The oil shocks of the 1970s are often remembered as a period of rising energy prices and economic instability. However, from an investment perspective, the more significant consequence was the exposure of a structural weakness within one of America’s most important industries.

For decades, the U.S. automotive sector had optimized around a single assumption: energy would remain abundant and inexpensive. Manufacturers designed vehicles, factories, supply chains, and business models around that premise. When oil prices rose sharply, the industry’s competitive position deteriorated rapidly relative to foreign competitors that had prioritized efficiency.

Today, as artificial intelligence, cloud infrastructure, and semiconductor demand reshape the technology sector, investors may benefit from asking a similar question:

The Real Lesson from the Oil Crisis

The common narrative surrounding the oil crisis focuses on fuel prices. Yet the more important investment lesson is that resource shocks rarely destroy industries directly. Instead, they expose inefficiencies that have accumulated over long periods of stability.

Prior to the 1970s, American automakers dominated global markets. The industry’s success was built on scale, manufacturing expertise, and consumer demand for larger vehicles. Fuel efficiency was often a secondary consideration.

When energy costs increased, consumers did not stop buying cars. They simply became less willing to pay for inefficiency.

Japanese manufacturers were uniquely positioned to benefit. Years before the crisis, they had invested heavily in smaller, more fuel-efficient vehicles. The oil shock did not create their advantage; it revealed it.

This distinction is critical.

The companies that emerge strongest following a structural disruption are often not the largest incumbents. They are frequently the organizations that have been optimizing for a future others have ignored.

Technology’s Equivalent of Cheap Oil

Since the emergence of modern computing in the 1980s and 1990s, the technology industry has operated under several assumptions that have largely held true:

• Computing power becomes cheaper over time.
• Electricity remains a relatively minor operating cost.
• Semiconductor manufacturing continues to advance.
• Digital infrastructure remains widely available.
• More computing resources generally produce better outcomes.

These assumptions have enabled extraordinary growth across software, cloud computing, and, more recently, artificial intelligence.

Yet history suggests that every dominant industrial paradigm eventually encounters constraints.

The question is not whether a constraint will emerge, but which constraint becomes economically significant.

Hypothesis One: Compute Becomes the New Scarce Resource

The most obvious candidate is computational capacity itself.

Artificial intelligence differs from previous software revolutions because it consumes resources at unprecedented scale. Training frontier models requires vast quantities of advanced semiconductors, energy, and infrastructure investment. For decades, technology firms benefited from a world where computing costs generally declined. Artificial intelligence may represent the first major technological shift where demand for computing power grows faster than supply.

In such a scenario, the key competitive advantage may shift away from raw scale and toward efficiency. Just as fuel economy became a strategic differentiator in the automotive industry, computational efficiency may become the defining metric for the next generation of AI companies.

Hypothesis Two: Energy Becomes the Constraint

A second possibility is that the bottleneck emerges through electricity rather than semiconductors. The rapid expansion of data centres has transformed energy consumption patterns globally. In several developed economies, forecasts now anticipate meaningful increases in electricity demand after decades of relative stability. If AI infrastructure continues to expand at current rates, energy could become a critical input rather than a background operating expense. Under this scenario, power generation, grid infrastructure, and energy security become increasingly important investment themes.

The comparison to the oil crisis becomes more direct:

• Oil powered industrial transportation.
• Electricity powers digital infrastructure.

As dependence increases, sensitivity to cost and availability increases as well.

The Data Centre Question

One of the most interesting parallels to Detroit may not be found in technology companies themselves, but in the infrastructure being built around them. Today, investors are witnessing an unprecedented wave of capital expenditure focused on data centres, AI clusters, semiconductor fabrication facilities, and supporting energy infrastructure. These investments are rational under current assumptions. However, history suggests that periods of rapid infrastructure expansion often coincide with extrapolation risk. What happens if future AI systems require significantly less computational power?

What happens if new architectures deliver equivalent performance using a fraction of today’s resources? What happens if AI increasingly moves to edge devices rather than centralized facilities? In such a world, existing infrastructure would not become obsolete overnight. Yet expected returns on invested capital could decline substantially. The automotive industry experienced a similar adjustment when efficiency became more valuable than scale.

The Search for Technology’s “Toyota Moment”

Perhaps the most compelling aspect of the analogy is not the resource shock itself, but the competitive response. The oil crisis did not elevate the largest automotive companies. It elevated the most efficient. This raises an important strategic question for investors:

Who is the Toyota of artificial intelligence?The answer may not be one of today’s largest technology firms. Instead, it could be a company developing:

• More efficient model architectures.
• Lower-power semiconductor designs.
• Novel approaches to inference.
• Edge computing solutions.
• New methods of reducing training costs.

In other words, the next decade’s winners may not be those with the greatest access to resources. They may be those capable of achieving similar outcomes with dramatically fewer resources.

A Different Way to Think About Risk

Market participants often focus on identifying the next breakthrough technology. History suggests that an equally important exercise is identifying the assumptions that underpin current market leadership. In the 1960s, few investors questioned the long-term viability of Detroit. By the late 1970s, the competitive landscape had changed fundamentally. Today, investors should consider whether artificial intelligence, cloud computing, and digital infrastructure are creating a similar dynamic.

The risk is not that technology adoption slows. The risk is that the industry’s definition of efficiency changes. When that happens, the companies best adapted to the previous era are not always best positioned for the next one. The oil shocks of the 1970s were not simply an energy event. They represented a reassessment of industrial efficiency.

The technology sector may eventually face a similar moment. Whether the constraint proves to be compute, electricity, semiconductor capacity, regulation, or an unforeseen technological breakthrough, the lesson remains the same: Industries rarely decline because demand disappears. They decline because the assumptions that made them successful cease to hold.