Brett Knoblauch

Kent Draper said pricing is consistent across live conversations and trending up; Dan clarified 3- to 5-year deals.

Kent Draper said PUEs are ambient-driven with incremental improvements; Dan noted spare-power headroom from conservative PUEs.

Kent Draper said Mirantis brings faster deployment, the ability to service software-heavy enterprise customers, strong internal engineering, and a well-built customer-support function; also asked about the 730MW 2027 ramp split (450MW Childress, 280MW Sweetwater), which management confirmed.

Kent pointed to longer tenures, air-cooled interest and prepayments rather than a specific pricing figure.

Kent said GPUs are the vast majority of the $5.8B order, with substantial InfiniBand networking; reusability for future GPU generations remains to be seen.

Daniel Roberts said customers want on-demand compute now, so ordering hardware ahead is warranted while each incremental order is a small share of overall risk; Kent added color on power-per-GPU rising.

Kent agreed it demonstrates internal capability, pointing to building 50 MW of data centers a month as proof supporting customer conversations trending toward IREN retaining ownership.

Kent confirmed the post-DeepSeek uptick factored in, citing a large demand uptick for liquid-cooled capacity across cloud and colocation and a delivery-timeline advantage.

Management dismissed Bitmain delivery concerns, citing six years without material issues and the last 31 exahash miners arriving within about seven days.

Management said it is less inclined to invest in GPUs near term given cost of capital and softer demand ahead of the Blackwell transition, favoring Bitcoin CapEx.

Management acknowledged more perceived demand after the FERC decision but cautioned nothing is real until a deal is signed.

Nathan Frankovitz, on for Brett Knoblauch, asked how mining could be incorporated into future site acquisitions and development; Matt described using modular immersion-cooled mining to monetize interruptible or early-energized power within 90 days, then converting sites to data centers.

Matt Schultz said the portfolio approach does not change lease timing and that assets have independent value, framing it as upside rather than a delay; a direct answer.

Asked by colleague Gareth on behalf of Brent Noble (operator-introduced). Management gave specifics on Sealy (285MW, first tranche 1H27) and Brazoria energization (Q4 27-Q1 28).

Matt declined to disclose the exact price but said it was a mix of equity and cash at the low end of the market range.

Knoblauch asked whether scale lets CleanSpark sell less Bitcoin to cover costs; Vecchiarelli cited operating leverage dropping to the bottom line.

Zach said expansion is possible across all operating regions but expects the majority of growth in Tennessee and Wyoming, citing utility willingness to adopt interruptible-load energy tariffs.

Gary attributed OpEx discipline to a strict budgeting process and low incremental corporate overhead, expecting indirect expenses to stay relatively flat with professional fees the only M&A-driven wildcard.

Zach expects a second wave of private-miner sales as held-back fleets age, but sees little public-to-public consolidation given premiums.

Zach declined to commit to one capital source, citing open capital markets, and quantified roughly $200 million of infrastructure spend to reach beyond 50 exahash.

Les declined to list specific gating items, describing the scale of the multi-year, 756 MW development and the design, legal, and commercial work required from both sides.

Management confirmed full-site interest persists but noted 756 MW is unprecedented scale for a single lease. Core and shell procurement is well-advanced with the same AMD general contractor secured, and Tier 3 long-lead items (generators, chillers) already being procured.

Management confirmed Phase 1 builds on unused land preserving mining cash flows, with Phase 2 eventually supplanting mining buildings for the final 168 MW, paced by customer demand.

Management described hiring as following a pendulum from engineering to sales to operations, with key development hires in place and recent SVP of AI and Hyperscale Sales addition shifting focus toward leasing.

Management described phased approach to Corsicana development, noting large-scale capacity is attractive for growth-oriented tenants seeking expansion pipeline beyond initial lease.

Management noted 4% network share is not a mandate, expects miners face power constraints similar to data centers, and outlined ~26% YoY growth for 2025 followed by ~10% for 2026.

Management confirmed openness to leasing the full 1 GW if the economics are accretive to shareholders.

Management clarified power supply agreements do not require demand response participation; all curtailment is voluntary and specific to Bitcoin mining.

Fred cited competition among multiple prospective tenants across 90% of MARA's capacity and surging demand for compute as model providers cap usage and raise prices.

Fred said the first 200-megawatt facility would take roughly 18 to 24 months, with behind-the-meter capacity coming online soonest and grid interconnection following.

Management declined to give a specific 2025 hash rate target, citing dependence on mining profitability and a focus on capital efficiency over scale.

Management defended using convertible debt to buy an appreciating asset (Bitcoin) rather than depreciating miners, framing it as opportunistic capital allocation.

Management said Bitcoin is now highly correlated to equities and is in a near-term risk-off phase, but maintained the long-term adoption trend is intact.

Management outlined a plan for roughly 50% international capacity by 2028, partnering with energy companies abroad and acquiring energy assets domestically.

Management answered directly, favoring small edge inference sites over large training campuses and full co-location leasing.

Direct, confirmed CapEx is in line with past projections and quantified cadence at $30-50 million a month for the first building.

Partial, pointed to more interest and higher counterparty comfort from world-class financing partners, but gave no concrete diligence-timeline acceleration.

Wes said the preferred is only a sliver of the capital stack and the blended cost with project-level debt leaves returns well north of cost of capital.

Wes said the ROFR applies to asset-level preferred financing, not the parent's public equity, which carries no such restriction.

Vlad framed Social as closing the loop between idea generation and trade execution. Verified trades across equities, options, crypto, prediction markets, and futures create unique content. Jason positioned it as flywheel to attract more customers rather than direct monetization.

Jason confirmed $6B in stakeable assets with rapid customer adoption. Broader opportunity framed as continuing to win crypto market share.

Jason indicated current pricing levels are satisfactory with experimentation continuing. Suggests near-term take rate stability after several quarters of increases.