AI CAN MAKE A COMPANY EXTRAORDINARILY EASY TO BUILD WITHOUT MAKING THAT COMPANY EXTRAORDINARILY DIFFICULT TO COMPETE WITH

The extraordinary valuations being assigned to young AI companies may be overlooking one of AI’s most important consequences: the same technology that makes companies easier to build may also make them easier to compete with.

Something extraordinary is happening in the private AI markets.

Companies that barely existed a year or two ago are attracting valuations measured in billions of dollars. Founders in their early twenties can seemingly move from an idea—or even no idea—to companies worth extraordinary sums in remarkably short periods of time.

The latest example is AfterQuery.

Forbes reported this week that AfterQuery, founded by two young entrepreneurs who entered Y Combinator roughly 18 months ago, has reached a valuation of $3.2 billion.

Five months earlier, the company announced a $30 million Series A financing at a valuation of approximately $300 million.

If the new valuation is ultimately confirmed, that would represent an increase of more than tenfold in roughly five months.

The numbers are breathtaking.

But perhaps we are focusing on the wrong number.

The most important number may not be $3.2 billion.

It may be 18 months.

LOOK AT HOW QUICKLY THEY BUILT IT

AfterQuery reportedly provides sophisticated human-generated reasoning and other data used to help train advanced artificial intelligence systems.

It appears to be a real company addressing a real and potentially valuable requirement of frontier AI development.

Its founders also appear to have executed extraordinarily well.

But the speed with which the company was created raises a question that receives considerably less attention than its valuation:

If two very young founders could build this company in approximately 18 months, what prevents dozens of other small teams from attacking the same opportunity over the next 18 months?

That is not criticism of the founders.

It may actually tell us something much more important about artificial intelligence.

AI is dramatically reducing the amount of organizational capability required to create sophisticated businesses.

Software development can increasingly be accelerated by AI.

Research can increasingly be accelerated by AI.

Analysis can increasingly be accelerated by AI.

Customer service can increasingly be automated.

Data can increasingly be processed automatically.

Infrastructure can be rented.

Advanced intelligence itself can increasingly be purchased through an API.

Capabilities that once required large organizations, substantial staffs and years of accumulated expertise can increasingly be assembled by remarkably small groups of people.

That is an extraordinary entrepreneurial development.

But investors should recognize the other side of it.

AI CAN MAKE A COMPANY EXTRAORDINARILY EASY TO BUILD WITHOUT MAKING THAT COMPANY EXTRAORDINARILY DIFFICULT TO COMPETE WITH.

AI COMMODITIZES CAPABILITY

The history of business is largely a history of scarcity.

Companies developed competitive advantages because certain capabilities were difficult to acquire.

Engineering expertise was scarce.

Software development was expensive.

Specialized knowledge accumulated slowly.

Infrastructure required capital.

Research required people.

Organizations required management.

Distribution required scale.

These capabilities became barriers separating established businesses from potential competitors.

Artificial intelligence is beginning to attack those barriers.

A small organization equipped with sufficiently capable AI may increasingly possess research, analytical, engineering and operational capabilities that previously required hundreds of employees.

That means AI isn't merely commoditizing individual products.

AI may be commoditizing capability itself.

And when the cost of acquiring capability collapses, some traditional barriers to entry collapse with it.

THE SAME TECHNOLOGY IS AVAILABLE TO THE COMPETITOR

This creates a peculiar characteristic of the AI economy.

The technology helping an entrepreneur build extraordinarily quickly is generally not available exclusively to that entrepreneur.

It is available to competitors too.

The next entrepreneur has access to increasingly capable models.

So does the entrepreneur after that.

And the ten entrepreneurs after them.

The competitive question therefore changes.

It is no longer simply:

Can somebody reproduce what this company has built?

It becomes:

What happens when dozens of capable teams can simultaneously try?

That is an entirely different competitive environment.

And it raises serious questions about some of the enormous valuations currently being assigned to young AI companies.

A MOAT IS NOT THE SAME THING AS A BOTTLENECK

There is another distinction investors should consider.

The AI buildout is producing extraordinary shortages and bottlenecks.

Compute has been scarce.

Advanced chips have been scarce.

Power is becoming scarce in important markets.

High-quality training data can be scarce.

Specialized human expertise can be scarce.

Scarcity produces extraordinary economics.

But scarcity does not necessarily produce a permanent moat.

A moat protects a company's economics from competition.

A bottleneck produces extraordinary economics precisely because everyone has an incentive to eliminate it.

The greater the profits created by the bottleneck, the greater the incentive to find another way around it.

That distinction matters enormously when valuing AI companies.

Investors should be asking:

Are we valuing a durable competitive advantage—or temporarily capitalizing an AI bottleneck?

EXTRAORDINARY PROFITS ATTRACT EXTRAORDINARY COMPETITION

We can already see this mechanism operating at the very top of the AI economy.

Nvidia developed one of the most valuable technological positions in modern business.

Its advanced accelerators became extraordinarily important to AI development, and demand for them produced exceptional economics.

But those same economics created enormous incentives for Nvidia's largest customers to develop alternatives.

Google developed TPUs.

Amazon developed Trainium.

Microsoft developed Maia.

Meta developed its MTIA accelerator family.

Elon Musk's companies are pursuing increasingly ambitious vertical integration of AI compute and semiconductor capabilities.

None of this means Nvidia's competitive advantages have disappeared. They plainly have not.

It demonstrates something else.

Extraordinary economic rents invite extraordinary attempts to eliminate them.

If this is happening to companies possessing some of the deepest technological advantages in the world, investors should be especially careful when assigning multibillion-dollar valuations to businesses whose underlying capabilities may be substantially easier to reproduce.

THE AFTERQUERY QUESTION

We do not know enough about AfterQuery to conclude that its reported valuation is unjustified.

And the Institute is not making that claim.

The reported $3.2 billion valuation itself also deserves qualification. Forbes attributed the figure to people familiar with the matter, while AfterQuery reportedly declined to confirm it. Until a financing is publicly completed or otherwise independently verified, it should be treated as a reported valuation rather than an established transaction.

Similarly, spectacular revenue claims deserve careful examination.

Revenue is not gross profit.

Recurring revenue is not necessarily contracted long-term revenue.

Rapid growth is not necessarily durable growth.

And a private financing valuation is not the same thing as realizable economic value.

AfterQuery may possess proprietary data, exceptional customer relationships, superior processes, long-term contracts or other competitive advantages that prove extremely difficult to reproduce.

If it does, those advantages matter enormously.

But they are precisely what investors should be examining.

The interesting question isn't whether AfterQuery has built something valuable.

Clearly, sophisticated customers appear willing to pay for what it provides.

The question is:

What prevents somebody else from providing it?

And then:

What prevents dozens of companies from trying?

SPEED OF CREATION IS NOT NECESSARILY EVIDENCE OF DURABILITY

This may be one of the most important investment distinctions of the AI era.

Markets naturally celebrate extraordinary entrepreneurial speed.

A company reaches $100 million in revenue faster than companies historically did.

A startup becomes a unicorn in months instead of years.

A tiny organization achieves something that once required hundreds of employees.

Those accomplishments deserve recognition.

But investors should be careful about extrapolating speed of creation into durability of competitive advantage.

Indeed, the opposite may sometimes be true.

The extraordinary speed with which a company can be created may itself be evidence that the barriers to creating competing companies have fallen.

That doesn't make the original company worthless.

It changes how its future economics should be evaluated.

WHERE IS THE MOAT?

Every generation of technological change creates enormous fortunes.

Artificial intelligence almost certainly will too.

But AI may introduce an unusual characteristic into the process.

It could simultaneously accelerate the creation of companies and the creation of their competitors.

That means extraordinary growth rates and extraordinary valuations should increasingly be accompanied by an extraordinarily simple question:

Where is the moat?

What does this company possess that competitors cannot reproduce?

How long will that advantage last?

What prevents customers from building the capability internally?

What prevents AI itself from automating the scarce capability being sold?

What happens when today's bottleneck disappears?

And perhaps most importantly:

If this company could be built extraordinarily quickly, why can't its competitors be built extraordinarily quickly too?

Artificial intelligence may create companies faster than any technology in history.

That does not necessarily mean those companies deserve valuations based upon decades of protected future profits.

Because AI may be doing something else at exactly the same time.

It may be making competition faster too.

Institute for Strategic Intelligence

Understanding where artificial intelligence is going—and what it means for business, investment and society.

Prepared by the Institute for Strategic Intelligence in collaboration with ChatGPT.