SEI introduction

[drew1618t]

Solaris Energy Infrastructure (SEI)

Solaris Energy Infrastructure (NYSE: SEI) is an energy infrastructure company headquartered in Houston, Texas. Originally focused on oilfield logistics services, SEI transformed itself through the September 2024 acquisition of Mobile Energy Rentals (MER) into a distributed power generation company serving data centers and AI infrastructure. The company provides mobile, natural gas-powered turbines and equipment to provide behind-the-meter power solutions to hyperscalers and commercial customers.

This company caught my attention because of its positioning in the AI data center power bottleneck space. With 760 MW operational and plans to scale to 2,200 MW by early 2028, SEI is riding the massive demand wave from AI infrastructure. From Microsoft’s CEO Nadella “I am power [constrained], yes, I’m not chip supply constrained”.

Financial Overview (Q3 2025):

Market Cap:	~$3.5B (as of January 2026)
Q3 2025 Revenue:	$167M (record quarter, +12% QoQ)
Q3 2025 Adj. EBITDA:	$68M (tripled YoY)
Q3 2025 Net Income:	$25M
2025 Revenue Est:	~$560M (~80% YoY growth)
Power Capacity:	760 MW operational
Capacity Target:	2,200 MW by early 2028
Pro Forma Earnings Est:	$600M+ (per management)

Who:

William Zartler (Founder, Chairman, Co-CEO) ~4.2M shares (~$200M)

30+ years of energy experience. Prior to founding Solaris in 2014, was a Founder and Managing Partner of Denham Capital Management, a global energy and commodities private equity firm. He led Denham’s global investing activity in midstream and oilfield services. Also serves as Executive Chairman of Aris Water Solutions. Recent insider buying (10,000 shares at $24.83 in Sept 2025) shows confidence.

Amanda Brock (Co-CEO, Director) - Appointed November 2025

Appointed to partner with Zartler in scaling the company for growth. Lead Independent Director at Coterra Energy. Started career as lawyer at Vinson & Elkins, managed global projects in power and water, served as CEO of Water Standard. Brings scaling expertise critical for SEI’s growth phase.

Kyle Ramachandran (President & CFO)

Joined Solaris in 2014. Prior experience at Barra Energia (Brazil exploration company) and First Reserve Corporation (energy-focused PE). Started career in Citigroup’s M&A Group. Also recently bought 2,000 shares showing alignment.

Revenue Ramp:

Quarter	Revenue	Adj. EBITDA
Q4 2024	$96M	$37M
Q1 2025	$126M (+31% QoQ)	$47M
Q2 2025	$149M (+18% QoQ)	$61M
Q3 2025	$167M (+12% QoQ)	$68M
Q4 2025 (guidance)	~$175M est	$65-70M
Q1 2026 (guidance)	~$185M est	$70-75M

Power Solutions now exceeds 60% of total revenue with over 75% of segment-level EBITDA. The legacy Logistics Solutions segment provides steady cash flow, but the growth story is entirely in Power Solutions.

Business Model:

SEI operates two segments:

Power Solutions (~60% of revenue): Provides configurable, all-electric natural gas-powered mobile turbines and ancillary equipment to data centers, energy companies, and industrial customers. This is the high-growth segment riding AI infrastructure demand. Contracts are typically 2-4 years with 80%+ of capacity committed.

Logistics Solutions (~40% of revenue): Legacy oilfield services business designing/manufacturing specialized equipment for oil & gas completions. Steady cash generator but tied to oil & gas activity levels.

Key Differentiator: SEI provides “behind-the-meter” power solutions - meaning they deliver power directly to the customer’s site rather than going through the grid. This bypasses grid delays and permitting bottlenecks that are holding back data center development. AI data centers need power NOW and the grid can’t deliver fast enough.

Stateline Joint Venture:

In Q1 2025, SEI finalized a joint venture with a “major data center client” (rumored to be a hyperscaler) to provide 900 MW of primary power to an AI data center campus. Key terms:

• 900 MW capacity (upsized ~80% from original plan)

• 7-year contract term

• Provides primary power with grid redundancy/backup

• JV structure provides balance sheet flexibility

This single deal represents ~41% of their 2,200 MW target capacity and validates their value proposition to major AI players.

Risks & Concerns:

Class Action Lawsuit: A securities class action was filed covering July 2024 - March 2025, alleging SEI failed to disclose issues with the MER acquisition (including that MER’s co-owner was a convicted felon with fraud allegations). The stock dropped 17% on March 17, 2025 when this came out. This is a real concern - however, the company has continued executing well since then and the stock has recovered significantly. The lawsuit appears to be standard ambulance-chasing but warrants monitoring.

Supply Chain Constraints: CEO Zartler explicitly noted “The supply chain is growing out. We are lucky to get the slots we have with our relationships.” Lead times for turbines and equipment are stretching, which could hamper execution.

Execution Risk: Scaling from 760 MW to 2,200 MW is nearly 3x growth in ~2.5 years. This requires significant capital expenditure and operational execution. Management has proven capable so far.

Negative Free Cash Flow: Due to heavy CapEx for growth, FCF is currently negative. Expected CapEx for remainder of 2025 was ~$295M. Watching this and depreciation on how it affects the companies profitability in the future.

Oil & Gas Exposure: The Logistics Solutions segment is tied to O&G activity. WTI crude has softened to <$60/barrel which may pressure this segment. However, this is the declining portion of the business.

Heavy Asset:

When looking at companies with large asset expenses, I want to verify that the company is actually profitable and not buying growth. Since the revenue is being driven by the usage of those assets, I want to account for that in gross margin. To do this I first start off with the companies Gross Margin (46.8%). I then add depreciation into the cost of goods sold.

Their Q3 2025 D&A breaks down as:

• Non-leasing D&A: $12.6M
• Depreciation of leasing equipment: $9.8M
• Total D&A: $22.4M

Scenario	Adjusted COGS	Adj. Gross Profit	Adj. Gross Margin
Add all D&A	$111.2M	$55.6M	33.4%
Add only leasing depreciation	$98.6M	$68.2M	40.9%

I broke apart the two options for depreciation because Non-Leasing D&A is not attached to the power solutions. But either way I see that the company has profitability from buying and running its assets.

Conclusion:

SEI is a growth company transitioning from oilfield services to AI data center power infrastructure. The numbers are compelling - revenue approximately doubling YoY with EBITDA tripling. Management has significant skin in the game with recent insider buying at lower prices.

The Stateline JV validates their value proposition to major hyperscalers. The 2,200 MW capacity target by 2028 represents massive growth from today’s 760 MW. If they hit their $600M+ pro forma earnings target, the current valuation looks reasonable.

The negative free cash flow is a yellow flag that bears watching, but hasn’t derailed execution. Supply chain constraints are real but manageable given their existing relationships.

I’m taking a mid-size position in this company. The AI power infrastructure theme has significant runway and SEI is well-positioned as an equipment provider rather than an operator. Next earnings call I’ll be focused on: capacity growth updates, Stateline progress, and any new customer announcements.

Drew,

Long

[GDavenport]

I designed and specified site emergency and stand-by power for greenfield sites and brownfield expansions where grid connect timelines and site development schedules required “behind the meter” developments. While almost all of my clients specified diesel engine driven generation, there is always the option of turbine generation if gas supply is collocated or readily available. Sizes (power delivery) for these modular units (40’ Sea Can or BiModal transport is typically the module for these) varies between 0.5MW to almost 3MW.

For this company, I can see nothing proprietary or moat related in their offer. This is strictly a peaking convenience play driven by urgent demand. Their margins are nice and are a direct clue to the supply/demand nature of these installations. If demand dries up or if DC strategy changes, quarterly delivery and contracted power could change quickly!

Because of the nature of this business, and the commodity nature of these installations, I would be reviewing the revenue features of these contracts, NOT necessarily ONLY the durations.

Once installed, what is the MIN power expected to be supplied (revenue recognition)? What is the portfolio median, range, average revenue recognition? etc. etc. etc.

High margins are a typical feature for peaking/standby/emergency/bridging power solutions. The down side of this is that those terms can be unfavorable enough that they actually promote grid connection sooner than later as soon as lower cost offerings are available. (This speaks to pricing pressure and fragile margins)

I see that they utilize battery installations as part of their offer to maintain service level agreement of 5 nines+. This is helpful to understand that they are potential supply bound by both the OEM for their turbine/engine generation train AND their battery supply vendor AND their integrator.

While I don’t see this as a limiting factor (look at their results), there is nothing particularly different than any other supplier/integrator and operator pairing other than, perhaps their desire to own all of the assets.

Their revenue drivers are now associated with new contracts and new installments. As their opportunity matures, it will pivot to owner/operator for their installed base.

Their service and power agreements will be the second (Power AAS) and third (Service/maintenance contract) profit drivers for them over time. Success (and durable growth in EBIT/Profit) will depend on shifting dynamics.

I would have to look deeply into their contracts to understand how to model this company. If they have lowish minimum power supply lower limits, they could be in the unenviable position of installing equipment that bridges center power from start up to grid connect, but then losing opportunity as that installed equipment sits at minimum revenue generation while they have to procure, integrate, deliver and connect entirely new modular power elsewhere.

This has nothing to do with technology, production or delivery and everything to do with contractual terms. They have to build/install power everywhere they go. This is not a long term guarantee of power sold.

(think about NVDA efficiency, AI service effectiveness and power consumption efficiency continuing to scale rapidly - Do we have dark data centers soon standing idle in minimum power states? In that scenario, does SEI continue to maintain equipment with low/no power sold?)

They are asset heavy (their supply chains are) so any bottlenecks or product pivots (between vendors/integrators/service supply contractors) will potentially hit their bottom line unless they are really nimble or growth is stable enough to allow strategy to drive consistency.

This is a power arbitrage opportunity, nothing else. I’m don’t think this is a great fit for a hypergrowth portfolio.

[drew1618t]

@GDavenport

think about NVDA efficiency, AI service effectiveness and power consumption efficiency continuing to scale rapidly - Do we have dark data centers soon standing idle in minimum power states? In that scenario, does SEI continue to maintain equipment with low/no power sold?)

First thank you for your inside view on this.

From my understanding of NVDA efficiency, is the compute per unit of electricity is going up. But its power demand per chip is going up. So 10,000 GPUs now use more power than 10,000 GPUs in the past. So we are seeing an increase of energy use in the sector even with the outputs increase in efficiency. We are also seeing a huge amount of new data centers being built. So energy demand is increasing as we move along as well.

From the conference call on SEI when asked about competitors they said this field will need lots of different companies to provide the needs of the growing power demand. CFO Kyle Ramachandran said this.

Yes. And I’d say just to put it in context, so the 2,200 megawatts we’re talking about here, to satisfy the leading-edge sort of incremental data center. I’m not sure that satisfies even 2 at this stage. So the sizes of these infrastructure projects continue to grow, to Bill’s point, such that it’s just a large market.

@GDavenport

The down side of this is that those terms can be unfavorable enough that they actually promote grid connection sooner than later as soon as lower cost offerings are available.

The point SEI is making is that instead of waiting 5+ years for a grid connection, you can pay a premium for power today and get billions of dollars’ worth of chips working immediately. A B200 draws roughly 1 kW; at $0.0865/kWh that’s about $900 per year in electricity. If a B200 costs $45,000 and is depreciated over five years, that’s $9,000 per year in depreciation. Electricity is a rounding error compared to the cost of the silicon itself.

I think with the current market conditions heavily favoring electrical producers and SEI growing revenue and profit make it an interesting AI play.

Drew

[GDavenport]

Drew,

I agree with your points. The background in my post was to provide context to how DURABLE this revenue and growth is. I don’t think this is a near term concern (next 18 months - there are simply too many requests for local generation). I should clarify that I did not consider this company as being a hypergrowth fit for my portfolio and not that I’m bearish on their plans. This was confusing and I’ll edit above to clarify.

A rising tide lifts all boats. For the sake of adding some research on company performance, I did a comparison between SEI and GNRC

SEI compares quite favorably in many metrics and provides similar ecosystem product as GNRC but with a different sales/ownership model

There is much more detail below:

Check out the Concerto Platform from Generac
The Concerto platform - Generac Grid Services

From Cushman & Wakefield regarding Datacenter real estate growth and development trends.

Six for 2026: U.S. Real Estate Trends to Watch | US | Cushman & Wakefield

and for locations and sizes, click the tabs in this infographic:

Americas Data Center Update | Cushman & Wakefield

[anthonyms]

This is an interesting one. I notice Beth Kindig picked a behind the meter AI DC energy play - Bloom Energy, as one of the 2 picks for the year (alongside MercadoLibre).

Bloom Energy is doubling its capacity this year to deliver onsite power using solid oxide fuel cells and is growing revenues at over 50% reaching the $500m/quarter level.

I haven’t looked into it much at this point but will alongside SEI.

Ant