The Hidden Dangers in Canada’s Oil and Gas Ambitions

Carney, Hodgson, Smith led astray by misguided energy demand forecasting

Danielle Smith thinks global demand for oil will grow to 2050, perhaps beyond, and the Alberta industry will be viable for a hundred years. She has publicly said many times that she agrees with Saudi Arabia’s view of oil’s future. “Simply put, there is no ‘peak in oil demand’ on the horizon,” OPEC Secretary General Haitham Al Ghais told the 2025 Global Energy Show. He added that OPEC expects mid-century global oil consumption to rise from its current 103 million barrels per day to over 120 million barrels per day.

The Alberta Premier agreed with him. And Prime Minister Mark Carney and federal Energy Minister Tim Hodgson agree with Smith.

How do we know this? In September, I asked the Department of Natural Resources which modelling studies underpinned Hodgson’s bullish public comments about the future demand for Canadian crude oil. This is the Department’s response (I also asked about the future of LNG and received a similar reply).:

Natural Resources Canada regularly consults a wide range of reputable international sources to inform its analysis of crude oil demand. In addition to the International Energy Agency’s Stated Policy Scenarios (STEPS), NRCan reviews forecasts and scenarios developed from the US Energy Information Administration, the Organization of the Petroleum Exporting Countries (OPEC), the Institute of Energy Economics (Japan), as well as industry outlooks from companies such as Exxon Mobil, Shell, and BP.

These organizations are all known for their bullish oil demand outlooks. With the exception of the Japanese Institute, they are also all major oil producing companies or nations. Despite Al Ghais’ claims that OPEC’s forecasts are “based on robust analysis of the data, and not on ideology,” we can be forgiven for assuming some home team bias from the oil exporters. In fact, as we shall see, ideology plays a prominent role in OPEC’s view of oil’s future.

These views matter because Prime Minister Carney is leading the charge to diversify Canadian oil and gas customers, principally to Asia, and to a lesser extent, Europe. Ottawa has already included LNG Canada’s Phase 2 expansion (Phase 1 cost $40 billion) among the first five “nation-building” projects designated for fast tracking under the One Canadian Economy Act. The recent federal budget mentioned Canada becoming an energy superpower in “conventional energy” no less than five times. This will require at least one major pipeline to the West Coast. In her recent mandate letter to her energy minister, though, Smith tasked him with promoting five pipelines.

More shipping capacity, in turn, requires more oil sands supply. In the letter to Jean, Smith repeated her ambition for Alberta to double its oil production from 4 million barrels per day to 8 million barrels per day by 2035. Pathways Alliance, the oil sands company trade group, has said that the cost to decarbonize Alberta’s emissions-intense heavy crude is $75 billion, of which industry expects governments will pay $50 billion. If production doubles, does that boost decarbonizing costs to $150 billion, and the public subsidy to $100 billion?

Even a quick calculation shows the potential cost to the public purse could be hundreds of billions. That doesn’t include government support for expanded LNG production. Carney was quoted in an interview saying that he would like to see Canadian LNG supply rise from 19 million tonnes per annum (mtpa) to 100 million or 150 million in the near future.

This is why the arcane matter of oil and gas demand modelling is so important to Canadian taxpayers. If the industry and Canadian governments get it wrong, taxpayers could drown in the extra debt for decades. Who will pay for stranded hydrocarbon assets? Alberta already has a massive unfunded liability of close to $300 billion. Double production and exports, and the numbers become mind numbingly large.

The danger posed by this risk becomes alarming when we realize that many oil and gas demand modelling studies see a bearish future for oil and gas demand. The IEA’s APS scenario, for example, sees oil demand dropping to just 57 million barrels per day by 2050, almost half of current supply.

Can Canadian oil compete in a rapidly declining market? The Canadian Energy Regulator doesn’t think so. Low prices caused by falling demand shrink oil output considerably by 2050 in Canada’s Energy Future annual report.

Imagine this scenario. Global oil and gas demand falls instead of rises after 2030, which is about the time that new Canadian hydrocarbon infrastructure would come online. That infrastructure becomes a white elephant. Instead of tens of billions in new government revenue, taxpayers are on the hook for hundreds of billions in subsidies and clean-up costs. Future generations are impoverished because of bullish, imprudent spending by today’s politicians.

That’s why energy demand modelling matters. Get it wrong and the consequences are mind-bogglingly awful for Canada.

Energi Media’s reporting and analysis suggest that Canada has got it wrong. In this essay, I will show you why we think that.

How Energy Demand Models Work (and Why Assumptions Rule)

Oil and gas demand forecasts typically cover two to five years ahead. Past that horizon, analysts move from forecasting to scenario modelling. Every model rests on assumptions: Are electric vehicles cost-competitive with internal combustion cars? If not now, will they be by 2050? Will adoption come faster in China than in Africa? Each answer becomes a parameter. When economists build bearish, baseline, and bullish scenarios, those parameters shift again. With hundreds of variables feeding their equations, modellers can conjure a multitude of possible futures. This is one reason why projections for 2050 diverge so sharply among agencies and oil companies.

These models are not neutral; they reveal the modeller’s worldview. Some privilege continuity; others expect disruption. For decades, OPEC, ExxonMobil, the U.S. Energy Information Administration, and other major forecasters have served as custodians of continuity. Their models rest on the premise that hydrocarbons dominate well into mid‑century, that transitions unfold only slowly, and that policy ambition will always lag technological change.

The International Energy Agency, long accused of its own conservatism, broke ranks in 2021 with the introduction of the Announced Pledges Scenario (APS). The APS rests on two key assumptions: first, that governments deliver on climate and energy commitments already made; second, that renewable energy supply (wind, solar, batteries) and electrotechnologies on the demand side (EVs, heat pumps) are now competitive with incumbent oil‑and‑gas technologies.

Under the APS, global oil demand peaks before 2030 and declines steadily thereafter—not a collapse, but a structural transformation driven by new technologies. Increasingly, this is the world we inhabit.

The evidence lies in the real economy. Over the past decade, China’s scale‑up of clean‑technology manufacturing, the plunge in renewable‑energy costs, and the rapid deployment of electrotech have all defied the assumptions embedded in the OPEC, ExxonMobil, and EIA models. The transition unfolding before us validates the logic of the IEA’s APS and exposes the limits of the old faith in hydrocarbons’ eternal reign.

OPEC and the Architecture of Continuity

OPEC’s World Oil Outlook 2023 is an elegant document, thick with data tables and careful language. Beneath that precision lies a narrative of denial. Its framework rests on four interlocking premises: steady global economic growth led by the Global South; a conservative reading of policy, limited to what has already been enacted; technological progress that remains incremental rather than disruptive; and an energy mix in which hydrocarbons remain the backbone through 2045.

Those pillars yield a comforting storyline. Oil demand rises to around 117 million barrels per day by mid‑century. In this future, electric vehicles remain costly, charging networks thin, and developing economies perpetually short of the capital to electrify. The world changes, but not too much, and certainly not too quickly.

The flaw is structural. OPEC treats technology costs as exogenous to industrial scale. It assumes innovation follows an external curve rather than responding to learning, manufacturing volume, and policy diffusion. A world in which cost declines are driven by cumulative output, a world of feedback loops, barely exists in the World Oil Outlook. And because OPEC is a producers’ organisation, its brief is to signal stability, not transformation. The model therefore performs a political function as much as an analytical one: to reassure members and customers that oil remains indispensable.

ExxonMobil’s Outlook for Energy tells a similar story, cloaked in the language of investor prudence. It assumes hydrocarbons remain essential to global development, policy pledges are aspirational and unlikely to be enforced, and renewables, though cheaper, will not scale fast enough to displace oil and gas. Carbon capture and LNG are cast as core solutions, while EV adoption is assumed to plateau well before half of global sales. ExxonMobil’s future keeps oil demand above 110 million barrels per day through 2050, slightly lower than OPEC’s reference case.

The U.S. Energy Information Administration’s International Energy Outlook is no less cautious. Its “reference case” is formally policy‑neutral but functionally conservative: moderate GDP growth, stable population trends, minimal policy tightening. Renewables gain share, but gradually. Efficiency improves, but not enough to offset consumption that rises with income. The result again: oil drifts upward, gas expands, coal declines only marginally.

Shared across these models is a hidden premise: the global energy system changes slowly and disruption is exceptional.

The Counter-Scenario: A Different Logic

The IEA’s World Energy Outlook 2024 offers a different architecture. The Stated Policies Scenario (STEPS) describes a low‑ambition world in which only existing policies are implemented; oil demand plateaus near 103 million barrels per day and holds to mid‑century. The APS assumes governments deliver on national targets: net‑zero commitments, emissions caps, and industrial strategies already on the books. Under APS, oil demand peaks this decade, begins falling during the early 2030s, and then rapidly declines.

The divergence is not only policy; it is structure. APS embeds empirically observed learning curves: solar costs falling roughly 11 percent per year since 2015; batteries dropping about 18 percent with each doubling of cumulative production. The conservative models either dilute or ignore these dynamics.

The Empirical World: Learning at Scale

In 2010, a battery pack cost more than $1,000USD per kilowatt‑hour. By 2023, it cost $130USD, with credible projections of sixty by 2030. Every doubling of cumulative production has cut costs by nearly a fifth. Solar module prices are down more than ninety percent since 2010; onshore wind, roughly seventy. These are not marginal efficiencies; they are phase changes.

China reached the inflection point first. By fusing state planning, consumer incentives, and industrial finance, Beijing created the fastest‑scaling clean‑energy ecosystem in history. In 2024, China sold a record 12.8 million New Energy Vehicles, driving the global total to 17.4 million, and exported 1.3 million. BYD, SAIC, Geely, and Changan now manufacture at a scale that turns innovation into arithmetic.

Conservative models still treat China as a consumer, not a producer. In OPEC’s world, technology diffuses from the rich North to the developing South, constrained by income and infrastructure. In reality, diffusion has reversed. Chinese automakers are flooding Southeast Asia, Latin America, and Africa with sub‑$20,000 EVs and two‑wheelers. Charging networks, battery recycling, and joint‑venture factories follow, often financed through the Belt and Road Initiative. What OPEC calls “infrastructure limitation” is being tackled through what China calls “industrial diplomacy.”

The same pattern holds for solar and batteries. China now accounts for roughly 60 percent of new renewable capacity installed globally each year. Its solar‑panel manufacturing capacity alone is on track to exceed the United States’ total electricity demand by the early 2030s. Industrial learning has become geopolitical strategy.

In this empirical world, technological gradualism collapses. Once scale passes a threshold, parity follows; once parity arrives, substitution accelerates. Road‑transport oil demand is already bending. IEA data show that each 10 million EVs on the road displace roughly half a million barrels per day. With global EV sales projected to exceed 39 million annually by 2030, according to BloombergNEF, and the cumulative reduction could reach 10 million barrels a day—an entire Saudi Arabia of demand erased within a decade.

Conservative Models vs. Feedback Dynamics

Conservative models are linear by design: demand rises with GDP; technology improves incrementally; costs edge down over decades. Continuity is thus a built‑in feature, not a forecasted outcome. The clean‑energy system does not obey those rules. It evolves through feedback: each doubling of output lowers cost, spurring adoption and new investment, which in turn drives the next doubling. Once the cycle begins, the curve bends.

OPEC, ExxonMobil, and the EIA struggle to capture that curvature because their equations sever feedbacks. Policy is cast as the prime mover, technology as passive. In the real world, industrial momentum has become autonomous. China’s “overcapacity” in solar, batteries, and EVs is pushing prices down worldwide, accelerating deployment. Costs fall because production rises—and production rises because costs fall.

Capital Tells the Story

Money moves before policy. Since 2015, clean‑energy capital expenditure has grown about six percent annually, surpassing 2 trillionUSD in 2024. Under APS, it grows seven percent per year through 2030, while fossil‑fuel investment declines five percent annually. Generation still dominates clean investment, but the next wave is grids and storage—the very infrastructure that enables high renewable penetration.

This inversion matters. The world is no longer allocating capital as if oil demand will rise indefinitely. OPEC’s scenario, which requires rising fossil‑fuel capex through the 2030s, is already contradicted by

The Record of Prediction

History is an unforgiving auditor. Between 2010 and 2023, OPEC’s projections overestimated global oil demand by roughly ten million barrels per day. ExxonMobil consistently underestimated wind and solar deployment by more than half. The EIA’s 2014 projection for global solar capacity in 2023 missed by over two hundred percent. The IEA’s APS—although newer—has tracked more closely with observed trajectories in renewables, batteries, and EVs.

The direction of bias is clear. Producer‑centred models protect incumbents by undercounting disruption. APS may at times overestimate policy success, but its structural realism—its incorporation of feedback and learning—confers predictive strength.

The Moral Dimension

Forecasts are not just numbers; they are instruments of power. They shape investment, policy priorities, and public expectations. When an institution like OPEC projects rising oil demand for 20 years, it is not only forecasting, it is lobbying. The signal to markets and governments is that oil will remain indispensable, discouraging investment in alternatives and shaping reality to fit the model.

By contrast, APS aligns assumptions with explicit governmental commitments that encompass the vast majority of global GDP and emissions. It describes where political and industrial will are converging. In that sense, it is not optimistic; it is empirically faithful.

Canada’s Reckoning

For Canada, the implications are stark. The national energy narrative remains anchored to OPEC’s worldview. Alberta’s political and industry leaders routinely cite projections of rising global demand to justify expanding oil‑sands output to eight million barrels per day and building new pipelines to the Pacific. The assumption is that customers will always be there.

But what if they are not? The International Institute for Sustainable Development’s 2022 reading of IEA scenarios shows that under APS, global oil demand peaks around 2025 and falls steadily thereafter. Even the Canada Energy Regulator’s Evolving Policies Scenario—conservative by global standards—projects Canadian oil production peaking around 2032 and declining thereafter. The mathematics of global demand do not support long‑lived export growth.

Building new pipelines into a flattening market is not foresight; it is denial. None of the IEA’s credible scenarios, least of all APS, justify a single additional Canadian export line. By the time one could be permitted and financed, global demand will have passed its peak. Alberta’s vision of endless expansion belongs to OPEC’s imagined future, not the one taking shape.

The risk is financial, not ideological. Refining‑capacity growth is flattening. Upstream investment risk is rising as long‑term demand certainty erodes. Each new dollar committed to greenfield oil projects is a bet against the mathematics of technological change. Canada’s exposure is triple: misallocated capital, fiscal vulnerability, and lost time for diversification.

A Different Playbook

To remain competitive in the real economy, the one shaped by feedback and learning, Canada should rebase planning on APS or stronger. That means shifting public investment from extraction to electrification, from pipelines to grids, from resource rents to manufacturing capacity.

It also means integrating industrial policy with the energy transition, as China has done: linking critical‑minerals development to battery production, battery production to EV manufacturing, and EV manufacturing to export markets.

And it means updating the modelling culture itself. The Canada Energy Regulator should publish scenarios that internalize cost‑learning dynamics rather than treating technology as static.

Competitiveness can no longer be measured by cost per barrel. It must be measured by adaptability per megawatt‑hour: how effectively an economy translates energy into innovation, value, and flexibility. In the Canadian context, consider the oil sands’ fixed, capital‑heavy model (low adaptability) versus an electrotech ecosystem (EVs, heat pumps) that can pivot and scale (high adaptability).

The Shape of the Transition

The global energy system is not following the script of its incumbents. It is being rewritten by industrial learning curves and state‑directed investment. The physics of cost decline have overtaken the politics of continuity.

China’s EV revolution demonstrates the new logic. Industrial scale is now the engine of transition. Policy accelerates it, but even without policy, the economics of cheaper, cleaner technology would drive substitution. The result is a self‑propelling system that aligns with APS far better than with its rivals.

OPEC’s world is equilibrium; ExxonMobil’s is inertia; the EIA’s is caution. APS is feedback, momentum, and diffusion. That is the world visible in the data.

Oil will not vanish. It will endure as a legacy fuel and as a feedstock for chemicals, aviation, and industrial heat. But its supremacy is ending not because activists decreed it, but because arithmetic did. Every doubling of solar manufacturing cuts costs by another twenty percent. Each policy ratchet accelerates the learning that enables the next ratchet. The math compounds faster than politics can resist it.

Canada’s Choice

For Canada, this is warning and invitation. The warning is obvious: a high‑cost, export‑oriented petroleum system faces a structural demand decline it cannot hedge with rhetoric. The invitation is more profound: pivot from being a price‑taker in a shrinking commodity market to a value‑creator in the electrified economy.

That pivot requires honesty. It requires provincial and federal leaders to admit that the modelling assumptions underpinning their energy strategies—drawn from OPEC, ExxonMobil, and the EIA—no longer correspond to empirical reality. It requires a new industrial compact among governments, firms, and workers built on the logic of learning, not depletion.

The alternative is to be caught, again, on the wrong side of arithmetic.

The Forecast That Fits the Facts

When historians look back on the 2020s, they will see the decade when the world’s energy future broke free of its old equations. The conservative models—OPEC’s World Oil Outlook, ExxonMobil’s Outlook for Energy, and the EIA’s International Energy Outlook—describe a stability that no longer exists. Their assumptions are too linear, their politics too cautious, their treatment of technology too static.

The IEA’s Announced Pledges Scenario fits the world as it is becoming: fast, networked, feedback‑driven, industrially ambitious. It captures the interplay between policy and manufacturing scale, between cost decline and global diffusion. It is the scenario whose architecture best matches the physics of learning and the sociology of ambition.

For Canada, accepting that reality is not pessimism; it is realism. The question is no longer whether the global energy transition will accelerate, but whether Canada will move with it or against it.

The evidence sides with change. And change, in this century, is destiny.