Superfly

WHY THE MIDTERMS COULD ACTUALLY BE BULLISH FOR $CRWV AND $IREN The market is making an important mistake with CoreWeave and IREN following midterms because a policy that slows how quickly new data centers can be built is fundamentally different from a deterioration in the demand for AI computing and those two outcomes can have very different implications for companies that already control valuable infrastructure. Gina Hinojosa has called for a moratorium until lawmakers establish stronger community protections, which I take seriously given the potential implications for projects awaiting approval but what investors seem to be overlooking is that Greg Abbott had already ordered ERCOT to audit pending grid connections and halted environmental permitting before the election while pushing developers to cover their own electricity and infrastructure costs. The political direction was already moving toward greater regulation regardless of who won, which means the real incremental risk is how long restrictions remain in place and what conditions developers must satisfy rather than the sudden disappearance of an otherwise unrestricted market. I would also be careful assuming every permitted facility is automatically protected because electrical interconnection, environmental approvals and construction authorizations are different things but I think the market is moving too quickly from legitimate concerns about development timelines toward an assumption that the economic opportunity itself has somehow deteriorated. What makes this particularly interesting to me is that electricity is increasingly becoming one of the most valuable inputs in the AI economy because hyperscalers can purchase additional GPUs much faster than developers can secure hundreds of megawatts of reliable power which means the value of infrastructure that can support computing today should rise as the time and cost required to bring competing capacity online increases. The distinction matters because an operating megawatt with available cooling, networking and computing equipment has an entirely different economic value from a megawatt that exists only in an announced development pipeline especially if customers are willing to pay more for immediately available capacity. IREN is particularly interesting in this environment because Childress is already operational and its first Horizon deployment has been accepted by Microsoft while Sweetwater 1 has an energized high voltage substation and an initial 300MW of gross capacity under construction creating an infrastructure advantage that becomes harder for competitors to replicate as permitting becomes more restrictive. However, I would not treat the entire 2GW Sweetwater development as secured operating capacity because its inclusion in ERCOT's Batch Zero process remains conditional, which is why I think the opportunity is really about the increasing value of IREN's existing infrastructure alongside the execution risk associated with developing its remaining capacity. CoreWeave presents a somewhat different opportunity because its advantage is less about owning the underlying electricity infrastructure and more about its ability to turn available computing capacity into contracted customer revenue which becomes increasingly important when new competitors face longer development timelines. Galaxy has already delivered 133MW of critical computing capacity at its Texas Helios campus to CoreWeave and has the next phase financed and under construction but the distinction is that Galaxy owns the physical infrastructure while CoreWeave accesses it through long-term leases, meaning some of the scarcity value naturally accrues to the property owner before CoreWeave captures anything through its cloud business. Thats where I think investors need to be more sophisticated because CoreWeave's roughly $104 billion reported revenue backlog provides meaningful visibility into customer demand but does not represent cash already earned while its 9.625% unsecured debt financing demonstrates how expensive development delays can become when significant capital has already been committed. IREN faces its own version of this challenge given the substantial spending required to deploy GPUs and its approximately $13.8 billion of disclosed contractual commitments at the end of June so I do not view either company as an automatic winner from tighter regulation but I do think IREN offers greater exposure to the appreciation of scarce physical infrastructure while CoreWeave offers more direct exposure to monetizing contracted AI computing demand. Ultimately, this is why I view the current selling as an opportunity to buy the fear because I believe the market is increasingly pricing these companies as though restrictions on future development automatically undermine the economics of infrastructure that is already operating, when a more constrained supply environment could actually strengthen the competitive position of companies capable of delivering computing capacity today. The financial question is whether the additional value created by scarce power and higher utilization can exceed the present value of delayed expansion, increased construction costs and financing obligations, particularly since fixed customer contracts may prevent existing operators from capturing higher market prices immediately. For IREN, I want to see continued customer acceptance of its Horizon deployments, progress at Sweetwater, improving cash generation and evidence that its development commitments can be funded without excessive dilution, while for CoreWeave I will be watching whether its substantial backlog continues converting into recognized revenue and whether new capacity generates enough cash to comfortably support lease payments, interest expense and ongoing GPU investment. I would reconsider my conviction if permitting restrictions begin materially delaying existing projects, financing costs overwhelm the returns on new deployments or declining GPU rental prices prevent either company from benefiting economically from tighter infrastructure supply but my broader view remains that as electricity becomes an increasingly scarce resource in the AI economy, the companies with access to usable power and the ability to convert that access into profitable computing capacity could become considerably more valuable than investors currently appreciate.