Darren Aftahi

Wes covered initial power (100 MW in 2028) and positive local receptivity but did not say whether a customer was in mind.

Wes put Base Electron's initial capacity at 2029-2030, ramping after; other regional transmission/generation projects fill 2026-2028 to keep expanding the ND campuses.

Wes said he expects a Delta Forge 1 lease in the near term to hit the RFS date, and confirmed they still have 3 sites in exclusivity with a hyperscaler after pausing the South Dakota campus over a missing tax exemption.

Wes said the evergreen Macquarie facility funds groundwork ahead of leases; construction on a new campus starts by end of January on high confidence of a new hyperscaler lease.

Wes said pricing has been stable to slightly better with extraordinarily robust demand, and contract terms (make-whole, transferability) have grown more favorable.

Management said Building 2 is designed for about half of Building 1's build time and confirmed standard late-delivery penalties apply.

Direct, said hyperscaler demand is at least steady to higher versus three months ago, though it rotates between players and hasn't shifted in recent weeks.

Direct, confirmed the multi-tenant technical issues were resolved in early March and should be behind them for most of the May quarter.

Wes declined to detail the deal timing but said he strongly expects a single customer to take the entire 400 MW campus.

Partial: declined to give yield targets (called it premature) but reaffirmed the economics fit the long-stated model and that buildings 2-3 won't see major CapEx before a lease is signed.

Direct: initial 100MW lease plus a reservation on the remaining 300MW, with the single tenant expected to take the whole campus across two leases.

Management favors AI cloud for higher dollars per megawatt but stays open to compelling colo deals.

Kent confirmed Sweetwater is likely in batch 0, securing the full 2 GW of power.

Dan said the Microsoft talks evolved over time and IREN chose to focus on cloud; hinted hyperscalers may prefer colo but declined specifics on Microsoft's preference.

Daniel Roberts explained Horizon can be monetized via cloud (own/operate) or colocation, that reconfiguring for lower rack densities widens GPU flexibility, and displacing mining is an accepted cost of business.

Kent answered directly, attributing below-market build cost to seven years of power-dense design experience and reuse of existing shells/electrical infrastructure rather than concrete new-builds.

Dan framed the decision as strategic rather than pure ROI, deliberately not projecting numbers, arguing the platform value from a signed liquid-cooled design unlocks Sweetwater conversations.

Dan said disclosing the late-stage project gives investors context on the Sweetwater 1 single-site strategy and the West Texas multi-gigawatt thematic, despite its unsigned status.

Management said building out Bitcoin mining remains the preferred use of capital, weighing each AI/colocation pathway as an opportunity cost while keeping optionality.

Management emphasized flexibility across customer types and described the basis of design as essential for serious counterparties to advance discussions substantially.

Management prioritized Corsicana for initial data center buildout but confirmed Rockdale remains in the pipeline for future transition, with Bitcoin mining monetizing power in the interim.

Management estimated 100-200 MW on existing pad with higher density possible at higher CapEx, with additional phases requiring civil and electrical work.

Management ruled out retrofitting the existing building but noted potential for new data center development on surrounding land using Rhodium's power capacity.

Jason Chung declined to give a specific acreage figure but confirmed procurement in the hundreds of acres to maximize optionality.

Jason Chung confirmed willingness to convert Rockdale fully if economics support it, noting 1.7 GW would represent about 5% of hyperscaler demand.