"When you get to the middle of 2028 we start to break very materially below in a historical way where gas available in storage has ever been before." - Patrick's Guest (Matt) [00:23:04]
"There is a tremendous inertia around natural gas being the primary fuel to power AI." - Patrick's Guest (Matt) [00:20:53]
"We think that complacency is going to take us right up to the point where it's too late." - Patrick's Guest (Matt) []
Disclaimer: Orignal content owned by or sourced from third parties. It does not represent the views of 'Nuggets' platform or it's team. AI is used extensively across this platform including for summaries. Accuracy is not guaranteed, there can be mistakes. Any info or content on this platform is not a financial, legal, or investment advice. Do your own research. Refer for complete disclosures:- Terms of Use · Full Disclaimer
"The US consumer is probably going to pay the bill in the meantime." - Patrick's Guest (Matt) [00:34:39]
"If you plug in all of this compute... it could end up being 20 or 30% of the cost of compute by 2029." - Patrick's Guest (Matt) [00:38:19]
"In order to make more Permian natural gas, you also have to be incentivized to make more Permian oil, and those incentives didn't exist until Iran." - Patrick's Guest (Matt) [00:41:39]
Speakers & Credentials
Patrick O'Shaughnessy: Host of Invest Like the Best, facilitating a deep-dive conversation on energy markets and infrastructure constraints.
Matt: An energy market investor and subject matter expert with 20 years of experience, who recently completed an 18-month, well-by-well analysis of US natural gas supply constraints in the face of rising AI compute demand.
1. Executive Summary
The United States is currently sleepwalking into a historic and structural natural gas deficit that is modeled to manifest materially by the middle of 2028 [00:23:04].
A collision of two massive demand drivers is underway: legally contracted LNG export terminal expansions and an unmitigated surge in AI data center power requirements [00:05:09].
While aggregate gas resources exist underground, the physical limits of flow mechanics, well decline curves, and surface-level processing infrastructure cap maximum deliverability at roughly 132 Bcf/d [00:15:39].
Market participants are entirely complacent, trading off a flat forward curve near $3.50 that falsely assumes perpetual abundance dating back to the 2010 shale revolution [00:22:12].
This supply-demand mismatch will likely trigger convex and unbounded price spikes for natural gas, directly passing through to exorbitant electricity bills for the US consumer by 2029 [00:34:39].
Long-term mitigation requires an immediate, federally backed mobilization to construct large-scale nuclear reactors, alongside aggressive residential solar deployment to shield consumers from peak daytime power costs [00:30:30].
2. Chronological Table of Contents
[00:00:00] Introduction: The 2028 Natural Gas Deficit Thesis
[00:02:48] The Shale Revolution and the Rise of LNG Exports
[00:05:09] Calculating Maximum Deliverability vs. AI Compute Demand
[00:11:59] Stock vs. Flow: Why Resource Abundance is a Mirage
[00:17:11] Surface Infrastructure Constraints (Processing and Pipelines)
[00:20:31] The Forward Curve Complacency and Storage Cliffs
[00:25:16] Assessing the Winners: Upstream Producers and Solar
[00:29:30] The Nuclear Imperative: AP-1000s vs. SMRs
[00:33:41] Assessing the Losers: Consumers, Hyperscalers, and Turbine OEMs
[00:39:28] Debating the Bear Case: Permian Output and Battery Tech
[00:45:24] Strategic Mitigations and Global Geopolitical Fallout
3. Detailed Thematic Summary
The LNG Export Trajectory and Structural Baseline
The domestic natural gas landscape fundamentally shifted in 2010 when shale extraction technologies transitioned the US from a net importer to an era of absolute abundance [00:02:48].
This abundance catalyzed massive capital investments starting with Cheniere, taking the US from zero exports to roughly 15 billion cubic feet per day (Bcf/d) of nameplate capacity, representing 12 to 15% of the total US daily production base of 110 to 112 Bcf/d [00:03:22].
Current project pipelines, backed by tens of billions in project financing and international contracts, guarantee that US LNG exports will scale to 35 Bcf/d by the end of 2030 [00:03:50].
Simply halting these exports to protect domestic prices is legally and geopolitically unfeasible, as the US will soon control one-third of the global gas supply, making international allies under free trade agreements heavily reliant on these shipments [00:07:53].
The Collision of AI Compute and Flow Constraints
Even excluding AI compute demand, the maximum capacity the US can realistically add from mature basins like Appalachia, Haynesville, and the Permian is exactly 20 Bcf/d, which perfectly matches the scheduled 20 Bcf/d growth required by approved LNG exports [00:05:09].
By filtering for projects with signed Power Purchase Agreements and active interconnection agreements, a P50 base case probability model assigns 5 Bcf/d of highly credible incremental gas demand solely from AI data centers [00:06:26].
If unchecked, unmitigated extreme-case AI demand could require 12 to 15 Bcf/d of natural gas by the early 2030s, creating an absolute mathematical deficit in the supply chain [00:07:21].
Technologies touted as localized behind-the-meter solutions are highly gas-intensive; for example, Bloom Energy's 6 series fuel cells consume 150 million cubic feet of gas per day just to generate a single gigawatt of power [00:06:52].
The Illusion of Underground Abundance
While deep resources exist underground, production is tightly constrained by flow metrics, specifically well decline curves and the exhaustion of prime acreage in maturing corporate portfolios [00:14:10].
Stacking well performance parameters against the existing captured acreage of major operators yields a hard maximum theoretical deliverability of 128 to 132 Bcf/d for the entire country before accounting for midstream friction [00:15:39].
Extracting this gas faces severe midstream bottlenecks, starting with processing facilities required to strip natural gas liquids and sulfur to meet the strict 1,030 BTU specification required by regulated interstate pipelines [00:17:46].
Regulatory friction and environmental permitting regimes have effectively killed the expansion of the interstate pipeline network, evidenced by the Mountain Valley Pipeline being the sole major connectivity project completed in the last decade [00:19:47].
Market Complacency and the 2028 Storage Cliff
The market is suffering from a normalcy bias born of 15 years of abundance, keeping the forward curve anchored flat around $3.50 through the late 2020s and early 2030s [00:22:12].
This depressed pricing environment actively disincentivizes production growth today, leading major operators like EQT to shut in natural gas production rather than sell it at spot rates they believe will be more valuable later [00:21:27].
By mid-2028, the mathematical overlap of new data centers and LNG exports will force the grid to draw down from the nation's 4 TCF of working gas storage at unprecedented and unsustainable rates [00:22:47].
By 2029, aggregate storage will drop below all known historical evidence, creating a scenario where natural gas pricing becomes completely unbounded, potentially spiking to $10 or $20 to force demand destruction much like the $8 to $10 spikes seen during the Russia-Ukraine crisis or the 2014 Polar Vortex [00:23:04].
Winners, Losers, and Collateral Damage
Upstream producers holding tier-one rock are severely mispriced; Expand Energy controls 70% of remaining core Haynesville wells yet trades at a highly depressed 4x forward EBITDA on a complacent forward curve [00:25:23].
Because natural gas sets the marginal clearing price for power markets, free-fuel generation assets like utility-scale solar stand to harvest massive windfall margins with zero incremental capital expenditure [00:27:06].
The primary victim will be the US consumer, forced into a brutal political tradeoff between powering domestic AI, fulfilling international LNG contracts, or paying devastatingly high residential electricity bills [00:34:39].
Hyperscalers, currently modeling energy as roughly 10% of total compute costs based on faulty forward curves, could see energy expenditure aggressively scale to 20 or 30% of their total cost structure by 2029 [00:38:19].
Original Equipment Manufacturers like Caterpillar are walking into a trap, doubling their solar turbine manufacturing capacity right at the moment when exorbitant fuel costs will render new gas generation economically unviable to operate [00:36:02].
The Nuclear Imperative and Strategic Mitigations
To solve the structural deficit post-2032, the US government must utilize the $260 billion loan program office to fully fund two to four large-scale Westinghouse AP-1000 nuclear reactors to de-risk the supply chain and break the first-mover hesitation among utilities [00:46:22].
Small Modular Reactors (SMRs) are viewed cynically by grid experts as unscalable science experiments lacking the manufacturing footprint to deliver the tens of gigawatts required to plug the grid's impending shortfall [00:31:41].
To shield against intraday price volatility, state governments must urgently reinvigorate tax incentives for residential solar and battery installations, protecting consumers during peak 10:00 AM to 6:00 PM pricing windows [00:47:30].
Building a dedicated 1 to 2 Bcf/d pipeline from Canada's 0.7 TCF of trapped gas reserves directly into the US MISO power market is identified as a high-leverage geopolitical workaround to alleviate midwestern strain [00:48:57].
The Reference Vault
4. Data & Figures
Data Point
Value
Context
Timestamp
Current US LNG Export Capacity
15 Bcf/d
Nameplate daily export capacity currently serving global markets.
Stock vs. Flow Resource Constraints: A critical error in energy forecasting is confusing resource stock (the total volume of hydrocarbons trapped underground) with resource flow (the physics-bound maximum daily rate of extraction and processing). Plentiful acreage in the Permian basin means nothing if gathering pipes and processing plants lack the throughput capacity to bring it to market. The market assumes abundance because the absolute stock is high, completely blind to the reality that the flow is violently constrained by physics and infrastructure [00:15:39].
The Dispatch Curve Windfall: In deregulated power markets, the price of electricity is set by the marginal fuel required to meet the last megawatt of demand, which is almost always natural gas. If gas prices triple due to structural shortages, electricity prices triple alongside it. Assets operating on the lower end of the dispatch curve with zero fuel costs, such as utility-scale solar farms, experience total margin expansion. They sell power at gas-inflated prices while their input costs remain at zero, generating massive, capital-free windfalls for operators [00:27:06].
Bring Your Own Generation (BYOG): Driven by an intense need for "time to power," hyperscalers are bypassing traditional utility timelines by opting for localized, behind-the-meter generation solutions. This framework relies on deploying distributed assets like fuel cells directly at the data center site. The irony of BYOG is that nearly every proposed solution consumes natural gas, severely amplifying the macro deficit while attempting to solve a micro power-availability problem [00:34:43].
The P50 Probabilistic Demand Filtering: To separate authentic market evolution from tech-sector hype, infrastructure forecasters apply rigorous gating mechanisms. A P50 demand constraint only maps power requests that have achieved a 50% viability threshold, defined as possessing signed Power Purchase Agreements and active interconnection queues. This strips out the noise of aspirational press releases and zeroes in on the exact volume of physical gas that will be legally and contractually pulled from the grid [00:06:01].
The Forward Curve Complacency Trap: Energy procurement relies heavily on forward futures curves to forecast levelized costs of energy. Because the current curve is suppressed by short-term oversupply, hyperscalers and utility planners are underwriting massive, multi-decade capital expenditures based on the illusion of cheap, permanent gas. This complacency prevents preemptive hedging and physical supply contracting, guaranteeing a highly volatile "knife fight" for physical molecules once the math inevitably fails in 2028 [00:22:12].
The Convexity of Commodity Deficits: Unlike equity markets, physical commodity deficits are unbounded. When working gas storage breaches historical lows, demand destruction becomes the only rebalancing mechanism. Because utilities and hyperscalers possess highly inelastic demand curves, the price required to force them to shut off their systems is exceptionally high. Therefore, upside price risk transitions from linear to convex, threatening to bankrupt entities caught without physical hedges [00:24:09].
6. Anecdotes
The Mountain Valley Pipeline Regulatory Nightmare: The speaker invokes the grueling history of the Mountain Valley Pipeline to illustrate the near-impossibility of mitigating gas shortages via new midstream infrastructure. Connecting Appalachia to the Mid-Atlantic required navigating a labyrinth of hostile environmental permitting and regime changes, resulting in it being the only interstate pipe built in over a decade. The anecdote proves that even if gas prices spike, regulatory paralysis prevents a rapid infrastructure response to move molecules where they are needed [00:19:47].
The Iran Conflict and Permian Oil Incentives: To push back against the narrative that the Permian basin can simply scale gas production to solve the crisis, the speaker points out that Permian gas is largely a byproduct of oil drilling. Therefore, to get more gas, operators must be financially incentivized to drill for more oil. The speaker notes that oil was stagnating at $55 a barrel until geopolitical tensions with Iran surfaced, proving that natural gas deliverability is precariously tethered to global oil conflicts rather than domestic gas demand [00:41:39].
The Vogtle 3 and 4 Cost Overruns: To explain why US utilities are terrified to build large-scale nuclear power, the speaker points to the Plant Vogtle expansion in Georgia. Bereft of institutional memory after 30 years without building a reactor, the project suffered from astronomical cost overruns and took roughly 15 years to reach commercial service. This historical scar tissue explains why the US government must absorb the first-mover risk of new AP-1000 builds to entice risk-averse capital back to the nuclear sector [00:30:25].
The Memory Chip (DRAM) Cycle Parallel: When sharing this natural gas thesis with tech executives, a parallel was immediately drawn to the recent DRAM memory shortage. Just like silicon manufacturing, the energy supply chain moved slowly, starved of capital expansion for years due to low margins. The anecdote highlights how capital starvation in upstream infrastructure operates silently until it bites "slowly at first, and then all at once," crushing downstream tech margins unexpectedly [00:37:40].
The Gas Turbine Boom and Bust of the Early 2000s: To caution investors enamored with companies selling distributed power generators, the speaker references the early 2000s turbine craze. Capacity was drastically overbuilt in response to a momentary grid signal, leading to a prolonged industry bust. Today, OEMs are repeating history by doubling factory capacity just as the actual fuel required to run those turbines is about to become prohibitively expensive, effectively building hardware that customers won't be able to afford to turn on [00:35:34].
7. References & Recommendations
Upstream & Midstream Energy Companies
Expand Energy: Highlighted as the preeminent winner of the impending crisis, possessing 70% of core Haynesville inventory while trading at a severely depressed 4x EBITDA multiple [00:25:23].
Range Resources: Identified alongside Expand Energy as a high-quality Appalachian upstream operator poised for immense margin expansion as resource scarcity takes hold [00:26:17].
EQT: Referenced to demonstrate current market complacency; the company is actively shutting in natural gas production today because spot prices are too low to justify resource depletion [00:21:27].
Cheniere: Mentioned as the pioneer of the US LNG export revolution, catalyzing the massive outflow of domestic gas to international markets [00:03:14].
Power Generation & Technology Companies
Xcel Energy / NextEra Yield Co (XIFR) / Clearway Energy: Mentioned as prime beneficiaries of the dispatch curve windfall. As legacy solar and wind asset owners, they will mark their power purchase agreements to market at higher values without any incremental capital expenditure [00:27:31].
Bloom Energy: Used as a cautionary tale of behind-the-meter generation. Their solid oxide fuel cells require massive volumes of continuous natural gas, making their deployment at gigawatt scale physically impossible under the 2028 supply constraints [00:06:52].
Caterpillar (Solar Turbines) & Generac: Cited as potential losers for perfectly mistiming the capital cycle by aggressively doubling gas-turbine manufacturing capacity straight into an impending fuel scarcity cliff [00:36:02].
GE Vernova: Referenced as a provider of large-scale, highly efficient combined cycle power assets that represent the top-tier of the time-to-power hierarchy [00:05:34].
Nuclear Infrastructure Entities
Westinghouse (AP-1000): The specific large-scale nuclear reactor design identified as the only viable baseload solution to the 2030s energy deficit [00:29:38].
Cameco / Brookfield: The joint owners of Westinghouse (Cameco 49%, Brookfield 51%), positioned as deeply undervalued equities levered to the inevitable nuclear renaissance [00:32:59].
BWXT: The primary nuclear supplier for the US Navy, noted as a highly strategic company with vast dollar content exposure to future AP-1000 deployments [00:33:28].
SCANA: The utility company nearly bankrupted by the failed VC Summer nuclear project in South Carolina, serving as a cautionary tale for utility-led nuclear development [00:30:02].
Energy Basins & Power Markets
Appalachia (Marcellus & Utica): The primary gas-producing basin in the US, facing mature decline curves and a lack of remaining top-tier inventory [00:12:13].
Haynesville: Identified as a crucial swing basin for natural gas production where prime acreage is highly concentrated [00:12:13].
Permian & Eagle Ford: Oil-directed basins where natural gas is extracted as a byproduct, meaning their gas output is fundamentally reliant on high crude oil prices to incentivize drilling [00:12:18].
MISO, PJM, SPP/ERCOT: Major regional power markets highlighted as targets for wheeling Canadian natural gas imports to alleviate midwestern and southern grid strain [00:49:04].
Historical & Weather Events
Fukushima, Chernobyl, Three-Mile Island: The triad of historical nuclear disasters responsible for destroying the institutional memory and workforce required to build large-scale nuclear power in the West [00:30:17].
2014 Polar Vortex & Winter Storm Elliott (Dec 2022): Cited as weather anomalies that caused temporary, extreme natural gas price spikes, serving as a preview for the structural spikes expected in 2028 [00:23:46].
Russia-Ukraine Conflict: Referenced as the catalyst that established the US as the primary LNG supplier to Europe and pushed global spot gas prices to structural highs [00:23:37].
Miscellaneous Technologies
Lithium-Ion vs. Sodium Batteries: Mentioned when assessing technological disruption risks to the thesis; while sodium batteries are emerging, lithium-ion remains the primary economic deployment for fast-discharge grid stabilization [00:42:48].
Jul 21, 2026
Selling the India Growth Story with Amitabh Kant | 21 Jul 2026 | Political Economy with Shweta Punj
1. Executive Briefing TL;DR Targeting 9% GDP Growth via Investment Rate Expansion: India must raise its investment rate from 30% to 38–40% of GDP over the next 2.5–3 decades to sustain a 9% annual growth rate. Achieving this requires bridg…
Base Case AI Compute Demand
5 Bcf/d
Gas demand from highly credible P50 data center projects possessing PPAs and interconnection rights.