AI investment illustration showing semiconductor chips, data centres, energy infrastructure and fluctuating financial market graphs
Home » Blog » Expert View » Beyond the AI Investment Hype

Beyond the AI Investment Hype

Why profitability is replacing market enthusiasm

By Utkarsh Ahuja, Founder and Managing Partner of Moon Pursuit Capital

Artificial intelligence has attracted extraordinary investment over the past few years, with companies commanding premium valuations based on expectations of transformative growth. But recent market volatility suggests investors are becoming increasingly selective about which businesses can justify those valuations.

I see this as both a valuation correction and a maturation of the AI investment cycle.

The market is not questioning whether AI will transform industries. It is questioning how much of that transformation has already been priced in and, more importantly, who will ultimately capture the economic value it creates.

One area I believe investors need to examine more closely is the interdependency of valuations across the AI ecosystem.

You cannot properly value an AI company without understanding the economics of the businesses it relies on. Semiconductor manufacturers, cloud providers, data centre operators and energy suppliers all influence its underlying cost structure and potential profitability.

The market is moving beyond simply rewarding AI exposure. Investors increasingly want to understand how much revenue these businesses can retain after paying the companies that make their technology possible.

That distinction will become increasingly important in determining which businesses deserve premium valuations.

From AI Hype to Economic Reality

The next phase of AI investing will be less about how quickly companies can grow and more about the economics behind that growth.

Many AI companies have significant financial commitments to a relatively concentrated group of infrastructure and service providers. Long-term compute agreements, cloud contracts, energy requirements and inference costs can create substantial obligations well before sustainable profitability is established.

What interests me is whether these businesses can become materially more efficient over time.

Can they reduce the cost of delivering their services? Can improvements in model architecture, semiconductor technology and infrastructure utilisation translate into better margins? And how much of those efficiency gains will they actually retain?

A company can double its revenue and still have a fundamentally unattractive business model if its underlying costs grow at a similar rate.

Investors need to look beyond adoption and revenue growth to understand whether scale is genuinely improving the economics of the business.

This distinction is particularly important when considering the extraordinary valuations that have emerged across the AI industry.

As highlighted in Tech Trends’ recent examination of AI unicorns and the extraordinary valuations being achieved by emerging technology companies, investor enthusiasm continues to generate enormous expectations for future growth.

But growth and profitability are not interchangeable.

For AI companies to justify their valuations over the long term, they must demonstrate that expanding adoption can translate into sustainable financial returns.

Finding Opportunities Beyond the AI Giants

Investor appetite for AI has become considerably more sophisticated and selective throughout 2026.

I remain constructive on AI, but some of the most compelling investment opportunities may be in companies solving the bottlenecks that are limiting the industry’s economics.

These include energy generation and efficiency, semiconductor innovation, inference optimisation, networking, data infrastructure and technologies that improve the utilisation of existing compute capacity.

As AI adoption expands, reducing the cost of delivering intelligence could become an enormous commercial opportunity in its own right.

This is an important distinction because much of the attention surrounding AI investment has focused on the companies developing increasingly sophisticated models and applications.

However, the infrastructure required to support those technologies represents a substantial commercial opportunity.

Businesses that make computing more efficient, reduce energy requirements or improve the performance of existing infrastructure could become increasingly valuable as AI adoption accelerates.

I also see attractive opportunities in enterprise AI applications that deliver measurable productivity gains and cost savings.

Businesses that become embedded in critical enterprise workflows and can demonstrate a tangible return on investment for their customers are particularly interesting.

The commercial potential of enterprise adoption is already influencing broader economic expectations. For example, research examining the UK’s AI readiness and its potential economic impact highlights the scale of the productivity gains that could emerge from more widespread adoption.

However, being positioned within an attractive segment does not automatically make a company an attractive investment.

Valuation, competitive positioning and capital efficiency still matter.

Some of the best opportunities may emerge in less celebrated businesses that quietly enable the broader AI ecosystem to operate more efficiently.

The Infrastructure Economics of AI

One of the most important developments shaping AI investment is the growing recognition that technological progress and commercial profitability are not necessarily the same thing.

The cost of developing and operating AI systems remains closely connected to the infrastructure supporting them.

Computing capacity, energy availability, semiconductor performance and data infrastructure all influence the economics of delivering AI services.

This means that advances in efficiency could become just as commercially significant as advances in model capabilities.

Consider the growing importance of technologies designed to improve the performance and efficiency of enterprise AI systems.

Developments such as Redis IRIS and its approach to improving AI agent memory and contextual capabilities illustrate how innovation is extending beyond the models themselves into the infrastructure and supporting technologies that make AI applications more useful.

From an investment perspective, the question is whether these technological improvements can ultimately translate into better financial performance.

Can businesses reduce the cost of serving each additional customer?

Can they improve infrastructure utilisation?

Can they diversify suppliers and reduce exposure to pricing pressures?

And perhaps most importantly, can they retain enough of those efficiency gains to improve their margins?

These are the questions that will increasingly determine which AI companies can build sustainable businesses.

What AI Companies Must Prove Before Going Public

Recent IPO performance offers important lessons for privately held AI companies considering public listings.

A successful IPO is not necessarily a successful investment.

Initial enthusiasm can quickly give way to scrutiny of a company’s underlying financials, particularly when valuations depend heavily on expectations of future growth.

Private AI companies preparing to go public need to articulate not only how they intend to grow, but how their cost structures will evolve as they scale.

Investors will increasingly examine infrastructure commitments, dependence on individual suppliers, contractual obligations and the assumptions underpinning future margins.

These considerations are especially important for capital-intensive AI businesses.

A company may demonstrate strong demand for its products while simultaneously accumulating substantial financial obligations to infrastructure providers.

If the cost of delivering those products remains elevated, revenue growth alone may not be sufficient to support the company’s valuation.

I expect some companies to delay their listings or recalibrate their valuation expectations.

Others may accelerate their plans to raise capital while market conditions remain favourable, particularly given the capital-intensive nature of the industry.

The companies best positioned to succeed will be those that can demonstrate credible commercial demand, greater visibility into future costs and a realistic path towards sustainable cash generation.

Public markets will be considerably less forgiving of valuations built predominantly on the promise of future growth.

Following the Money Through the AI Ecosystem

Disciplined capital allocation will become increasingly important as investors attempt to distinguish AI companies with genuine long-term potential from those benefiting primarily from market enthusiasm.

I would start by looking at the true cost of delivering the product or service.

Traditional metrics such as recurring revenue, gross margins, free cash flow and return on invested capital remain essential, but AI requires an additional layer of analysis.

How much does it cost to serve each incremental customer?

What percentage of revenue is effectively being transferred to infrastructure and service providers?

How exposed is the company to supplier pricing, energy costs and compute availability?

What financial commitments has it made against revenue it has yet to generate?

Equally important is the trajectory of those costs.

I want to understand whether a company can improve its economics through technological innovation, greater utilisation, supplier diversification or stronger negotiating power.

We also need to examine how capital flows through the ecosystem.

When infrastructure providers invest in AI developers that subsequently spend significant amounts with those same providers, investors should distinguish independently generated customer demand from revenue supported by interconnected financing arrangements.

These relationships can complicate the assessment of commercial performance.

A company may report substantial revenue growth, but investors need to understand the underlying financial relationships supporting that growth.

This does not automatically mean that such arrangements are unsustainable.

However, it does mean that traditional financial analysis must account for the interconnected nature of the AI ecosystem.

Ultimately, I am looking for businesses where revenue can scale faster than the capital and operating expenditure required to generate it.

That is where meaningful operating leverage and sustainable shareholder value can emerge.

Could an AI Correction Create New Opportunities?

Looking ahead to 2027, I expect further valuation volatility.

However, I view that as a potentially healthy development and a compelling opportunity for disciplined investors.

The next major opportunity in AI may come not simply from greater adoption, but from making AI dramatically more economical to deploy.

If advances in computing, model efficiency, energy infrastructure and competition meaningfully reduce the cost of delivering AI services, that could unlock an entirely new wave of commercial applications and profitability.

This is particularly important because many of the industry’s most ambitious growth projections depend on AI becoming increasingly accessible and affordable.

Reducing the cost of delivering intelligence could expand the range of commercially viable applications, creating opportunities for businesses that might otherwise struggle to justify the expense of implementation.

Conversely, if infrastructure bottlenecks persist and operating costs remain elevated, some of today’s valuations will become increasingly difficult to justify.

I remain bullish on the long-term trajectory of AI.

However, being bullish on a technology does not mean being bullish on every valuation associated with it.

Periods of market repricing can create exceptional entry points into fundamentally strong businesses.

For investors willing to examine the underlying economics rather than simply following market enthusiasm, volatility may reveal opportunities that are less apparent during periods of rapid expansion.

The Next Phase of AI Investment

The AI industry is entering a period in which commercial discipline will become increasingly important.

The extraordinary investment that has accelerated technological development has also created complex infrastructure dependencies, substantial financial commitments and ambitious expectations for future profitability.

As the market matures, investors will need to distinguish between companies that are benefiting from the broader enthusiasm surrounding AI and those capable of building genuinely sustainable businesses.

That distinction will not necessarily be determined by which companies develop the most advanced models or attract the largest funding rounds.

Instead, it may depend on which businesses can improve their operating efficiency, demonstrate measurable customer value and capture a meaningful share of the economic benefits created by AI.

The defining investment question heading into 2027 is not simply how large the AI market can become, but who can capture the most economic value from it, how efficiently they can do so and what price investors are paying for that opportunity.

The next chapter of AI investment may ultimately be defined less by how much capital the industry attracts and more by how effectively that capital is converted into sustainable returns.

About the Author

Utkarsh Ahuja, Founder and Managing Partner of Moon Pursuit Capital, wearing a blue blazer

Utkarsh Ahuja is the Founder and Managing Partner of Moon Pursuit Capital, an investment management firm specialising in digital assets, quantitative investment strategies and emerging technologies.

With more than a decade of experience in cryptocurrency markets and blockchain infrastructure, he focuses on identifying investment opportunities across digital assets, artificial intelligence and emerging technologies. His work encompasses investment strategy, capital allocation, market valuations and risk management.