked at other cities, and states." (optimal location: ✅)

CoreWeave $CRWV co-founder Brannin McBee @branninmcbee on Odd Lots discussing how the shortage of skilled workers is one of the most acute bottlenecks in bringing GPU clusters online, because without tradesmen such as electricians, powered shells can't be built. "Just think about the electricians for these sites and getting the accreditation on the electrician side to be able to participate in these builds. I mean, there's a five year plus apprenticeship to be able to go through that program. Right. We can't just make new electricians leveraging a supply chain.. that's a trade that you can't really scale efficiently. So that is absolutely the bottleneck for us." podcasts.apple.com/us/podcas…

Ruth Porat to Google's new CFO earlier this week $GOOG

If Morgan Stanley and Goldman are modeling $200B of SpaceX "AI" revenue in 2030, one can only imagine what they're going to show in their Anthropic and OpenAI models 😂