$SEI KEY READ-THROUGHS FROM SOLARIS ENERGY INFRASTRUCTURE Q1 2026 EARNINGS CALL
Solaris Energy Infrastructure’s Q1 2026 call is a high-signal read-through for the AI/data-center power ecosystem, U.S. natural gas demand, electrical equipment, distributed generation, regulated utilities, and infrastructure construction. The key market implication is that behind-the-meter natural gas power is moving from a niche workaround to a bankable, multi-GW procurement model for investment-grade technology customers. Solaris now has >2 GW of long-term contracted generation capacity with 3 technology customers, including a new >600 MW contract with balance of plant, a prior >500 MW February contract, and the ~900 MW Stateline JV. Contract terms are 10 to 15 years, and management described a path to 3.1 GW of secured generation capacity and >$1 billion of annual adjusted EBITDA once fully delivered and operating. The most important cross-sector message is that the data-center power bottleneck is not being solved only by utility interconnections, grid PPAs, renewables, or nuclear scarcity; it is increasingly being solved by dedicated, site-level power infrastructure that includes gas supply, generation, emissions controls, electrical distribution, storage, controls, and operations. This is positive for turbine OEMs, electrical balance-of-plant suppliers, gas midstream, SCR/emissions-control vendors, and specialty construction contractors. It is negative or at least dilutive to the investment narratives of utilities, renewable-only developers, grid-dependent IPPs, and subscale powered-land developers that lack integrated power execution capability.
POWER GENERATION OEMS AND TURBINE SLOT HOLDERS: DATA-CENTER DEMAND IS CONVERTING INTO SCARCE, LONG-CYCLE GAS TURBINE ORDERS (READ-THROUGH 1)
Affected companies: Caterpillar (CAT: US), GE Vernova (GEV: US), Baker Hughes (BKR: US), Siemens Energy (ENR: Germany), Rolls-Royce (RR: UK), Mitsubishi Heavy Industries (7011: Japan), Cummins (CMI: US), Generac (GNRC: US).
Directional impact and magnitude: Positive / High for OEMs and suppliers with exposure to mobile, modular, aeroderivative, industrial gas turbine, and large natural gas generator platforms. Positive / Medium for diversified manufacturers where data-center distributed power is still a small percentage of total revenue.
Specific call support: Solaris stated that it added 2 significant long-term contracts with 2 investment-grade global technology companies for >1 GW of contracted power generation capacity and associated balance-of-plant equipment. Management also closed 2 strategic transactions that expanded secured generation capacity by >40% to 3.1 GW. Kyle Ramachandran said: “Demand for our solutions continues to outpace our committed and on-order capacity.” He also highlighted the acquisition of Genco Power Solutions, adding ~400 MW of incremental capacity between 2026 and 2028, and the purchase of 30 turbine delivery slots, adding ~500 MW between early 2027 and 2029. Later, management stated that “the turbines and quite frankly, the SCRs continue to be the long-lead item in the scope.”
Transmission mechanism: Hyperscaler willingness to sign 10 to 15-year behind-the-meter power contracts is converting AI power demand into financed turbine procurement. This benefits turbine OEMs through order conversion, delivery-slot scarcity, backlog durability, pricing power, and aftermarket service demand. The call also suggests that delivery slots themselves are becoming monetizable assets: Solaris purchased slots because customers need faster deployment timelines than standard OEM lead times allow. That dynamic supports a higher value for near-term manufacturing capacity and installed/available equipment.
Near-term trading catalyst: Additional announcements by Solaris or peers securing turbine slots, OEM backlog commentary, price increases, and evidence that delivery dates can be pulled forward. Any OEM commentary that data-center-related orders are now a discrete demand vertical would be positive for the group.
Longer-duration fundamental shift: Data centers are becoming a structural demand category for distributed gas turbines, not just emergency backup generators. This increases the strategic value of modular gas turbine manufacturing capacity and changes the load profile from episodic standby demand to long-duration, high-utilization generation demand.
ELECTRICAL BALANCE OF PLANT, SWITCHGEAR, TRANSFORMERS, AND POWER MANAGEMENT: THE VALUE POOL IS EXPANDING BEYOND TURBINES (READ-THROUGH 2)
Affected companies: Eaton (ETN: US), Schneider Electric (SU: France), Vertiv (VRT: US), ABB (ABBN: Switzerland), Siemens (SIE: Germany), Powell Industries (POWL: US), Hubbell (HUBB: US), nVent Electric (NVT: US).
Directional impact and magnitude: Positive / High for electrical equipment suppliers levered to medium-voltage distribution, switchgear, transformers, protection systems, e-houses, power conditioning, and data-center power management. Positive / Medium for broader electrical manufacturers due to diversification and mix.
Specific call support: Amanda Brock stated that customers increasingly value a “turnkey and rapidly deployable solution” that includes “not only generation capacity, but the power-related distribution, conditioning, storage and management capabilities as well as the equipment needed to supply fuel and minimize emissions.” She also stated that the newest >600 MW contract includes “last mile gas delivery as well as natural gas fuel generation assets in the associated distribution storage and balance of plant infrastructure.” In Q&A, management confirmed: “The last 2 contracts include balance of plant.” Steve Tompsett reiterated that balance-of-plant scope can drive a 20% to 50% uplift, with an additional $800 million to $1 billion of capex potentially translating into $160 million to $200 million of incremental EBITDA.
Transmission mechanism: Behind-the-meter power generation requires a complete electrical architecture, not just turbines. Every large deployment needs switchgear, transformers, cables, e-houses, power conversion, protection, controls, monitoring, and distribution systems capable of delivering power at the required voltage and reliability standard. Solaris’ commentary that customers are asking for more turnkey scope implies higher electrical content per MW of data-center capacity. This should drive incremental order flow to suppliers of power distribution and power management hardware.
Near-term trading catalyst: Backlog and book-to-bill strength at electrical equipment companies, particularly in data-center, utility, and industrial power segments. Commentary around medium-voltage equipment, transformer lead times, and data-center customer demand should be watched closely.
Longer-duration fundamental shift: Data-center power architecture is becoming more decentralized and site-specific. That favors electrical equipment companies that can support both grid-connected and behind-the-meter architectures. The value pool shifts from “server building only” to a broader external power-infrastructure layer surrounding the data center.
SCR, NOX CONTROL, AND EMISSIONS-CONTROL SUPPLIERS: EMISSIONS HARDWARE IS AN UNDERAPPRECIATED BOTTLENECK (READ-THROUGH 3)
Affected companies: CECO Environmental (CECO: US), Fuel Tech (FTEK: US), Babcock & Wilcox Enterprises (BW: US), Johnson Matthey (JMAT: UK), BASF (BAS: Germany).
Directional impact and magnitude: Positive / Medium to High for emissions-control and catalyst suppliers. Magnitude is high for niche suppliers with meaningful SCR exposure, but medium for diversified chemical and industrial companies.
Specific call support: Solaris repeatedly emphasized that behind-the-meter power must be deployed “rapidly and compliantly.” Kyle Ramachandran stated that “the turbines and quite frankly, the SCRs continue to be the long-lead item in the scope.” Bill Zartler also said Solaris made an investment in an SCR manufacturing business and described it as “very strategic,” adding that SCRs and “the catalysts associated with the SCR” are bottlenecks.
Transmission mechanism: Onsite gas generation for data centers requires emissions compliance, particularly where projects are located near communities concerned about air quality, electricity affordability, and industrial development. SCR systems and catalyst supply become gating items for permitting, deployment, and community acceptance. As data-center customers increasingly demand rapid, behind-the-meter gas power, emissions-control hardware moves from a secondary procurement item to a critical path component.
Near-term trading catalyst: Evidence of longer SCR/catalyst lead times, order growth from distributed power projects, and supplier commentary around data-center or industrial gas-generation demand. Smaller public companies with SCR/NOx-control exposure could see outsized sensitivity if orders materialize.
Longer-duration fundamental shift: The market may begin valuing emissions-control suppliers as beneficiaries of AI power demand rather than as legacy industrial compliance vendors. The more behind-the-meter gas generation is deployed, the more emissions-control technology becomes embedded in the AI infrastructure supply chain.
NATURAL GAS MIDSTREAM AND GAS-LEVERED PRODUCERS: DATA CENTERS ARE BECOMING A NEW BASELOAD GAS DEMAND CLASS (READ-THROUGH 4)
Affected companies: Williams Companies (WMB: US), Kinder Morgan (KMI: US), Energy Transfer (ET: US), ONEOK (OKE: US), Enterprise Products Partners (EPD: US), EQT (EQT: US), Range Resources (RRC: US), Antero Resources (AR: US), Coterra Energy (CTRA: US), CNX Resources (CNX: US), Archrock (AROC: US), USA Compression Partners (USAC: US).
Directional impact and magnitude: Positive / Medium for midstream and compression companies; Positive / Medium for gas producers in basins with deliverability into data-center growth regions. Magnitude becomes High if Solaris-style behind-the-meter gas deployments scale broadly across the hyperscaler market.
Specific call support: Solaris described its strategy as “molecule-to-electron.” The company’s new contract includes “last mile gas delivery” and “natural gas fuel generation assets.” Bill Zartler stated that Solaris can deliver “everything from gas through the delivery of the electron into the building.” He also said the company may build “small pipes into the facility” if needed and if returns justify the capital.
Transmission mechanism: Dedicated behind-the-meter gas generation creates new, high-load-factor gas demand attached directly to data-center sites. Unlike peaking power, AI compute loads are structurally baseload or high-utilization. At illustrative high utilization, a fully deployed 3.1 GW gas turbine fleet could represent several tenths of a Bcf/d of natural gas demand, depending on heat rate and dispatch profile. The more important investment implication is replicability: if multiple hyperscalers adopt this model at multi-GW scale, data centers become a material incremental demand vertical for pipeline laterals, pressure control, compression, and gas supply.
Near-term trading catalyst: Announcements of data-center gas supply agreements, new pipeline laterals, compression projects, and producer commentary tying demand growth to AI/data-center load. Midstream companies with assets near gas-rich, power-constrained regions should be most sensitive.
Longer-duration fundamental shift: Power availability may increasingly determine data-center site selection, and gas deliverability becomes a competitive advantage. This favors regions and midstream systems that can support large, rapidly deployable gas-fired loads without requiring years-long grid interconnection upgrades.
REGULATED ELECTRIC UTILITIES: BEHIND-THE-METER POWER IS A THREAT TO THE DATA-CENTER LOAD-GROWTH NARRATIVE (READ-THROUGH 5)
Affected companies: Dominion Energy (D: US), American Electric Power (AEP: US), Duke Energy (DUK: US), Southern Company (SO: US), Exelon (EXC: US), FirstEnergy (FE: US), Entergy (ETR: US), CenterPoint Energy (CNP: US).
Directional impact and magnitude: Negative / Medium for utilities whose equity narratives and rate-base growth expectations rely on large data-center load additions. Impact is not uniformly negative because some utilities may still benefit from transmission, backup service, interconnection support, and long-term grid integration, but the call is directionally adverse to a pure utility-served load-growth thesis.
Specific call support: Bill Zartler stated: “Grid interconnection delays have continued to expand, which given the market’s focus on speed to compute, has accelerated adoption of long-term behind-the-meter power solutions.” He also said: “Electricity affordability for residential grid customers remains at the forefront of every politician and community leaders’ minds, which reinforces the need to bring your own power solutions like ours and in some cases is even essential in too many communities.” Solaris also emphasized the ability to operate “fully behind-the-meter in island mode if needed.”
Transmission mechanism: If hyperscalers can self-supply power through long-term behind-the-meter contracts, regulated utilities lose some portion of prospective load growth, generation investment, and grid-upgrade rate-base opportunities. Regulators and politicians may also pressure large loads to bring dedicated supply to avoid socializing costs across residential customers. This creates risk to utility load forecasts, interconnection assumptions, and capex plans predicated on data-center demand.
Near-term trading catalyst: Utility regulatory proceedings that require large-load tariffs, dedicated generation, self-supply structures, or upfront contribution requirements. Additional hyperscaler behind-the-meter announcements would likely pressure utilities most exposed to data-center load-growth expectations.
Longer-duration fundamental shift: A new large-load framework is emerging in which hyperscalers may be expected to bring power with them, pay dedicated infrastructure costs, or accept limited grid reliance. This could structurally reduce the utility monopoly over serving new large industrial loads.
MERCHANT POWER AND IPPS: BEHIND-THE-METER GAS CREATES COMPETITION FOR GRID-CONNECTED DATA-CENTER PPAS (READ-THROUGH 6)
Affected companies: Constellation Energy (CEG: US), Vistra (VST: US), Talen Energy (TLN: US), NRG Energy (NRG: US), AES (AES: US).
Directional impact and magnitude: Negative / Medium for the scarcity-premium component of the data-center PPA thesis; Neutral to Positive / Medium for broader power demand fundamentals. The read-through is nuanced: AI load growth remains positive for power markets, but behind-the-meter alternatives reduce the exclusivity of grid-connected generation assets.
Specific call support: Solaris stated that behind-the-meter power will play “a significant role in the long-term powering of data centers and other large industrial power loads.” Amanda Brock said Solaris can deploy “at a speed and reliability level that the grid and traditional procurement channels will have difficulty matching.” Management also said technology customers are increasingly asking Solaris to deliver broader turnkey scope rather than only generation.
Transmission mechanism: Merchant generators and IPPs have benefited from the market view that hyperscalers must secure scarce grid-connected power through premium long-term PPAs, nuclear contracts, or colocated structures. Solaris’ success suggests that hyperscalers have another route: dedicated onsite or behind-the-meter gas power with contracted infrastructure partners. This could moderate the negotiating leverage of grid-connected generators and reduce the scarcity option value embedded in some IPP valuations.
Near-term trading catalyst: Additional long-duration behind-the-meter contracts with hyperscalers, particularly if customers choose onsite gas instead of grid PPAs. Any sign that data-center customers are using behind-the-meter options as leverage in grid-connected PPA negotiations would be incrementally negative for IPP pricing power.
Longer-duration fundamental shift: The data-center power market may bifurcate into 2 procurement channels: grid-connected clean or firm power for longer-duration decarbonization objectives, and behind-the-meter gas for immediate speed-to-compute. That reduces the probability that any 1 category of generator captures all AI-driven power economics.
RENEWABLE DEVELOPERS AND STORAGE: FIRM GAS IS WINNING THE SPEED-TO-POWER USE CASE (READ-THROUGH 7)
Affected companies: NextEra Energy (NEE: US), AES (AES: US), Brookfield Renewable Partners (BEP: US), Fluence Energy (FLNC: US), First Solar (FSLR: US).
Directional impact and magnitude: Negative / Medium for renewable-only data-center power narratives; Mixed / Medium for storage companies because storage can be part of balance-of-plant architecture but is not presented as the primary energization solution.
Specific call support: The call repeatedly emphasized “speed to compute,” expanding grid interconnection delays, behind-the-meter operation, natural gas-fueled turbine capacity, and the ability to operate “in island mode if needed.” Solaris’ strategic transactions added ~900 MW of natural gas-fueled turbine capacity, not renewable generation. Amanda Brock also framed storage as part of broader power infrastructure, but generation was centered on natural gas turbines.
Transmission mechanism: Hyperscalers still have decarbonization goals, but the call indicates that immediate data-center energization is being driven by firm, dispatchable, deployable gas power. Renewable projects and storage may remain important for long-term emissions strategy and grid integration, but they are disadvantaged when the primary customer objective is rapid deployment and reliable island-mode operation. This can defer or reduce the share of AI load captured by renewable-only PPAs in the near term.
Near-term trading catalyst: Hyperscaler power procurement announcements that emphasize natural gas, behind-the-meter generation, or dedicated onsite power instead of renewable PPAs. Renewable developers with data-center demand narratives could face multiple compression if investors conclude that interconnection delays push meaningful load to gas.
Longer-duration fundamental shift: Renewables may become a second-stage optimization layer rather than the first-stage solution for AI compute power. The first stage is increasingly firm onsite power; the second stage may involve renewables, storage, grid integration, and emissions mitigation.
AI ACCELERATOR, SERVER, AND NETWORKING SUPPLY CHAIN: POWER DE-BOTTLENECKING SUPPORTS MULTI-YEAR AI HARDWARE DEMAND (READ-THROUGH 8)
Affected companies: NVIDIA (NVDA: US), Advanced Micro Devices (AMD: US), Broadcom (AVGO: US), Marvell Technology (MRVL: US), Arista Networks (ANET: US), Dell Technologies (DELL: US), Super Micro Computer (SMCI: US), Taiwan Semiconductor Manufacturing Company (2330: Taiwan).
Directional impact and magnitude: Positive / Medium. The read-through is not a direct order signal for any one semiconductor or hardware vendor, but it supports the durability of AI infrastructure capex by showing that hyperscalers are actively solving the power bottleneck with long-duration contracts.
Specific call support: Solaris said it is now operating, constructing, designing, and planning multiple large behind-the-meter power projects for 3 distinct large technology companies across several data centers. Kyle Ramachandran stated that Solaris now has >2 GW of power generation under long-term contracts with 3 leading technology companies, with >50% of that contracted in the last 2 months and contract terms extending to 10 to 15 years. Amanda Brock also said Solaris was asked by a large technology customer to participate in a pilot research program related to “mobile distributed compute.”
Transmission mechanism: AI accelerator demand is constrained not only by chip supply but also by available powered data-center capacity. Multi-GW behind-the-meter power contracts imply that hyperscalers are underwriting the infrastructure needed to energize additional compute clusters. That supports future GPU, custom ASIC, networking, server, and storage procurement by reducing the risk that power availability becomes a hard cap on AI hardware deployment.
Near-term trading catalyst: Hyperscaler capex commentary and AI infrastructure order guides that reference improving power availability or new data-center capacity. Hardware vendors may benefit if investors interpret power-procurement progress as support for sustained 2027 and 2028 AI capex.
Longer-duration fundamental shift: Energy procurement is becoming part of the semiconductor demand chain. The pace of AI hardware growth will increasingly be linked to the ability of hyperscalers and infrastructure partners to secure power, not only to silicon capacity or server supply.
HYPERSCALERS AND CLOUD PLATFORMS: OPERATIONAL POSITIVE, FREE-CASH-FLOW NEGATIVE (READ-THROUGH 9)
Affected companies: Microsoft (MSFT: US), Amazon (AMZN: US), Alphabet (GOOGL: US), Meta Platforms (META: US), Oracle (ORCL: US).
Directional impact and magnitude: Operationally Positive / Medium; financially Mixed to Negative / Medium due to higher long-duration infrastructure commitments and energy complexity. No specific Solaris customer identity was disclosed, so the read-through applies to the hyperscaler cohort rather than to any named customer relationship.
Specific call support: Solaris said customers selected it as a “trusted long-term partner” and that contracts are extending to 10 to 15 years. Amanda Brock said customers increasingly want Solaris to do more, including generation, last-mile gas delivery, distribution, storage, and balance-of-plant infrastructure. Bill Zartler said technology companies “have been forced into contracting for power in ways like this,” reflecting the urgency of the power bottleneck.
Transmission mechanism: For hyperscalers, behind-the-meter power contracts can accelerate compute deployment and reduce dependence on utility interconnection queues. However, they also increase the capital and operating complexity of AI infrastructure. Even if assets are owned by Solaris or other third parties, hyperscalers are likely underwriting long-term contracted costs, availability requirements, fuel pass-through structures, and potential emissions/ESG trade-offs. The result is faster capacity but potentially lower free cash flow conversion and more complex infrastructure risk.
Near-term trading catalyst: Hyperscaler earnings commentary around capex, data-center energization, power procurement, and AI infrastructure returns. Investors should distinguish between “capacity solved” and “ROIC solved.” Solaris’ call supports the former more clearly than the latter.
Longer-duration fundamental shift: Cloud platforms are evolving from software/infrastructure-light businesses into vertically integrated industrial power consumers. Power strategy may become a core differentiator in AI cloud competition.
