Joulent

1.1 Identity and business model

Joulent is a Houston, Texas-based, technology-driven energy company that builds large-scale, co-located power infrastructure for AI data centers and industrial users, publicly launching in June 2026 after a multi-year incubation by investment firm Engine No. 1 in collaboration with GE Vernova. Its productized 'Across-the-Meter' power stack co-locates gas generation, battery storage and renewables with customer demand and manages grid interconnection to deliver reliable electricity at industrial scale. The company positions itself around delivering energy at the 'speed and scale of American innovation,' targeting the acute power bottleneck facing hyperscale computing. Joulent is privately held; its stage is best described as growth-phase development, with its first flagship project still pre-Final-Investment-Decision as of July 2026.[CO001, CO002, CO003, CO015, CO024, CO030]

1.2 Leadership and governance

Joulent's public face is founder and chief executive Chris James, who simultaneously leads Engine No. 1, the activist-investment firm he founded in 2020 and that famously won three ExxonMobil board seats in 2021. That dual role gives Joulent an experienced, credible energy-transition leader but also concentrates key-person risk, because no other named executives have been disclosed publicly. James frames leadership in the AI era as a race to deliver energy and compute the fastest, most reliably and at the lowest cost—the thesis underpinning Joulent's strategy. Governance detail beyond the founder is thin: the senior operating team, board composition and the balance of control between Engine No. 1 and new investor National Grid Ventures are not yet public, leaving a material diligence gap around management depth.[CO004, CO005, CO026, CO036, CO032]

1.3 Funding, valuation and ownership

On July 1, 2026, National Grid Ventures—the commercial, non-regulated arm of UK utility National Grid plc—agreed to invest $1.75 billion in Joulent for an approximately 35% stake, a transaction Kirkland & Ellis advised on. That price implies a headline valuation of roughly $5 billion, vaulting the young company to unicorn scale, though the figure is derived from the deal rather than an audited mark. National Grid characterizes the investment as balance-sheet funded and incremental to its record capital program, having earmarked about $1 billion of near-term capital for Ventures projects. The capital is intended to fund Project Kilby and additional large-scale projects, giving National Grid direct exposure to fast-growing US 'large load' contracted-power demand while leaving Joulent's residual ownership and full cap table undisclosed.[CO006, CO007, CO008, CO021, CO022, CO023]

1.4 Flagship project and partnerships

Joulent's flagship is Project Kilby, a 2.67-gigawatt co-located power campus on a 2,000-plus acre site near Pecos in Reeves County, Texas, in the Permian Basin—capacity roughly equivalent to two million homes. Chevron, through subsidiary Energy Forge One LLC, holds a 20-year power purchase agreement with Microsoft, whose adjacent Pecos data-center campus is planned to add about 2 GW of compute load over five to seven years; Joulent holds a 50% equity option in the joint venture. GE Vernova supplies the core turbine technology and Caterpillar's Solar Turbines division adds supplemental generation. The project, projected to cost $7-9 billion, targets a Final Investment Decision by end-2026 and first power in 2028, and is claimed to generate over $10 billion in state and local tax revenue and roughly 2,000 jobs.[CO009, CO010, CO011, CO012, CO013, CO014]

1.5 Milestones and trajectory

Joulent's short history traces to Engine No. 1's 2020 founding and its 2021 ExxonMobil campaign, followed by a roughly three-year incubation of the power platform with GE Vernova beginning around 2023. Momentum accelerated sharply in 2026: a January announcement that Chevron, GE Vernova and Engine No. 1 would co-locate up to about 4 GW of gas plants for US data centers; the June 22 signing of the Chevron-Microsoft 20-year Kilby agreement alongside Joulent's public launch; and the July 1 National Grid Ventures investment. Ahead lie the targeted end-2026 Final Investment Decision and 2028 first power. Independent observers, however, caution that large new gas plants built for data centers carry execution, emissions and stranded-asset risks, tempering the otherwise steep upward trajectory.[CO029, CO031, CO006, CO009, CO032]

1.6 Exhibits

2.1 Market boundary and substitutes

The market Joulent addresses is the supply of large-scale, contracted electricity and the physical power infrastructure that serves 'large load' data centers, above all AI compute campuses. In-scope spend covers new generation capacity, grid interconnection, co-located and behind-the-meter plants, and long-dated power purchase agreements; it explicitly excludes IT hardware such as GPUs, servers and cooling. The status-quo substitute is a conventional grid interconnection through a regulated utility, an option increasingly constrained by multi-year queues that have become a central pain point for hyperscalers. Adjacent supply paths—grid-scale independent power producers, nuclear restarts and small modular reactors, and fuel-cell or hybrid on-site systems—compete at the margin. Behind-the-meter and co-located gas generation has emerged as the fastest route to reliable power, which is exactly the segment Joulent productizes. The boundary matters for diligence because headline 'data-center' spending figures blend IT hardware, real estate and power; isolating the power-and-interconnection layer is what makes Joulent's opportunity legible, and it is that narrower slice—new dispatchable capacity delivered on a hyperscaler's timeline—where the company competes rather than the far larger compute-hardware budget.[CM001, CM002, CM003, CM004, CM005, CM030]

2.2 Market sizing across multiple lenses

No single number captures this market, so it must be triangulated across lenses. On a consumption basis, the IEA projects global data-center electricity use roughly doubling to about 945 terawatt-hours by 2030. On a capacity basis, blended estimates put data-center power demand near 132 gigawatts in 2026, rising toward roughly 290 gigawatts by 2030, with Gartner alone expecting about 26% consumption growth in 2026. On a capital basis, McKinsey and JP Morgan frame $5-7 trillion of global data-center investment this decade, with McKinsey's headline near $6.7 trillion, while Goldman Sachs models a 165% jump in power demand by 2030. Hyperscaler capex of roughly $690 billion in 2026—Microsoft alone near $190 billion—underscores the spend behind these curves. Crucially, only a fraction of this is addressable by any one co-located developer, so Joulent's serviceable and obtainable market is far smaller than the headline totals and is not publicly quantified.[CM006, CM007, CM008, CM009, CM010, CM011]

2.3 Buyer segmentation and adoption path

Demand concentrates among hyperscale cloud and AI operators—Microsoft foremost—alongside colocation providers, REITs and large industrial users. The economic payer is typically the hyperscaler, which contracts power under 10-to-20-year purchase agreements, with budget ownership sitting in infrastructure and energy-procurement teams steering multi-billion-dollar capex programs. The adoption path is long and capital-intensive: site selection and land control, then generation and interconnection, followed by a long-term power agreement, financing and a final investment decision, and finally construction and energization over several years. Project Kilby—where Chevron holds a 20-year PPA with Microsoft and Joulent an equity option—illustrates this motion end to end. Increasingly, AI operators co-locate compute directly with dedicated generation to compress time-to-power, turning speed into a procurement differentiator and making contracted, long-duration offtake the linchpin of the investment case.[CM016, CM017, CM018, CM019, CM020, CM035]

2.4 Growth drivers and constraints

The dominant driver is the AI-compute surge, which is far more power-hungry than prior workloads and is colliding with grid interconnection queues that often stretch for years—together the core reason hyperscalers turn to co-located generation, where speed-to-power and long-dated offtake de-risk new plants. Against these drivers sit real constraints. Texas's Senate Bill 6 and ERCOT's new large-load interconnection standards impose study fees, per-megawatt financial security and disclosure obligations on loads of 75 megawatts and above, while regulators grapple with the systemic strain of concentrated demand. Capital intensity—power plants costing millions of dollars per megawatt—limits how quickly supply can scale, tight gas-turbine availability adds a physical bottleneck, and mounting emissions scrutiny pressures new gas-fired capacity. Finally, the demand forecasts themselves carry wide error bars: sustained AI adoption, efficiency gains or a financing pullback could each move the trajectory materially, and the divergence of published estimates underscores that uncertainty. For diligence, the constraint set is as important as the demand curve, since it determines which developers can actually convert demand into energized, contracted megawatts.[CM021, CM022, CM023, CM024, CM025, CM026]

2.5 Exhibits

3.1 Competitive landscape

Joulent competes in a crowded, fast-moving field to supply reliable power to AI data centers. The landscape spans four archetypes: vertically integrated AI-power developers such as Crusoe; behind-the-meter and on-site specialists like VoltaGrid and Bloom Energy; gas and nuclear independent power producers including Vistra, Constellation and Talen; and comparable co-located gas developers such as Homer City with Kiewit. Underpinning all of them is the status-quo alternative—a standard grid interconnection through a regulated utility—which remains the default but is increasingly constrained by multi-year queues. Traditional IPPs now face direct competition from distributed and co-located power, and hyperscalers can also 'build internally' by contracting their own generation, an implicit competitor to every third-party developer. This breadth means Joulent must win not on a single axis but across speed, siting, fuel and capital simultaneously. The field is also fluid: nuclear restarts, gas co-location and behind-the-meter fuel cells are all being underwritten by the same handful of hyperscalers, so competitors frequently overlap on customers even when their technologies differ. For a pre-FID entrant, the practical question is less who the competitors are than which of them can actually deliver energized megawatts on a hyperscaler's timeline, and on that test the incumbents with operating fleets start ahead.[CP001, CP002, CP003, CP008, CP012, CP013]

3.2 Peer profiles and funding

The most direct scaled peer is Crusoe, a vertically integrated AI-infrastructure company that reached roughly a $10 billion valuation after a $1.375 billion Series E in late 2025 and has been reported raising a pre-IPO round in 2026; its blend of data-center development, power sourcing and AI cloud gives it a revenue base Joulent lacks. Behind-the-meter specialists VoltaGrid and Bloom Energy compete on distributed gas and fuel-cell generation, while colocation developers like Vantage and Equinix arrange dedicated on-site power. Among IPPs, Vistra runs a large gas-and-nuclear fleet, Constellation is restarting Three Mile Island for Microsoft, and Talen has struck nuclear deals with Amazon—carbon-free, round-the-clock offerings that contrast with gas. Homer City's roughly 4.5-GW GE-Vernova-powered Pennsylvania campus shows the co-located gas model being replicated at gigawatt scale by others.[CP004, CP005, CP006, CP007, CP009, CP010]

3.3 Differentiation and pricing

Joulent differentiates less on any single feature than on a bundled package: Chevron's fuel supply and balance sheet, GE Vernova's scarce turbine slots, National Grid's strategic capital and a Microsoft anchor—an assembly no single rival replicates. At a roughly $5 billion implied valuation it sits below Crusoe's ~$10 billion, a gap that reflects Crusoe's broader operating platform as much as stage. Competition across the field turns on speed-to-power, siting and access to firm fuel or turbines rather than published price, since pricing is set through bilateral 10-to-20-year PPAs; in Kilby's case Chevron holds the Microsoft agreement. Access to scarce gas turbines is a shared gating factor, and notably Joulent's own supplier, GE Vernova, equips competing projects, so the equipment advantage is real but not exclusive.[CP019, CP020, CP021, CP022, CP023, CP024]

3.4 Moat durability and competitive risk

Joulent's most durable moat is contractual: long-dated PPAs create high switching costs once signed, and site control, interconnection rights and turbine delivery slots are hard-to-replicate assets. But durability is conditional. Distribution power and capitalization favor incumbents like Vistra and Constellation that already own fleets and grid relationships, and hyperscalers multi-home across nuclear, gas and grid to avoid dependency. The co-located gas model is being commoditized as many developers copy it, and carbon-free rivals—nuclear and renewables-plus-storage—could displace gas on emissions and long-run cost, a threat amplified by rising sustainability scrutiny. Turbine scarcity shared with rivals could delay Joulent relative to nuclear peers with operating plants, and better-capitalized incumbents can outbid a single-project entrant. The moat is genuine but rests heavily on reaching Kilby's final investment decision and retaining its partner bundle. Until first power in 2028, every element of that moat is prospective rather than proven, which is the central competitive vulnerability an investor must weigh against the strength of the assembled partners.[CP025, CP026, CP027, CP028, CP029, CP030]

3.5 Exhibits

4.1 Revenue streams and pricing

Joulent's revenue model is built on long-term power purchase agreements for electricity supplied to data centers and industrial users. For Project Kilby, Chevron—through its Energy Forge One subsidiary—holds a 20-year PPA with Microsoft, and Joulent participates via a 50% equity option in the joint venture, its primary route to project economics. Such agreements typically run 10-to-20 years with take-or-pay-style commitments, and pricing is negotiated bilaterally rather than published, commonly cited in the tens of dollars per megawatt-hour for the segment. Revenue would be recognized as energy is delivered and capacity is made available once the plant operates; today Joulent is pre-revenue on Kilby, with first power targeted for 2028. The result is a structurally clear but quantitatively opaque top line: the contract exists, but price, escalation and Joulent's exact economic share are undisclosed.[CI001, CI002, CI003, CI004, CI005, CI006]

4.2 Go-to-market, cost structure and unit economics

Joulent's go-to-market inverts the usual sequence: it secures an anchor hyperscaler offtake before committing capital, as the Microsoft agreement preceding Kilby's final investment decision illustrates, and leans on partner channels—Chevron, GE Vernova, National Grid—rather than a conventional sales force. Customer acquisition is therefore highly concentrated, with a single anchor underpinning the first project, and the sales cycle for gigawatt-scale deals stretches across months to years of siting, interconnection and contracting. On costs, the model is intensely capital-intensive: Kilby is projected at roughly $7-9 billion, and co-located gas generation runs into the millions of dollars per megawatt, with GE Vernova turbines and Caterpillar units as major line items. Margins hinge on the spread between contracted PPA revenue and natural-gas fuel plus operating costs, amortized over high utilization and a long contract to recover heavy front-loaded capex. Importantly, an off-grid, behind-the-meter design lets Kilby avoid socializing grid-upgrade costs and sidestep multi-year interconnection queues, which both supports standalone project economics and improves time-to-revenue relative to a grid-dependent build.[CI008, CI009, CI010, CI011, CI012, CI013]

4.3 Public traction and capital adequacy

As a privately held company, Joulent discloses no audited revenue, ARR or profit; its public traction signals are the signed Microsoft PPA and the $1.75 billion strategic investment, with Kilby's 2.67 GW of contracted capacity serving as the headline forward metric in lieu of revenue. On capital, National Grid Ventures agreed on July 1, 2026 to invest $1.75 billion for roughly a 35% stake—implying about a $5 billion valuation—describing the deal as balance-sheet funded and incremental to its record capital plan, having earmarked around $1 billion of near-term capital for Ventures. That equity, however, is small against a $7-9 billion Kilby build, which implies substantial additional project finance or partner capital, plausibly debt secured against the 20-year Microsoft PPA. The next major financing trigger is Kilby's targeted end-2026 final investment decision, and Chevron's fuel supply and balance-sheet depth materially reduce Joulent's standalone financing burden.[CI019, CI020, CI021, CI022, CI023, CI024]

4.4 Financial verdict and diligence blockers

The financial picture is a high-conviction option rather than a proven business. On the positive side, once operational, Kilby's cash flows would be anchored by a 20-year agreement with an investment-grade counterparty, supporting high revenue quality and project-finance leverage. Against that, the capital intensity and long build cycle back-end returns and expose the model to cost overruns and delays, and skeptics warn that large new gas plants for data centers carry stranded-asset and return risks if AI demand or climate policy shifts. The single largest diligence blocker is the near-total absence of public financials—revenue, margins, burn, the full cap table and project-finance terms are all undisclosed—so the implied ~$5 billion valuation is a pre-revenue, derived mark resting on contracted pipeline rather than current cash flows. The verdict: promising structure, unproven economics, and disclosure gaps that must close before underwriting, with the end-2026 final investment decision the first hard checkpoint at which the investment thesis either de-risks materially or stalls outright.[CI032, CI033, CI034, CI035, CI036]

4.5 Exhibits

5.1 Product definition in customer workflow terms

Joulent's product is not a device or a piece of software but firm electrical power delivered as a service to hyperscale data centers and industrial users. The company finances, builds, owns and operates the generation and sells electricity under long-term agreements, so from the customer's perspective it is buying guaranteed capacity rather than equipment. The flagship, Project Kilby, is engineered to deliver roughly 2.67 GW to serve Microsoft's AI compute. The job it solves is acute: hyperscalers need multi-hundred-megawatt firm power on a timeline the congested public grid cannot meet, and interconnection queues stretch for years. By co-locating generation with compute in an off-grid, behind-the-meter campus that islands its load from ERCOT, Joulent removes the transmission bottleneck entirely and shortens the electrical path from turbine to server. In workflow terms, the customer swaps a multi-year wait for grid capacity for a purpose-built plant delivered on a private schedule.[CE001, CE002, CE003, CE004, CE005, CE006]

5.2 Asset and product-line map

The generation fleet centers on GE Vernova 7HA-class heavy-duty gas turbines configured in combined-cycle blocks, supplemented by Caterpillar Solar Turbines units for firming and flexibility. The campus occupies more than 2,000 acres near Pecos in Reeves County, Texas, in the heart of the Permian Basin, which places the plant atop abundant, low-cost natural gas and gives it local fuel access without long pipelines. The asset is held through a joint venture in which Chevron carries the 20-year Microsoft power purchase agreement while Joulent holds a 50% equity option—the structure that converts the physical plant into Joulent's economic product. Beyond Kilby, the product line is best understood as a pipeline of additional co-located campuses that the $1.75 billion National Grid Ventures capital is meant to seed, though no second project has yet been named. The asset map is therefore concrete for Kilby's turbine fleet and site but still thin on exact equipment counts and future-project specifics.[CE007, CE008, CE009, CE010, CE011, CE012]

5.3 Operating architecture and technical mechanism

Kilby's architecture stacks proven components into an islanded system. Gas turbines burn Permian fuel to spin generators, and a heat-recovery steam cycle captures exhaust heat to drive a steam turbine, lifting fuel-to-power efficiency well above simple-cycle operation; GE Vernova's 7HA platform reaches combined-cycle efficiencies in the 63-64% range, among the highest in commercial gas generation, and is engineered to co-fire hydrogen blends for a partial decarbonization pathway. Because the campus runs off-grid, grid-orchestration software in the GridOS class must continuously balance generation, storage and load, and the plant must self-supply frequency regulation, spinning reserve and black-start capability that the public grid would normally provide. Behind-the-meter delivery shortens the electrical path from turbine to server, cutting transmission losses and congestion-pricing exposure. The operating model front-loads capital into turbines, heat recovery and controls, then earns over decades of high-utilization dispatch. Every layer is individually proven, but integrating them at 2.67 GW for a single AI customer is the central technical unknown.[CE013, CE014, CE015, CE016, CE017, CE018]

5.4 Deployment, reliability, support and roadmap

The deployment path runs from a targeted end-2026 final investment decision to a 2028 first-power goal, implying a compressed multi-year engineering-procurement-construction cycle with more than 6,000 workers at peak. A critical gating dependency is turbine availability: GE Vernova's heavy-duty gas-turbine backlog is effectively sold out into the late 2020s, so securing delivery slots matters as much as capital or permits to hitting 2028. Reliability for an AI training campus depends on very high availability, so redundant turbine blocks and on-site backup are integral to the design, and a phased buildout lets capacity come online in blocks—allowing partial revenue before the full 2.67 GW is complete. Long-term support and maintenance are expected to lean on OEM service agreements from GE Vernova and Caterpillar rather than an in-house field organization. Taken together the roadmap is milestone-clear but every post-FID date remains a target rather than a committed, financed schedule, so slippage risk is real.[CE020, CE021, CE022, CE023, CE024, CE025]

5.5 Differentiation and competitive moat

Joulent's primary differentiation is speed-to-power: delivering firm gigawatt-scale capacity years faster than a grid-interconnected alternative by bringing generation to the site rather than waiting on transmission. Around that sits a partner-access moat—privileged supply of GE Vernova turbines, Chevron fuel and balance sheet, and a Microsoft anchor offtake—that is hard for a new entrant to assemble. Physical site control over 2,000-plus acres with co-located gas is itself a scarce, hard-to-replicate asset, and Engine No. 1 incubation plus Chris James's energy-transition network give Joulent capital-markets and dealmaking reach uncommon for a young developer. The important caveat is that the behind-the-meter concept is not proprietary: competitors from Crusoe to traditional IPPs are pursuing the same co-location thesis, so the durable edge is first-mover execution and integration quality, not the idea. On a maturity map, the components are proven while the differentiated, still-unproven layer is islanded integration at full scale.[CE027, CE028, CE029, CE030, CE031]

5.6 Trust, safety, compliance and quality controls

As large combustion sources, Kilby's turbines require federal and state air-quality permitting for pollutants such as NOx and CO2, and islanding off ERCOT reduces interconnection-approval exposure but does not exempt the plant from environmental and safety regulation. The design draws pointed criticism: analysts and environmental groups argue that a large gas-fired plant for data centers carries meaningful carbon and stranded-asset risk if AI demand or climate policy shifts, and no public lifecycle-emissions model has been released. Hydrogen co-firing offers a compliance hedge, but current blends only modestly reduce lifecycle emissions. Reliability and safety quality are governed largely by OEM engineering standards and long-term service agreements rather than independent audit, and no third-party review of Kilby's design, emissions modeling or reliability engineering is yet public. Water sourcing for combined-cycle steam and cooling in arid West Texas is a further unresolved operational and compliance question. In short, compliance obligations are identifiable but publicly under-documented.[CE032, CE033, CE034, CE035, CE036, CE037]

5.7 Exhibits

6.1 Customer base and segmentation

Joulent aims at two customer segments: hyperscale data-center operators and large industrial power users. In practice the base today is a single, decisive account—Microsoft—which is simultaneously the buyer, user and payer for the roughly 2 GW of Project Kilby capacity it has contracted to feed its AI compute. Geographically the footprint is concentrated in Texas at the Kilby campus in Reeves County, and the core use case is powering AI training and inference that cannot obtain timely grid capacity. Industrial users are named as a target segment, but no specific industrial customer has been disclosed, and no second hyperscaler has been announced. The result is a segmentation story that is aspirational beyond one account: the market opportunity spans multiple verticals, yet the concrete customer reality is one hyperscaler, mediated commercially through Chevron's joint-venture PPA. This shapes every downstream question about adoption, retention and concentration.[CU001, CU002, CU003, CU004, CU005, CU006]

6.2 Adoption trajectory and demand

Adoption is contractual rather than operational. Kilby is contracted but not yet delivering power, so there is no active usage, no repeat purchase and no utilization data; the headline metric is contracted capacity—about 2.67 GW in total, of which Microsoft has effectively reserved roughly 2 GW—rather than deployed or utilized megawatts. First power is targeted for 2028, meaning any deployment, account-growth or utilization figures simply do not exist yet, and operating capacity today is zero. What gives the reservation weight is the demand behind it: Microsoft's broader AI buildout, including its large Fairwater-class campuses, creates the compute-power appetite that the Kilby contract is meant to satisfy. In diligence terms, the adoption trajectory must be read as a forward book underwritten by a single hyperscaler's capital-expenditure plans, with all of the realization risk concentrated in the construction period between the end-2026 investment decision and 2028 commissioning.[CU007, CU008, CU009, CU010, CU011, CU012]

6.3 Named customer proof and reference quality

The Microsoft relationship is the single named customer proof point, evidenced by a 20-year power purchase agreement held through Chevron's Energy Forge One entity rather than a pilot or letter of intent—a production-grade commercial commitment. Microsoft has publicly discussed securing dedicated power for AI infrastructure, corroborating the strategic intent, and the proof is fresh, tied to Joulent's June-July 2026 launch and investment announcements. Reference quality is high on contract certainty but necessarily low on operational outcomes, which cannot exist before first power in 2028. Beyond Microsoft, no second named customer proof point has been made public, and pricing, volume-ramp and service-level terms of the agreement remain undisclosed. The proof therefore establishes that a marquee hyperscaler has committed for two decades, which is a strong signal, while leaving open every question about realized performance, unit economics and whether the model can attract additional named accounts.[CU013, CU014, CU015, CU016, CU017, CU018]

6.4 Retention, durability and switching costs

Durability is the strongest part of the customer story. The 20-year PPA gives exceptional contractual longevity and revenue visibility once power flows, and take-or-pay-style structures common to such agreements further lock in commitment. Switching costs for the customer are very high because the plant is co-located and purpose-built for the campus, so the counterparty cannot easily shift load elsewhere, and Microsoft's investment-grade credit reduces default risk. Conventional retention metrics—net revenue retention, gross retention, churn and renewal cohorts—are all null because the customer is not yet live; contract length is the primary durability signal in lieu of cohort data. The honest caveat is that even a strong 20-year contract does not eliminate construction- and delivery-phase risk to the relationship: a delayed or over-budget build could strain the agreement before it ever generates revenue. Durability is thus contractually excellent but operationally untested.[CU019, CU020, CU021, CU022, CU023, CU024]

6.5 Expansion, concentration and skeptical views

Customer concentration is extreme: essentially one customer underpins the entire flagship, making single-customer dependence the dominant commercial risk. The land-and-expand path is to replicate the co-located model with additional hyperscaler campuses funded by National Grid Ventures capital, but expansion depends heavily on partners—Chevron for fuel and PPAs, GE Vernova for turbines—and on long, high-friction procurement cycles for gigawatt-scale deals. Skeptics add two adverse lenses: they question whether hyperscalers' reliance on new gas plants is durable given climate commitments, and they note that Microsoft has multiple alternative power-sourcing options—nuclear restarts, renewables and rival developers—which limits Joulent's leverage. Community and environmental pushback around large gas-for-data-center projects could further pressure customer optics. A downturn in Microsoft's AI capital expenditure or a strategy shift would directly threaten Kilby's economics, so diversifying beyond Microsoft is the key milestone to de-risk the customer base, and as of mid-2026 that remains unmet.[CU025, CU026, CU027, CU028, CU029, CU030]

6.6 Exhibits

7.1 Severity-ranked risk overview

Joulent's risk profile is dominated by two forces: execution and concentration. The single largest risk is delivering a $7-9 billion, gigawatt-scale gas plant on a compressed 2026-2028 schedule, with first power not expected until 2028—meaning years of pre-revenue construction exposure before any offsetting income. Capital intensity magnifies the downside, because cost overruns on a multi-billion-dollar build erode equity returns directly. Alongside execution, customer concentration is a top-tier risk: essentially one customer, Microsoft, underpins the flagship. Residual exposure remains high across most categories because the project is pre-construction and largely undisclosed, and mitigation maturity is uneven—strong contractual and partner mitigants coexist with weak public transparency and unproven operations. Ranked by residual severity, the priority risks are execution/construction, customer concentration, turbine supply, financing gap, evolving Texas large-load regulation, and carbon/stranded-asset exposure. Each is examined in turn, with monitorable triggers that would break the thesis.[CR001, CR002, CR003, CR004, CR005, CR006]

7.2 Regulatory and legal risk

The Texas regulatory backdrop shifted with Senate Bill 6 (2025), which tightens rules for large electric loads interconnecting to the grid, and ERCOT's large-load process (Project 58481) imposes study fees, disclosure and security requirements on loads at and above roughly 75 MW. SB6 also gives ERCOT and utilities expanded authority to curtail or disconnect large loads under grid-stress conditions, and legal analysts flag heightened disclosure and demand-transparency obligations for data-center loads. Joulent's off-grid, behind-the-meter design reduces—but does not fully remove—exposure to these interconnection rules, since the campus islands its load. Separately, air-quality permitting for large gas turbines (NOx and CO2) is a required, time-consuming step; environmental litigation or permit challenges are a plausible legal risk for large fossil projects; and water-use permitting in arid West Texas adds a further dependency for combined-cycle cooling. Federal energy oversight can also bear on co-location and wholesale-power arrangements. The register is ordered by severity, but Kilby-specific filings are not yet public, so residual exposure is only partially sized.[CR007, CR008, CR009, CR010, CR011, CR012]

7.3 Operational and supply-chain risk

Operational risk concentrates in turbine supply. GE Vernova's heavy-duty gas-turbine backlog is effectively sold out into the late 2020s, and strong global demand creates delivery-slot risk that could slip Kilby's timeline. Construction-phase risks include labor availability at a 6,000-plus-worker peak, EPC coordination and West Texas weather. Once built, operating an islanded 2.67 GW microgrid introduces reliability risk in self-supplying the frequency regulation, reserve and black-start services the grid would normally provide, and a single-site plant means an outage or equipment failure would directly threaten the anchor customer's power. Natural-gas supply disruptions or price spikes are an operational and margin risk despite Permian proximity. Compounding all of this, no public reliability, safety or quality-audit record yet exists for Kilby's specific design, so operational risk remains largely unquantified until commissioning. Rows in the operational register are ordered by severity, with turbine delivery, construction overrun and islanded reliability at the top.[CR016, CR017, CR018, CR019, CR020, CR021]

7.4 Partner and dependency risk

Joulent sits at the center of a web of high-concentration dependencies. Chevron dependence is high—it holds the Microsoft PPA and supplies fuel, so its commitment is load-bearing—and GE Vernova dependence is high for both turbines and grid-orchestration technology. National Grid Ventures is the key capital provider, so any change in its commitment would strain financing, while Microsoft is simultaneously the key customer and the demand rationale, compounding concentration. Regulators—ERCOT, the Texas PUC and environmental agencies—are external dependencies that can gate the project. The joint-venture structure means Joulent's economics depend on partner performance it does not fully control, and failure of any single critical partner in fuel, turbines, capital or offtake could stall the flagship. On the people side, execution hinges on a young organization delivering a first-of-scale megaproject, with key-person dependence on founder-CEO Chris James and a need for deep project-development and EPC-management bench strength. Rows are ordered by the severity of a partner or execution failure.[CR023, CR024, CR025, CR026, CR027, CR028]

7.5 Financial, carbon and mitigation triggers

Financially, the $1.75 billion equity raised is small versus a $7-9 billion build, implying substantial additional financing risk, and the long pre-revenue period means sustained cash outflow before any income. Margin-compression risk arises if gas prices rise or PPA pricing is fixed without adequate pass-through, and leverage against the 20-year PPA—though plausible—exposes the model to interest-rate and credit conditions. The near-total absence of public financials is itself a diligence risk that limits independent verification, and a downturn in AI data-center capex would undermine both demand and the valuation basis. The most poorly mitigated tail risk is carbon: critics warn that large new gas plants for data centers carry stranded-asset risk if AI demand or climate policy shifts, environmental groups highlight the climate tension, and hyperscaler climate commitments could pressure the customer to limit gas power over time. Key mitigants—the 20-year PPA, an investment-grade offtaker, strong partners and hydrogen-capable turbines—reduce but do not eliminate these risks. The clearest thesis-break triggers are FID slippage past end-2026, loss of the Microsoft contract, or a major cost blowout.[CR030, CR031, CR032, CR033, CR034, CR035]

7.6 Exhibits

8.1 Investment thesis and anti-thesis

The investment thesis is that Joulent is an early leader in the scarce, fast-growing market for co-located power for AI data centers, with a signed 20-year Microsoft power purchase agreement and a $1.75 billion strategic investment that validate demand and de-risk the anchor project. Privileged partner access—Chevron for fuel and offtake, GE Vernova for turbines, and National Grid for capital—forms a defensible ecosystem advantage that a new entrant would struggle to assemble. The anti-thesis is equally clear: Joulent is a pre-revenue, single-project, single-customer developer valued at roughly $5 billion on execution optionality, where capital intensity, carbon exposure and construction risk could impair returns if execution slips. Both cases hinge on the same variables: the thesis holds if Kilby reaches final investment decision and first power on schedule with the customer intact, and the anti-thesis holds if any of those fail. The valuation debate is therefore really an execution debate wearing a price tag.[CV001, CV002, CV003, CV004, CV005, CV006]

8.2 Recommendation, confidence and stance

On balance the recommendation is to monitor: Joulent is a high-quality option rather than an underwritable business today. Confidence is medium, constrained above all by the near-total absence of public financials, which prevents any independent underwriting of revenue quality, margins or returns. The risk rating is high, driven by execution, customer concentration and capital-intensity risk, and the valuation stance is premium because the roughly $5 billion mark prices in successful execution that has not yet been demonstrated. The supportable posture is to track the end-2026 final investment decision and the 2028 first-power milestone before committing capital, re-rating as each de-risks. Framed against the alternatives, an outright invest call is premature given the opacity, while a pass would be too dismissive of a genuinely strong, well-partnered option on a scarce resource. Monitor threads that needle: it preserves access to the upside while declining to pay the full execution premium up front, and it defines concrete milestones that would justify moving to a firmer stance.[CV007, CV008, CV009, CV010, CV011, CV039]

8.3 Financing and valuation context

The July 2026 National Grid Ventures investment—$1.75 billion for about 35%—implies an approximate $5 billion post-money valuation. It is a strategic minority investment rather than a control transaction, which shapes preference and governance dynamics and could leave dilution or preference overhang that weighs on future common returns. Entry discipline is challenged because the valuation rests on contracted pipeline rather than current earnings: public evidence supports the existence and size of the deal but offers no independent valuation cross-check, and the mark is a private, primary-round figure rather than a market-tested or liquidity-backed one. In diligence terms this means the price should be read as an option premium on execution, not a multiple on cash flow. An investor paying at or above this level is underwriting flawless delivery of a first-of-scale megaproject, so the burden falls on verifying the contracted economics and the financing plan behind the $7-9 billion build before treating the ~$5 billion as a floor rather than a hopeful mark.[CV012, CV013, CV014, CV015, CV016, CV017]

8.4 Bull, base and bear scenarios

The outcomes are unusually binary because value concentrates in a single megaproject. In the bull case, Kilby executes on time, Joulent replicates the co-located model with additional campuses, and it becomes a multi-gigawatt power platform worth well above $5 billion. In the base case—assigned the highest probability given contracted demand—Kilby is built with some delay or cost overrun and Joulent remains a valuable but single-asset developer near its current mark. In the bear case, the final investment decision slips or Microsoft demand softens, cost overruns bite, and the equity is impaired below the entry price. Downside triggers include FID slippage, PPA renegotiation, capex blowout above $9 billion, and AI-capex retrenchment. Probability signals favor the base case, but the tail risk on execution is material, and the resulting valuation range is wide—from roughly impaired equity in the bear case to several times the current mark under successful scaling. That spread, more than any single point estimate, is the honest summary of Joulent's valuation.[CV018, CV019, CV020, CV021, CV022, CV023]

8.5 Comparables, exit readiness and diligence

Comparables bracket the mark. Crusoe Energy, a comparable AI-data-center power and compute developer, was valued around $10 billion in its late-2025 round; data-center infrastructure assets have traded at roughly 15-19x EBITDA; and listed independent power producers exposed to data-center demand—Vistra, Constellation, Talen—provide public-market reference points, while private strategic rounds in AI-power infrastructure have clustered at multibillion-dollar valuations through 2025-2026. Together these suggest Joulent's ~$5 billion mark is within range for the theme but rich for a pre-revenue single-asset developer, and every comparable has limits: Crusoe has compute revenue, IPPs are operating utilities, and infrastructure multiples assume cash-flowing assets. A DCF on Kilby would hinge on PPA pricing, gas cost, utilization and discount rate, all largely undisclosed. Exit readiness is early—pre-revenue status and single-asset concentration limit near-term liquidity—though an IPO, strategic sale, or a National Grid buyout are plausible paths. Final diligence asks center on financials, the cap table, PPA economics and the project-finance plan, and skeptics rightly caution that theme-driven valuations can overshoot fundamentals.[CV024, CV025, CV026, CV027, CV028, CV029]

8.6 Exhibits