Reginald Smith

Management identified depth of technical and product-level engagement as the key signal of counterparty seriousness, noting they are engaging multiple counterparties simultaneously to move toward high-quality outcomes quickly.

Management reframed location premium as primarily driving tenant quality rather than direct pricing uplift, noting proximity to Tier 1 markets widens conversations to the highest-credit counterparties who unlock the best financing.

Management acknowledged evaluating the neocloud model but prioritized build-to-suit colocation as the best first step, noting the design could serve an internal neocloud business if decided later.

Management emphasized both location and ready-for-service timing as dual advantages, noting they can command strong economics and attract high-quality tenants who unlock favorable project financing terms.

Management acknowledged high demand for powered land leases but argued the value extracted would be significantly below what the build-to-suit approach can capture with their assets, balance sheet, and team.

Management described Corsicana as near a Tier-1 market (Dallas) and Rockdale near an emerging market (Austin/San Antonio), both commanding better economics than remote sites.

Management outlined milestones: internal team buildout, basis of design completion, tenant lease, and financing as key markers to watch.

Management noted domestic switchgear production mitigates direct tariff risk, but high-voltage transformers from China face exposure. Situation too fluid for definitive impact assessment.

Management confirmed an LOI would likely precede any formal lease, providing an intermediate signal for investors.

Management attributed elevated SG&A to approximately $22 million in one-off Q4 litigation, M&A, and advisory costs, guiding to $30-33M quarterly cash run-rate.

Management expressed optimism but acknowledged unpredictable timeline, noting multiple steps could occur in various orders without a hard deadline.

Fred confirmed MARA prioritizes near-term convertible power (MPLX is longer-term) and is still pursuing sites both domestically and internationally, including with French energy majors.

Fred said Starwood's existing hyperscaler relationships and in-house EPC compress the courtship period and accelerate lease timelines versus MARA going alone.

Fred outlined permitting first, building behind-the-meter gas generation, with ERCOT interconnect later.

Fred pointed to energy-company partnerships, rising inference mix, and profit per megawatt hour as the KPI.

Fred argued the market undervalues mining infrastructure and the energy/AI story beyond the Bitcoin stack.

Fred framed the MPLX deal as securing reliable, low-cost, self-generated power and answered directly.

Fred explained the compute-efficiency progression and the 3-year ASIC depreciation versus longer hyperscaler schedules.

Fred distinguished MARA's air-cooled, ASIC-based inference model from GPU colocation risk.

Fred described Exaion as an asset-light, secure private-cloud operator of EDF's data centers.

Max said category choice is driven by consumer pull observed via the card; the goal is universal acceptance like a 21st-century Amex.

Max said Affirm's robust model-building process lets it quickly incorporate new data and enter new verticals confidently.

Max said AI underpins underwriting and fraud, and generative AI drives productivity in legal, compliance, and accounting without adding headcount.

Max said the underwriting engine and ML team transfer to the UK, though local data is still very early (about eight days).

Harley emphasized the GMV-based model over a la carte pricing, noting commerce components create on-ramps for enterprise brands to start with Checkout or Shop Pay and expand into full platform over time.