CoreWeave's risk is buildout speed: each new build needs capital before the last pays for itself. Its analysis shows $35.6B of loans lasting about nine months beyond customer contracts; $14.6B is due in 2031-32 as final contract cash arrives. IREN's Microsoft project loan matches its contract, but $6.3B of convertible bonds mature in 2031-33, mostly after contracted cash ends; low shares mean cash repayment. Slower AI demand growth could let both earn back projects before funding more. x.com 

AI infrastructure providers' heavy capital needs and aging equipment may leave little cash for shareholders despite revenue growth. Depreciation is among their most underappreciated costs: shorter equipment lives raise it, while replacement and financing erode returns. Leverage and dilution magnify economic risks. Some may prosper, but until pricing power, depreciation, financing and long-term returns on capital are clearer, their investment appeal remains hard to share. x.com  x.com 

Keel Infrastructure has a realistic path to 10x by leasing the 2.2GW of grid-connected power it already controls. At $150-200 per kW per month, the three leases the CEO targets this year, Panther Creek (350MW), Sharon (110MW) and Moses Lake (18MW), imply $850M-$1.1B of annual revenue. The full pipeline would exceed $4B, and at 6-7x sales for operators with hyperscaler contracts, a conservative multiple implies about $30B. Each signed lease re-rates the rest. x.com 

Keel Infrastructure looks expensive at $3.57, about 20% above the top of a $2.35-$2.98 fair-value range. The company, which runs digital and energy infrastructure for HPC and AI workloads, trades at 13.4x EV/sales with forward EPS of -$0.24. These are valuation levels from reported earnings and revenue estimates, not fundamental work, and the stock would be outright expensive above $4.80. x.com 

Keel Infrastructure has no signed hyperscaler lease, so it must fund its pre-lease buildout with expensive convertibles and ATM equity, and every dilutive dollar raised before a lease comes out of the bulls' 10x target. Cost of capital is the whole game, and the cheapest capital, debt against an investment-grade lease, only unlocks once a lease is signed. @CKCapitalxx's 10x math assumes $150-200 per kW per month, but DGXX's $190 rate only works for turnkey builds, not leasing a powered shell. … Show more

Keel Infrastructure trades at a large discount to peers mainly because it lacks a hyperscaler deal, and a single contract could close that gap. It has grid connections in place, a 2.2 GW pipeline and a market cap near $3.8B. Sites are Panther Creek, PA at 350 MW with a path to 500+ MW long term; Sharon, PA at 110 MW; and an 18 MW Moses Lake, WA pilot. CEO Ben Gagnon committed to sign three leases by year-end, and the stock should join the Russell 3000 in June. Leopold added to his position. … Show more

CoreWeave's financing model still works if it can keep funding capex with contract-backed loans at roughly 6-9%. Its existing debt costs about 8.4% blended, not 15%; the 15% figure is the market yield on its 2032 bond, which it locked at 9.625%. With roughly $20B of 2026 capex left and more needed in 2027, the next deal is critical: another $20B at 15% would cost $3B a year, more than current annual interest, and the model would not work. Spreads near 1,000 bp have made new borrowing pricier. … Show more

Keel Infrastructure's lease commitments, which would justify the stock, have slipped from a firm 2026 target to open-ended talks while its biggest site runs late. On the Q2 call CEO Gagnon would not reaffirm three signed leases this year when pressed by Northland and ATB analysts, citing only active negotiations at all three sites. Final permits for the 350 MW Panther Creek campus are taking months longer, and Moses Lake slipped a couple of months. Its secured 480 MW faces no policy risk. x.com 

Keel Infrastructure has multibagger potential because a single signed hyperscaler lease would reprice the whole company. Formerly Bitfarms, it is converting entirely to AI data centers, with a 2.2GW power pipeline that includes up to 1.3GW at its Pennsylvania site. Management is targeting three hyperscaler leases by year end. x.com 

AI infrastructure can keep expanding despite heavy issuance because most Big Five builders have strong operating cash flow and benign leverage. SpaceX is the cash-flow exception. Big Five capex could rise from over $1.3tn in 2027 to over $2.5tn in 2030, requiring multiple trillions in collective issuance and likely material dilution. Financing may cut CoreWeave and Nebius's spending, but even a $100bn cut would be only 3-4% of Big Five 2030 capex. Sustained spending requires visible returns. … Show more