HAIL THE PIONEERS

There is a peculiar irony in technological revolutions.

The pioneers who create something genuinely transformative may eventually discover that their greatest achievement is making their own position less essential.

Electricity offers a useful analogy.

When electricity first emerged as a practical force, its generation and distribution were anything but ordinary. It required enormous investments, specialized knowledge, expensive machinery and networks that few organizations could create. The companies that developed and controlled those systems occupied extraordinary positions in the emerging economy.

But electricity had a characteristic that ultimately worked against permanent control of electricity itself.

It could be distributed.

It could be standardized.

It could be reproduced.

And, most importantly, once it became broadly available, its greatest economic value moved outward.

Electricity ceased being primarily something to *possess* and became something almost everyone could ‘use’.

The consequences were enormous. Industries transformed and factories were reorganized. Homes changed; transportation changed; communications changed. Entire businesses that had been impossible became commonplace.

The pioneers did not necessarily disappear. But the economy no longer revolved around the proposition that a handful of companies controlled electricity.

Electricity had become infrastructure.

That may offer a useful way of thinking about artificial intelligence.

We are still early enough in the AI revolution that it is easy to confuse control of AI with control of the economic value created by AI.

Today, frontier AI requires extraordinary resources: specialized researchers, enormous computing capacity, sophisticated data centers, energy, capital and increasingly complex relationships with the companies that provide the underlying infrastructure.

Anthropic is an especially striking example.

Its reported 2025 revenue approached $4.6 billion, while its infrastructure commitments reportedly reach an extraordinary $518 billion over coming years. The scale is difficult to comprehend because it represents a wager not merely on the success of a product, but on the proposition that increasingly capable artificial intelligence will become one of the dominant productive forces in the global economy.

That wager may prove correct.

But there is another possibility.

The wager may prove correct in a way that ultimately undermines the scarcity on which its enormous valuation depends.

That is the paradox.

Imagine that Anthropic and its competitors succeed beyond today's expectations.

Models become extraordinarily capable.

They become inexpensive enough to be incorporated into almost everything.

Businesses, individuals and governments use them routinely. Software products contain them almost invisibly. AI becomes a normal component of economic life.

Through that adaption, something fundamental has happened.

Artificial intelligence is no longer primarily a product.

It has become a utility.

And utilities behave differently from scarce technological prizes.

As computing became cheaper, computing power became less important as something to own and more important as something to use. The internet followed a related path. What was once exotic infrastructure became a platform upon which millions of other businesses could be built.

AI could follow the same trajectory.

The ultimate triumph of artificial intelligence may therefore be its democratization.

And democratization creates a peculiar problem for the pioneers.

This is where the notion of a ‘Pyrrhic evolution’ becomes interesting.

The pioneers may win the technological race and nevertheless lose the economic race implied by their initial valuations.

Not because they failed.

Because they succeeded.

Suppose Anthropic develops extraordinarily powerful models. Suppose Claude becomes deeply embedded throughout the economy. Suppose millions of businesses depend upon Anthropic technology.

Then suppose competitors catch up.

Open models improve. Algorithms become more efficient. Hardware becomes dramatically more capable. Smaller models accomplish tasks that once required frontier systems. Customers gain bargaining power. AI becomes abundant.

The economic value of AI continues to explode — while the price of obtaining the underlying intelligence falls.

At that point Anthropic might remain an enormous and important company.

But it would no longer necessarily be THE AI company.

It would be AN AI company.

That small change in language could represent an enormous change in valuation.

Return to consider what happened with electricity.

The great achievement was not permanently concentrating electricity in the hands of the people who first learned how to generate it.

The achievement was creating a world in which electricity could flow almost everywhere.

Once that happened, the economic imagination shifted.

The question was no longer:

“Who controls electricity?”

It became:

“What can we do now that electricity is available?”

That distinction is enormous.

The same could happen with AI.

Today's question is:

“Who controls the most capable artificial intelligence?”

Tomorrow's question may be:

“What can humanity do now that such intelligence is available almost everywhere?”

If that transition occurs, much of the economic value may migrate away from the companies that created the original intelligence and toward the millions of enterprises and individuals that use it.

The greatest beneficiaries might be companies that barely exist today.

A medical company may use AI to discover treatments.

A small manufacturer may use it to compete with multinational corporations.

A farmer may use it to optimize every acre.

A teacher may effectively have an expert tutor available to every student.

A novelist may work with an artificial collaborator.

A retired person may have access to capabilities once reserved for governments and corporations.

The AI companies may still provide the electricity, so to speak.

Everyone else may build the civilization that runs on it.

None of this diminishes the pioneers.

There is something magnificent about a technology becoming so successful that it escapes the circumstances of its creation.

The Wright brothers did not need to own aviation.

The inventors of the transistor did not need to control every computer.

The people who built the early internet did not need to own everything that traveled across it.

Their historical importance came from making something possible.

The same standard may eventually apply to today's AI pioneers.

If Anthropic, OpenAI and their competitors succeed in making machine intelligence extraordinarily capable, reliable and accessible, their achievement may be greater than any individual corporate valuation can measure.

Their reward may not ultimately be permanent dominance.

Their reward may be that they opened the door.

There is therefore a potentially fascinating irony in the enormous financial bets now being placed on AI.

Investors are attempting to identify the companies that will control the future of intelligence

But the future may eventually belong to a world in which intelligence is no longer something that needs to be controlled.

It is simply available.

That does not mean Anthropic or its peers necessarily collapse. They could remain major technology companies for generations.

But their evolution might resemble the evolution of the electrical industry: from pioneering scarcity, through enormous infrastructure investment, toward broad availability and competition.

The pioneer becomes part of the landscape.

Important, profitable, and influential. But no longer alone.

That could be the ultimate paradox of artificial intelligence.

The companies building the most powerful systems today may spend hundreds of billions of dollars attempting to make intelligence scarce enough to command extraordinary economic value.

And if they succeed brilliantly, their success may eventually make intelligence too abundant to command that premium.

That would be a Pyrrhic victory only in the narrow language of corporate valuation.

In the larger language of history, it would be something else entirely.

It would be progress.

Hail the pioneers.

Then move beyond the pioneers.