Stranded Minds: The Knowledge Crisis Hiding Inside the Energy Transition

Executive Summary

The global energy transition has focused intensively on stranded assets—the fossil fuel reserves and infrastructure that climate policy may render worthless. This analysis identifies a parallel risk that has received far less attention: knowledge stranding, the potential loss of engineering expertise accumulated by the oil and gas industry precisely when net-zero technologies require it most.

Carbon capture and storage, geothermal energy, and underground hydrogen storage all depend on subsurface engineering capabilities developed over a century of hydrocarbon extraction. Yet these capabilities are concentrated in an aging workforce—48% of Society of Petroleum Engineers members are over 55—while educational pipelines have collapsed, with US petroleum engineering enrollment down 83% from 2015 peaks. The 8-10 years required to develop autonomous subsurface expertise means that students entering programs today will not reach competence until the late 2030s, after most senior professionals have retired.

This demographic scissors crisis is compounded by an epistemological divide. The fossil fuel industry operates from a "stock" way of knowing suited to finite underground accumulations; the renewable sector operates from a "flux" paradigm suited to variable atmospheric flows. When stock knowledge is applied to net-zero applications without adequate translation, projects fail—as evidenced by major CCS installations like Gorgon (30-50% of design capacity) and Snøhvit (emergency redesign required).

A critical window exists between now and approximately 2028 to transfer knowledge at scale. After that, the expertise walks into retirement, and the window closes.

In 2007, the United States Department of Energy confronted an enemy that could not be named, sanctioned, or bombed into submission. A classified material called FOGBANK, essential for the proper functioning of W76 nuclear warheads, needed to be replaced as part of a routine life-extension program. The original production facility at the Y-12 National Security Complex in Oak Ridge, Tennessee, had been shuttered in 1989, but this presented no apparent obstacle. The formula existed. Detailed chemical specifications had been meticulously documented. The equipment could be rebuilt. What could possibly go wrong?

Everything, as it turned out. Engineers discovered that the replacement material failed quality tests repeatedly. Batches that should have been identical to the 1980s originals performed erratically. Years of investigation and hundreds of millions of dollars later, scientists finally identified the culprit: an ingredient originally classified as an impurity—a contaminant that quality control had worked to eliminate—turned out to be performance-critical.[1] The veteran engineers who had produced the original material knew this instinctively, but they had never written it down because it seemed like common sense to anyone who had worked the process. By 2007, those engineers had retired or passed away. The documentation recorded what to do but not why it works. The tacit knowledge that connected procedure to outcome had vanished.

The FOGBANK episode delayed the warhead refurbishment program by years and triggered a wave of institutional soul-searching about knowledge management in the nuclear weapons complex.[2] But the deeper lesson extends far beyond Oak Ridge. It reveals something unsettling about the fragility of technical knowledge in complex systems: even with complete documentation, unlimited budgets, and the full weight of federal urgency, a sophisticated organization can find itself essentially reinventing an existing formula because the human carriers of contextual understanding are gone.

Now imagine that FOGBANK is not a material but a skill set: the ability to drill a wellbore to 4,000 meters through fractured carbonate formations, manage the pressure regime to prevent blowouts, and design cement barriers that will maintain integrity for centuries. Imagine that the people who know how to do this—who have accumulated decades of experience managing the unpredictable physics of the deep subsurface—are all approaching retirement age at the same moment. And imagine that this skill set is precisely what humanity needs to seal carbon dioxide underground for a thousand years, to tap superhot geothermal reservoirs, and to store hydrogen in depleted gas fields as a buffer against renewable intermittency.

This is not a hypothetical scenario. This is the situation confronting the global energy sector in 2025. The concern here is not whether to transition away from fossil fuels—climate science has settled that question—but how to transition without losing capabilities that net-zero technologies require.

The Blind Spot in Climate Finance

Since the Carbon Tracker Initiative published its landmark "Unburnable Carbon" report in November 2011, the concept of stranded assets has successfully anchored climate finance discourse.[3] The original analysis was stark: using data from the Potsdam Institute for Climate Impact Research, Carbon Tracker demonstrated that the world's proven reserves of coal, oil, and gas, if burned, would emit approximately 2,795 gigatonnes of carbon dioxide—more than three times the carbon budget compatible with limiting warming to 2°C above pre-industrial levels. If climate policy were to be enforced rigorously, up to 80% of declared fossil fuel reserves owned by the world's largest listed companies would become unburnable. The top 100 coal companies and top 100 oil and gas companies held a combined market capitalization of $7.42 trillion as of February 2011.[4] A regulatory or technological shift that rendered most of these reserves worthless would constitute one of the largest value destructions in financial history.

The stranded assets framework has proven remarkably influential. Mark Carney, then Governor of the Bank of England and Chair of the Financial Stability Board, endorsed Carbon Tracker's analysis in a 2015 speech at Lloyd's of London, warning that assets rendered worthless by climate action could trigger systemic financial instability.[5] The Task Force on Climate-related Financial Disclosures, which Carney championed, has pushed thousands of companies to assess and disclose their exposure to transition risk. The divestment movement, inspired in part by Bill McKibben's 2012 Rolling Stone article popularizing Carbon Tracker's findings, has mobilized over $40 trillion in committed divestment from fossil fuels.[6]

Yet this framing, for all its influence, contains a fundamental oversight. The stranded assets framework focuses entirely on physical capital and extractable reserves—on steel, concrete, pipelines, and molecules locked underground. It treats fossil fuel companies as if they were merely boxes containing commodities, ignoring the cognitive infrastructure that built and operates those boxes.

Consider an offshore platform scheduled for decommissioning. On a balance sheet, it may have zero or negative book value. But the platform is also the embodiment of thirty years of accumulated knowledge about how to manage deepwater completions, how to maintain well integrity in corrosive environments, how to respond when pressures deviate from predictions. That knowledge cannot be reduced to a line item. It walks off the platform when the crew disperses. And unlike steel, which can be melted and recast, tacit expertise that leaves the industry is rarely recovered.

This is knowledge stranding—the risk that engineering expertise, operational experience, and institutional memory accumulated by the fossil fuel industry may be disrupted or permanently lost due to industry decline, while a significant portion of this knowledge is precisely what net-zero technologies require.

Knowledge stranding is not the same as the "skills gap" that appears in workforce policy discussions. Skills gap framing assumes that training programs can bridge any deficit given sufficient funding. Knowledge stranding recognizes that some forms of expertise cannot be rapidly manufactured because they depend on years of experiential learning in specific operational contexts. The 8-10 years that industry practitioners estimate it takes for a new petroleum engineering graduate to reach autonomous decision-making capacity is not an artifact of inadequate curricula—it reflects the irreducible time needed to develop judgment through exposure to the variability of real-world geological systems.[7]

Knowledge stranding also differs from "brain drain," the migration of skilled workers from one region to another. Brain drain is spatial; the knowledge still exists somewhere, potentially recoverable if incentives change. Knowledge stranding involves generational rupture compounded by industry-wide contraction. When an entire sector shrinks simultaneously across all major economies, there is nowhere for the knowledge to migrate. It simply dissipates.

Knowledge stranding is the core phenomenon this analysis identifies. The sections that follow examine its two primary drivers: an epistemological mechanism—the fundamental mismatch between stock and flux ways of knowing that prevents automatic knowledge transfer—and a temporal accelerator—the demographic scissors crisis that compresses the window for any transfer to occur.

Two Ways of Knowing Energy

The difficulty of transferring knowledge from the fossil fuel industry to the renewable energy sector runs deeper than mismatched terminology or training deficits. It stems from something more fundamental: these industries have developed entirely different ways of seeing.

Every industry develops characteristic frameworks for interpreting evidence, managing uncertainty, and making decisions under conditions of incomplete information. A surgeon's way of knowing differs from an airline pilot's not because of educational differences but because the objects of their work—human bodies, aircraft—present fundamentally different kinds of problems and demand different modes of reasoning. The energy sector is no different. The nature of the resource—whether it sits underground waiting to be extracted or flows through the atmosphere waiting to be captured—shapes everything from risk assessment to financial valuation to regulatory philosophy.

The fossil fuel industry operates from what might be called a stock way of knowing. Hydrocarbons are finite accumulations deposited over geological time, invisible to direct observation, confined under pressure in porous rock formations. The central anxiety is running out, drilling dry holes, overestimating what lies underground. Success means extraction—getting value out of the ground before competitors do, before prices fall, before the reservoir pressure declines. The relevant timescale is geological: formations that took millions of years to fill, production profiles that span decades, wells that must maintain integrity for centuries after abandonment.

The renewable energy industry operates from a flux way of knowing. Solar radiation and wind are continuous flows, visible and measurable at the surface, variable over meteorological timescales, effectively inexhaustible but fundamentally intermittent. The central anxiety is not depletion but variability—forecasting errors, grid instability when the wind stops or clouds pass. Success means integration—matching variable supply to variable demand, coordinating across vast distances, converting ambient energy into dispatchable power. The relevant timescale is meteorological: weather patterns that shift hourly, demand cycles that peak seasonally, capacity factors that must be characterized statistically over years.

These differences manifest in concrete engineering practices. In the oil and gas world, the ultimate evidence is a well—a physical puncture through kilometers of rock that yields (or fails to yield) hydrocarbons. Seismic surveys, log analyses, and reservoir simulations are all approximations, useful for narrowing the search space but never definitive until steel enters the ground. The well test is the moment of truth. In the renewable world, the ultimate evidence is a statistical distribution—years of wind speed measurements extrapolated through power curves to yield probability distributions of annual energy production. No single measurement settles the question because the resource is inherently variable. Truth is not revealed by a puncture but constructed from aggregated observations.

The fossil fuel industry has learned, through a century of painful experience, to expect the unexpected from the subsurface. Every reservoir model is a simplification. Surprises are not exceptions but expectations. This hard-won humility is institutionalized in practices like history matching, where simulation models are repeatedly adjusted to reproduce observed production data, creating an iterative dialogue between prediction and performance.[8] A reservoir engineer who expects a model to work the first time has not been in the industry long. The mature professional knows that models are hypotheses, that heterogeneity is the rule, and that the earth keeps secrets until forced to reveal them through drilling and production.

This extends to an acceptance of fat-tailed distributions—the recognition that low-probability events (blowouts, unexpected pressure compartments, reservoir connectivity different from predictions) are not merely theoretical possibilities but lived realities on geological timescales. The industry designs for tail risks in ways that would seem excessive to someone accustomed to normal distributions. And it names its disasters: Piper Alpha, Deepwater Horizon, Texas City. These are not merely historical events but reference points in an ongoing conversation about what can go wrong and why. Root cause analysis is a cultural practice, not just a regulatory requirement. Young engineers learn the names of the disasters before they learn the details, absorbing a tradition of failure-consciousness that shapes how they approach every operation.[9]

The renewable energy industry has developed its own sophisticated approach to uncertainty—but it is a fundamentally different kind of sophistication. The wind industry treats variability not as a defect to be eliminated but as an inherent feature to be managed statistically. The IEC 61400 standards for wind turbine design specify target failure probabilities on the order of 10⁻⁴ per year for certain structural failure modes, rather than deterministic safety factors.[10] Uncertainty becomes a parameter to trade off against cost: higher reliability is achievable but expensive, and the optimal point depends on economic considerations.

A distinctive feature of this approach is what might be called learning curve faith—the conviction that renewable technology costs will continue declining along predictable trajectories. Wright's Law, which observes that manufacturing costs typically decline 20-24% for every doubling of cumulative production, is treated as something close to a physical law in solar industry projections.[11] The fossil fuel industry has no comparable expectation; extraction costs tend to rise over time as easy resources are depleted and production moves to more challenging environments. This directionality shapes investment decisions, policy designs, and competitive dynamics in ways that distinguish flux industries from stock industries.

Neither way of knowing is wrong. Each is a valid response to the physical properties of the resource each industry handles. But when applied to the wrong context, each can produce systematic errors.

The Dangerous Illusion of Shared Language

Despite their different foundations, the oil and gas world and the renewable energy world share some surface vocabulary that can create an illusion of mutual understanding. Both industries use probabilistic metrics like P90 (the value that has a 90% probability of being exceeded). Both employ independent technical reviewers to validate project parameters. Both rely on complex financial structures that allocate risk among equity, debt, and offtake counterparties.

But beneath this apparent convergence lie deep differences in what these practices mean.

Consider P90. In the oil and gas context, P90 refers to a conservative estimate of hydrocarbon volumes: there is a 90% probability that actual recoverable resources exceed this value, based on current geological understanding. The P90 estimate is a lower bound on a physical stock that exists underground regardless of human activity. It waits. If the model is wrong—if connectivity is worse than expected, if water encroachment is faster—the actual volumes may fall below P90, but the methodology treats this as a modeling failure to be corrected through history matching and model revision. The implicit assumption is that more data will reduce uncertainty over time.

In the wind and solar context, P90 refers to a percentile of a statistical distribution of energy production: there is a 90% probability that annual generation will exceed this value in any given year. This is not a physical stock but a description of variability. The P90 estimate does not bound a quantity that exists but characterizes the lower tail of an ongoing stochastic process. The underlying assumption is stationarity—the expectation that future weather patterns will resemble past patterns in their statistical properties.[12]

When geologists and bankers sit in the same room discussing a carbon capture project and both use the phrase "P90," they may believe they are speaking the same language. They are not. The geologist means "conservative lower bound on physical capacity given current data." The banker means "bankable base case for debt sizing." The geologist expects that injection performance may require adaptive management as the reservoir responds to pressure buildup. The banker expects that the number in the spreadsheet corresponds to the revenue line in the financial model.

This semantic slippage creates systematic optimism bias: parameters that geologists consider conservative become planning cases that financiers treat as near-certainties. P90 functions as what sociologists of science call a "boundary object"—a concept that different communities can use together without agreeing on its precise meaning.[13] Boundary objects enable cooperation across disciplinary divides, but they can also obscure disagreements that matter. In the context of project finance for carbon capture, the P90 convergence allows geologists and bankers to reach final investment decisions without confronting the fundamental question: what happens if the subsurface does not behave as modeled?

The stock way of knowing has answers to this question, grounded in decades of experience with history matching and adaptive reservoir management. The flux way of knowing has different answers, grounded in statistical inference from historical data. Neither set of answers maps cleanly onto the geological time horizons and long-term liability regimes that carbon capture projects require.

The Scissors Crisis

The fossil fuel industry is experiencing a demographic event without precedent in its history—a simultaneous aging of the workforce, collapse of the educational pipeline, and cultural delegitimization that together threaten to sever the intergenerational transmission of knowledge.

The Society of Petroleum Engineers (SPE) is the world's largest professional association for petroleum engineers, geoscientists, and related technical specialists, with approximately 143,000 members worldwide. Its membership provides a window into the state of the global oil and gas technical workforce—an imperfect window, since membership includes retirees who maintain their affiliation for journals and networking, and skews toward senior and internationally mobile professionals. Nonetheless, the picture is revealing. According to SPE's 2023 data, approximately 47-48% of professional members are over 55 years old.[14] Given that the typical retirement age in the industry ranges from 55 to 62, a substantial fraction of the credentialed petroleum technical workforce is within a decade of leaving.

This concentration of expertise in the pre-retirement cohort did not arise suddenly. It reflects the compound effects of two previous hiring freezes—one following the oil price collapse of the mid-1980s, the other following the 2014-2016 downturn—which created a "missing generation" of mid-career professionals. A healthy industry age structure would resemble a pyramid or cylinder, with roughly equal numbers of professionals in each age cohort. The oil and gas industry instead exhibits a pronounced bulge in the 55+ age bracket and a much thinner population of mid-career professionals who would normally be positioned to absorb knowledge from senior colleagues and transmit it to juniors.[15]

The 2014 downturn was particularly damaging because it occurred just as the post-1980s generation should have been entering senior roles. McKinsey research indicates that 42% of workers who left oil and gas companies during the 2015-2020 period left the energy sector altogether, taking their specialized knowledge to technology, finance, or entirely unrelated industries.[16]

Regional variations exist but do not fundamentally alter the picture. The Middle East retains significant operational capacity but faces its own demographic transitions as national oil companies prepare for diversification. China has expanded petroleum engineering education and launched major CCS initiatives—Sinopec's Qilu-Shengli project, the country's largest operational CCS facility, began injection in 2022—but faces the same 8-10 year time-to-autonomy constraint; rapid enrollment scaling cannot compress experiential learning, and early projects rely partly on expertise transferred from international oil service companies.[17] The global nature of the oil and gas labor market means that regional surpluses in one area have historically offset deficits elsewhere—but industry-wide contraction removes this buffer.

Meanwhile, the educational pipeline that would replenish the workforce has collapsed. Enrollment declines at major petroleum engineering programs—whether measured by declared majors, new enrollments, or graduates, depending on university reporting—have been dramatic. Texas Tech University, one of the largest programs in the United States, reports an 88% decline from its mid-2010s peak. The Colorado School of Mines reports 87.7%. The University of Oklahoma and Louisiana State University—institutions deeply embedded in their states' oil and gas economies—report declines exceeding 85%.[18] The national average across US programs exceeds 70% from mid-2010s peaks.

These are not statistical fluctuations. They represent structural collapse. At the graduate level, research funding for petroleum-related disciplines has declined by approximately 70% over the past decade. Faculty positions have been eliminated or converted to other specialties. Imperial College London suspended its petroleum engineering program. The University of Calgary paused its undergraduate program amid declining student interest.[19]

The relationship between oil prices and petroleum engineering enrollment has historically been robust, with enrollment changes lagging price movements by approximately 2.5 years as prospective students respond to perceived career prospects. But this relationship broke down after 2015. Despite significant oil price recoveries in 2017-2018 and again in 2021-2022, enrollment did not rebound. Something structural changed: the perception of oil and gas as a career path shifted from cyclical risk to terminal decline.

Deloitte's 2021 survey of young professionals found that 44% would not consider working for an oil and gas company, with climate concerns the most commonly cited reason.[20] A significant cohort of technically capable students who might previously have entered petroleum engineering are instead pursuing environmental engineering, data science, renewable energy technology—or leaving engineering entirely.

The significance of these trends can only be understood in light of the time required to develop expertise in subsurface engineering. The industry consensus on time-to-autonomy—approximately 8-10 years according to practitioners and professional associations—reflects not credentialing requirements but the irreducible time needed for experiential calibration.[21] A student entering a petroleum engineering program in 2025 will graduate in 2029. Add 8-10 years for autonomous capability, and we reach 2037-2039. By that point, the cohort of professionals currently over 55 will have largely retired. The mentors who could have supervised that student's development through the full learning period will be gone.

This temporal mismatch between supply and demand creates what might be called a scissors crisis. On one blade, the supply of experienced subsurface professionals is declining as retirements accelerate. On the other blade, demand for subsurface expertise is poised to increase as carbon capture, geothermal, and hydrogen storage projects move from demonstration to commercial scale. The two blades cross in the late 2020s. After that crossing point, projects will bid against each other for a shrinking pool of qualified professionals, driving costs higher and forcing compromises on technical quality.[22]

Between now and roughly 2028, there exists a unique window. The senior cohort—those with 25-40 years of experience, who carry the most concentrated deposits of tacit knowledge—remains active in the workforce. Many are contemplating retirement but have not yet left. At the same time, net-zero technologies are transitioning from pilot phase to early commercial deployment. If this window is missed, the expertise walks out the door into retirement, and net-zero projects face structural knowledge deficits that no amount of post-hoc training can quickly remedy.

What Happens When Knowledge Strands

The consequences of knowledge stranding are already visible in the performance of early carbon capture projects.

The Gorgon project in Western Australia—the world's largest dedicated CCS project, operated by Chevron—has consistently underperformed its design capacity since beginning injection in 2019. The capture rate has fluctuated between 30% and 68%, well below the 80%+ that was promised at final investment decision in 2009.[23] The causes include pressure management challenges in the target formation, sand production clogging water disposal wells, and equipment corrosion requiring repeated shutdowns. Each of these issues involves the kinds of geological surprises and operational challenges that experienced reservoir and completions engineers are trained to anticipate and manage.

Norway's Snøhvit project, often cited alongside Sleipner as evidence that CCS works, nearly failed in its second year of operation when injection pressure spiked unexpectedly. The target formation, a fluvial sandstone unit called Tubåen, turned out to be compartmentalized by low-permeability barriers that the pre-injection geological model had not identified. After 18 months, the project had to be restructured to inject into a shallower formation instead.[24] The emergency intervention succeeded—Snøhvit continues operating—but only because Equinor had experienced subsurface professionals who could diagnose the problem, redesign the injection strategy, and execute an expensive workover on an offshore well.

These cases might seem to undercut the argument for preserving oil and gas expertise. After all, Chevron and Equinor are among the most technically sophisticated hydrocarbon companies on Earth. If their experts struggled with CCS, what is the value of that expertise?

The answer lies in understanding what these failures actually demonstrate. Gorgon's struggles do not show that oil and gas expertise is irrelevant to CCS—they show that even deep expertise requires adaptation when applied to a fundamentally different problem. Extraction and injection obey different physics. The same reservoir that yielded hydrocarbons predictably may behave unpredictably when asked to accept CO₂ at scale. The pressure dynamics reverse. The geochemistry changes. The liability horizon extends from decades to centuries.

The experts at Gorgon knew how to extract. The question is whether anyone had fully learned how to inject at industrial scale—and whether the translation from one to the other was made explicit or merely assumed. If masters struggle during the rewriting of the textbook, novices without masters will struggle far more. Snøhvit survived because experienced professionals were available to improvise a solution. What happens when they are not?

A 2022 Institute for Energy Economics and Financial Analysis review of major CCS projects found that the majority had failed to meet their stated capture and storage targets.[25] This pattern is consistent with systematic translation failure: knowledge from oil and gas contexts being applied without adequate adaptation to the distinct challenges of long-term CO₂ storage.

The costs extend beyond individual project economics. If early CCS projects continue to disappoint—if they acquire a reputation for overpromising and underdelivering—the credibility of the technology as a climate solution will suffer. Investors will demand higher risk premiums. Policymakers will lose patience. The window for CCS to contribute meaningfully to mid-century decarbonization could close not because the technology is fundamentally flawed but because the knowledge to implement it properly was lost in the transition from fossil fuel extraction to carbon storage.

Which Knowledge Matters Most

Not all oil and gas knowledge is equally relevant to net-zero technologies, and not all relevant knowledge is equally at risk of stranding.

Well integrity represents perhaps the most critical domain. The ability to design, construct, and maintain wells that contain fluids under pressure over long time periods is foundational to carbon capture and storage, underground hydrogen storage, and geothermal energy. A CCS project that cannot guarantee the integrity of its injection wells cannot guarantee the permanence of its storage—and permanence is the entire point. Well integrity knowledge encompasses materials science (cement behavior under CO₂ exposure, steel corrosion in acidic environments), mechanical engineering (casing design, stress analysis), and operational judgment (pressure testing, recognizing warning signs, remediation techniques). This knowledge exists partly in standards documents but resides more fundamentally in the judgment of experienced completions engineers who have seen wells fail and learned to recognize the precursors.[26]

Reservoir modeling is similarly critical. The simulation of fluid flow through porous media—how injected CO₂ will migrate through an aquifer, how pressure will build and dissipate, how geochemical reactions will alter permeability—requires expertise developed over decades of oil and gas applications. The software tools (Eclipse, CMG, Petrel) are transferable; the interpretive judgment about when to trust a model and when to question it is not.

High-pressure gas handling becomes critical as hydrogen emerges as an energy carrier. The oil and gas industry has extensive experience with pressurized hydrocarbon gases, but hydrogen presents distinct challenges: smaller molecular size enabling leakage through seals that contain methane, embrittlement of certain steel alloys, wider flammability limits. Adapting gas-handling knowledge to hydrogen requires not just learning new parameters but understanding where old intuitions remain valid and where they mislead.[27]

Megaproject management—the ability to coordinate tens of billions of dollars of investment across multiple contractors, regulatory jurisdictions, and technical disciplines—is perhaps the least appreciated domain at risk. Large energy projects fail more often from organizational and managerial causes than from technical ones. The oil and gas industry has developed project management capabilities through painful experience with cost overruns and delays. These capabilities are embodied in processes, but more fundamentally in the tacit knowledge of experienced project directors who have learned to recognize warning signs, manage contractor relationships, and maintain schedule discipline.

Can Technology Fill the Gap?

Digital twins and machine learning are sometimes proposed as solutions to tacit knowledge loss—the idea that sufficient data can substitute for experienced judgment. This view misunderstands what tacit knowledge does.

A digital twin can replicate known physics; it cannot anticipate novel failure modes that fall outside training data. Machine learning excels at pattern recognition within distributions it has seen; it struggles with the fat-tailed surprises that characterize subsurface operations. The value of experienced professionals lies precisely in their encounters with anomalies that have not yet been digitized.[28]

Consider what a digital twin of a CCS injection site would require: accurate representations of geological heterogeneity at sub-seismic resolution, validated models of CO₂-brine-rock geochemistry over century timescales, and training data from enough injection operations to teach the system what going wrong looks like. We do not have this data. The CCS industry is too young to have accumulated the failure library that machine learning would need.

AI systems could substitute for experienced judgment only after accumulating decades of operational CCS data, validated physics models across diverse geological settings, and governance frameworks for algorithmic accountability. Until then, technology augments expertise; it does not replace it.

What This Is Not

Several clarifications are important.

This is not an argument that the energy transition should be slowed to accommodate knowledge transfer. The climate imperative is clear and urgent. The pace of decarbonization is, if anything, too slow. The argument here concerns not the direction of transition but its quality—how to transition without avoidable losses of capability that increase costs, delay deployment, and create safety risks.

This is not a claim that oil and gas professionals are uniquely valuable and should be privileged over other workers affected by energy transition. All displaced workers deserve support, and just transition frameworks address this moral obligation appropriately. The claim here is narrower: a specific body of technical knowledge is at risk of loss, and this loss would damage the net-zero project itself, not merely the welfare of the knowledge-holders.

This is not a claim that knowledge stranding is the only or even the primary barrier to net-zero deployment. Capital availability, policy stability, permitting delays, supply chain constraints, and public acceptance all matter enormously. Knowledge stranding is a contributor to these challenges—projects fail and budgets overrun partly due to knowledge deficits—but not a substitute explanation for them.

Who Owns This Problem?

The knowledge stranding problem does not have a single owner. It sits at the intersection of educational policy (who trains subsurface engineers and for what careers), industrial strategy (how companies manage knowledge assets during transition), financial regulation (how project finance structures price epistemic risk), and workforce policy (how just transition frameworks account for knowledge preservation alongside worker welfare). No single actor created the problem; no single actor can solve it. But the absence of clear ownership is itself part of the problem—it means that the window can close while everyone waits for someone else to act.

The Window Is Still Open

The knowledge that humanity has accumulated about interacting safely with the deep subsurface is one of the most sophisticated bodies of technical expertise ever developed. It took a century to build. It cannot be quickly rebuilt.

The senior generation of subsurface professionals—those who carry the densest concentrations of hard-won judgment—remains active, for now. Many are approaching retirement but have not yet left. Net-zero technologies that depend on their knowledge are scaling up, creating demand for expertise that did not exist a decade ago.

This conjunction creates a window for large-scale knowledge transfer that will not recur. If the senior cohort can be engaged in net-zero projects—whether through direct employment, consulting arrangements, or structured mentorship—their expertise can be transmitted to a new generation oriented toward decarbonization applications. If this window closes, the expertise dissipates into retirement, and projects that could have succeeded will fail for want of knowledge that once existed but was allowed to strand.

The clock is running. The question is whether we act.

References

[1] U.S. Government Accountability Office. "Nuclear Weapons: NNSA and DOD Need to More Effectively Manage the Stockpile Life Extension Program." GAO-09-385. Washington, DC: GAO, March 2009; Stockton, Peter, and Ingrid Drake. "The FOGBANK Fiasco." Bulletin of the Atomic Scientists, March 2014.

[2] Polanyi, Michael. The Tacit Dimension. Chicago: University of Chicago Press, 1966; Collins, Harry. Tacit and Explicit Knowledge. Chicago: University of Chicago Press, 2010.

[3] Carbon Tracker Initiative. Unburnable Carbon: Are the World's Financial Markets Carrying a Carbon Bubble? London: Carbon Tracker, November 2011.

[4] Carbon Tracker Initiative. Unburnable Carbon: Are the World's Financial Markets Carrying a Carbon Bubble? London: Carbon Tracker, November 2011. The $7.42 trillion figure represents combined market capitalization of top 100 coal and top 100 oil and gas companies as of February 2011.

[5] Carney, Mark. "Breaking the Tragedy of the Horizon—Climate Change and Financial Stability." Speech at Lloyd's of London, September 29, 2015.

[6] McKibben, Bill. "Global Warming's Terrifying New Math." Rolling Stone, July 19, 2012; Fossil Free. "Divestment Commitments." Accessed 2024.

[7] IOGCC Blue Ribbon Task Force. Petroleum Professionals: A Crisis in the Making. Oklahoma City: Interstate Oil and Gas Compact Commission, 2004.

[8] Oliver, Dean S., and Yan Chen. "Recent Progress on Reservoir History Matching: A Review." Computational Geosciences 15, no. 1 (2011): 185–221.

[9] Cullen, Lord. The Public Inquiry into the Piper Alpha Disaster. London: HMSO, 1990; National Commission on the BP Deepwater Horizon Oil Spill and Offshore Drilling. Deep Water: The Gulf Oil Disaster and the Future of Offshore Drilling. Washington, DC: GPO, 2011.

[10] International Electrotechnical Commission. IEC 61400-1:2019, Wind Energy Generation Systems—Part 1: Design Requirements. Geneva: IEC, 2019.

[11] Wright, T. P. "Factors Affecting the Cost of Airplanes." Journal of the Aeronautical Sciences 3, no. 4 (1936): 122–128; Farmer, J. Doyne, and François Lafond. "How Predictable Is Technological Progress?" Research Policy 45, no. 3 (2016): 647–665.

[12] AWS Truepower. Wind Resource Assessment: A Practical Guide to Developing a Wind Project. Albany, NY: AWS Truepower, 2014; International Electrotechnical Commission. IEC 61400-12-1:2017, Power Performance Measurements of Electricity Producing Wind Turbines. Geneva: IEC, 2017.

[13] Star, Susan Leigh, and James R. Griesemer. "Institutional Ecology, 'Translations' and Boundary Objects: Amateurs and Professionals in Berkeley's Museum of Vertebrate Zoology, 1907-39." Social Studies of Science 19, no. 3 (1989): 387–420.

[14] Society of Petroleum Engineers. SPE Membership Demographics: 2023 Year-End Summary. Richardson, TX: SPE, 2024.

[15] McKinsey & Company. "Talent Squeeze: Planning for the Energy Sector's Talent Transition." McKinsey Energy Insights, 2022.

[16] McKinsey & Company. "Talent Squeeze: Planning for the Energy Sector's Talent Transition." McKinsey Energy Insights, 2022.

[17] Global CCS Institute. Global Status of CCS 2023. Melbourne: GCCSI, 2023; Sinopec. "Qilu-Shengli CCUS Project Begins Full Operation." Press release, August 2022.

[18] American Society for Engineering Education (ASEE). Engineering enrollment data, various years; JPT Staff. "Petroleum Engineering Enrollment Decline Continues." Journal of Petroleum Technology, March 2023.

[19] "Imperial College to Suspend Petroleum Engineering Programme." Energy Voice, February 2022; "UCalgary Pauses Undergraduate Petroleum Engineering Admissions." CBC News, April 2023.

[20] Deloitte. 2021 Deloitte Global Millennial and Gen Z Survey: Energy Industry Perspectives. London: Deloitte, 2021.

[21] IOGCC Blue Ribbon Task Force. Petroleum Professionals: A Crisis in the Making. Oklahoma City: Interstate Oil and Gas Compact Commission, 2004. Internal career development frameworks at Shell, BP, and ExxonMobil consistently specify 8–12 years before autonomous decision-making authority for high-consequence operations.

[22] Robert Gordon University Energy Transition Institute. Powering Up the Workforce: The Future of the UK Offshore Energy Workforce. Aberdeen: RGU, 2023; International Energy Agency. Net Zero by 2050: A Roadmap for the Global Energy Sector. Paris: IEA, 2021.

[23] Institute for Energy Economics and Financial Analysis. "Gorgon Carbon Capture Facility Hits New Lows in 2023-24." IEEFA Report, September 2024; Chevron Australia. Gorgon Project: Environmental Performance Report 2023. Perth: Chevron Australia, 2024.

[24] Institute for Energy Economics and Financial Analysis. "Norway's Sleipner and Snøhvit CCS: Industry Models or Cautionary Tales?" IEEFA Report, February 2023; Equinor. Snøhvit LNG: Lessons Learned from CO₂ Storage Operations. Stavanger: Equinor, 2020.

[25] Institute for Energy Economics and Financial Analysis. "The Carbon Capture Crux: Lessons Learned." IEEFA Report, September 2022.

[26] Zoback, M. D., and S. M. Gorelick. "Earthquake Triggering and Large-Scale Geologic Storage of Carbon Dioxide." Proceedings of the National Academy of Sciences 109, no. 26 (2012): 10164–10168; Society of Petroleum Engineers. Well Integrity for CO₂ Injection Wells. Technical Report. Richardson, TX: SPE, 2020.

[27] DNV. Hydrogen Forecast to 2050: Energy Transition Outlook. Oslo: DNV, 2022.

[28] Marcus, Gary, and Ernest Davis. Rebooting AI: Building Artificial Intelligence We Can Trust. New York: Pantheon, 2019; Caers, Jef. Modeling Uncertainty in the Earth Sciences. Chichester: Wiley-Blackwell, 2011.

Copyright

© 2026 Alex Yang Liu. All rights reserved.
Publisher: Terawatt Times Institute | ISSN: 3070-0108
Version: 1.0 | Date: January 2026
Citation: Liu, A.Y. "Stranded Minds: The Knowledge Crisis Hiding Inside the Energy Transition." Terawatt Times, January 2026. ISSN 3070-0108.

License

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License (CC BY-NC-ND 4.0).
You are free to share and distribute this work with proper attribution, for non-commercial purposes, without modification. For permissions beyond this license, contact the author.

Disclaimer

This paper represents the author's independent analysis and does not necessarily represent the views of Terawatt Times Institute or any other organization. While every effort has been made to ensure accuracy, the author and publisher assume no responsibility for errors or omissions. Data and projections are based on publicly available sources as of January 2026.