AESI

Power Business Strategy & Operational Updates

• The power business is organized into two divisions: Oil Field Power (power generation for oil and gas operators) and Long-Term Behind-the-Meter Power (permanent primary power for large-scale users, primarily AI data centers). The company signed its first 120-megawatt behind-the-meter power purchase agreement (PPA) with an investment-grade technology infrastructure subsidiary this quarter.
• The Socorro, Texas project has a $190 million total capital cost, is expected to generate $55 million annualized adjusted free cash flow once operational, with a cash-on-cash payback of less than 3.5 years. A 26-megawatt temporary facility is already operational to support the customer's construction phase; full commissioning is expected to wrap by the end of Q1 2027, with revenue recognition starting in Q2 2027.
• Demand for long-term behind-the-meter private power has grown rapidly due to exploding AI compute demand and increasing difficulty securing timely grid access. The company now expects its remaining uncommitted Caterpillar generator capacity to be contracted via 2-4 large projects, down from the prior expectation of 8-10 smaller projects, and customers increasingly prefer 15-20 year contract tenors, up from the initial 10-year target.
• Oil Field Power delivered strong contracting momentum this quarter, with total 2026 year-end deployed capacity still guided to 180-200 megawatts, the majority under long-term agreements. The business is prioritizing longer-tenor contracts over near-term deployments, and growing interest in hybrid microgrid systems with Atlas battery technology.
• Total capacity scheduled for delivery includes 120 additional megawatts by the end of 2026 and 350 megawatts through 2027, aligning with current customer demand.

Sand & Logistics Market & Commercial Strategy Update

• Macro oil conditions have improved, with the Permian rig count up 17% since the start of the Iran conflict, but completion activity (the key driver of sand demand) has remained flat due to larger pad construction timelines and lingering Permian gas takeaway constraints.
• Management believes the West Texas sand market is much tighter than implied by public nameplate capacity estimates, as widespread underinvestment in maintenance over the past three years has reduced actual productive capacity. Truck availability is also increasingly tight, with acute driver shortages, rising diesel costs, and rising over-the-road freight rates pushing up hauling costs.
• After more than a year of gaining market share via aggressive pricing, the company is shifting to a deliberate strategy of holding firm on pricing in certain tenders, accepting lower near-term volumes to force market discovery of actual productive capacity, drive pricing recovery, and highlight the value of Atlas' reliable execution. The company already took over two mid-completion well-site jobs in June 2026 after competitors could not secure drivers at their bid rates.
• Autonomous trucking development with partner Kodiak has delivered significant productivity gains, with a target of public road operations by mid-2027 pending regulatory and operational milestones.

Core Competitive Advantage Context

• Management highlighted Atlas' track record of delivering large-scale innovative energy infrastructure (including the 42-mile Dune Express conveyor, the first autonomous frac sand delivery in the Permian, and local sand production in West Texas) that overcame industry skepticism, positioning the company to execute on the large behind-the-meter power opportunity.

• Capital Expenditure: Full second-half 2026 growth capex remains consistent with prior guidance; the $110 million second-half capex allocated to the Socorro project was already included in prior guidance, with no increase to the total capital budget. Total 2026 second-half maintenance capex for the legacy logistics/sand business is expected to be $5-7.5 million per quarter, down from prior levels as major legacy projects are completed.
• Q3 2026 Financials: Total sand sales volume is expected to be 5.3-6 million tons, with a wide range driven by planned customer completion breaks and the new pricing discipline strategy. Q3 adjusted EBITDA is guided to $30-45 million, with all variance coming from the sand and logistics segment. Q3 SG&A (excluding litigation and contracting costs) is expected to be $22-24 million, and per ton operating costs are expected to be flat to down sequentially. Logistics margins are expected to remain solidly in double digits.
• Q4 2026: The company expects meaningful sequential improvement in revenue and EBITDA, matching or exceeding Q2 2026 results, based on already allocated volumes and customer completion schedules.
• Power Deployment: The company maintained its guidance of 180-200 megawatts of total Oil Field Power deployed by the end of 2026.

1. Sand & Logistics: Total Q2 2026 revenue from the combined business was approximately $291.8 million (total company revenue is $293.2 million, with the remaining $1.4 million from Power). Sand sales volume hit 5.6 million tons, flat sequentially, with an average sales price of $17.70 per ton. Logistics posted 14% Q2 margins, with a quarterly record of 6 million shipments and a 70% quarter-over-quarter increase in autonomous deliveries to 4,600. Dune Express also set a quarterly volume record. Q2 cost of sales (excluding DD&A) for the segment was $220 million, and per ton sand plant operating costs fell sequentially to $12.39, including royalties. This segment contributes ~99.5% of total company revenue.
2. Power: The power segment is split into Oil Field Power and Long-Term Behind-the-Meter Power. Q2 power equipment operating costs were $1.4 million, and the segment reported strong sequential growth driven by improved oil field fleet utilization and the startup of temporary operations at the Socorro, Texas behind-the-meter facility. This segment contributes ~0.5% of total company revenue.

• Sand and logistics pricing and volume depend on upstream oil and gas completion activity, which has remained flat despite growing rig counts, and is constrained by Permian gas takeaway capacity and operator capital allocation discipline.
• Actual industry sand productive capacity may remain mispriced if the company's new pricing strategy fails to force market discovery, delaying expected pricing recovery.
• Widespread driver shortages and rising diesel/freight costs increase logistics costs even for Atlas, though the company is partially insulated via the Dune Express and autonomous trucks.
• Power contract negotiations have multiple moving parallel processes (including customer data center lease negotiations) that can delay or prevent final deal closing, even with strong customer urgency.
• Large-scale behind-the-meter power projects require long negotiation timelines and depend on customer market growth for AI compute capacity, which carries inherent demand uncertainty.
• Geopolitical volatility continues to impact commodity oil prices, which impacts upstream operator activity levels and resulting sand demand.
• Autonomous trucking expansion to public roads depends on achieving regulatory milestones which may be delayed.

Q: Jim Rolison (Raymond James) asked for additional detail on the size, timing, and pace of data center power deals, specifically noting the updated expectation that remaining capacity will be taken by only 2-4 projects rather than the prior 8-10 forecast. / A: John Turner confirmed that prospective customers are showing intense urgency to sign contracts to de-risk their power timelines, with negotiations moving on month-long rather than year-long timelines. He added that deal sizes and contract tenors are both larger than what the company was seeing six months ago, with deals running in parallel to complex customer data center lease negotiations, and the company will only announce deals once contracts are fully signed.

Q: Steven Gingaro (Stifel) asked about the range of capital expenditures per megawatt for power projects, inflation pressures, and how the company plans to fund its power buildout. / A: Management explained that total capex per megawatt ranges from $1.5 million to $2.5 million, driven mostly by project scope (required reliability metrics and load profile) rather than cost inflation. Following the April convertible issuance, the company currently has $168 million in cash and $125 million in undrawn ABL capacity, enough to cover all near-term obligations. Future project funding is expected to come primarily from off-balance sheet project debt, which will only be added once firm contracts with high-quality counterparties are signed.

Q: Doug Becker (Capital One) asked how the company balances capital allocation between shorter-return oil field power and longer-duration data center power projects. / A: Tim Ondrak explained the company will continue opportunistically deploying into oil field power, prioritizing longer-tenor contracts in high-density areas to capture efficiencies. Oil field power has much shorter scaling timelines than data center power, so the business remains opportunity-driven, with the company deploying capital there when attractive long-tenor opportunities arise.

Q: Chuck DeVore (private investor) asked if recent Texas public policy changes (a pause on grid-connected data center construction and delays to new Permian transmission lines) increase demand for Atlas' behind-the-meter power solutions. / A: Management confirmed these policies are a material tailwind for Atlas. Prospective customers have already been shifting to long-term behind-the-meter solutions for more than a year due to long grid backlogs and capital requirements, and these new policies further accelerate that trend. Customers now prefer 15-20 year contracts because they no longer expect grid power to be a viable long-term solution, and behind-the-meter power also delivers far higher reliability than the public grid.

Q: Scott Gruber (Citigroup) asked if the company's new pricing discipline strategy will lead to higher realized average sand pricing heading into 2027. / A: Blake McCarthy confirmed that 100%, the company's strategy is explicitly designed to deliver higher realized pricing. A large portion of the 2027 sales book is up for bid in the upcoming RFP season, and the strategy will shine a light on the industry's true limited productive capacity, strengthening the company's negotiating position for higher prices.