The Great AI Repricing: Why Data Moats Will Define Tech Leaders

From the compute arms race to data moats

By deploying capital when others are sellers, investors can acquire stakes in dominant technology franchises at discounts to intrinsic value.

Market sentiment is perpetually driven by extremes, and the current, breathless discourse surrounding Artificial Intelligence (AI) is no exception. One might conclude that the advent of large language models represents a terminal event for traditional commerce, monitoring today’s media. 

However, applying a dispassionate, analytical framework reveals a profoundly different reality: the development of AI is not an existential threat to business. Rather, it is simply the next logical phase in the structural evolution of organisational efficiency.

To capitalise on the genuine opportunities this technology presents, investors must separate the subsidised hype from the commercial reality, focusing on the idiosyncratic drivers of value creation and the structural tailwinds that define true market leadership.

The Reality of AI: Automata and Statistical Mimicry

To understand the commercial trajectory of AI, it is highly instructive to view it through the historical lens of the industrial automaton. During the industrial revolution, mechanical automatons were feared as the end of human labour rather they mechanised repetitive physical tasks and unleashed unprecedented productivity gains.

Current AI models operate on the exact same principle, albeit in the cognitive domain. At their core, these systems are not "thinking" machines; they are sophisticated engines of statistical mimicry. They predict patterns and generate outputs based on vast oceans of historical training data. They do not possess true understanding, nor do they innovate in a vacuum. Much like the mechanical loom, AI is a tool designed to automate repetitive, low-friction tasks. It is an instrument of productivity, magnifying the abilities of the user.

The Friction of Adoption: Compute Costs and the Cost of Capital

The AI industry is presently masked by significant frictional elements that distort its true economic viability. Chief among these is the staggering burden of AI compute costs. The training and continuous inference of large models require massive, capital-intensive infrastructure. Currently, these foundational models are loss making, heavily subsidised by venture capital and the deep pockets of big tech balance sheets attempting to capture early market share.

Investors must question the long-term viability of these profitless business models, especially given the current macroeconomic backdrop. We are operating in an environment characterised by inflation, tightening monetary policy, and most importantly, significant increases in the cost of capital. The spigots of 'helicopter money' that fuelled the speculative tech boom of the past decade are steadily closing.

As interest rates restrict funding and risk appetite normalises, the market's tolerance for cash-burning, heavily subsidised AI models will decline. The ultimate commercial winners will not necessarily be the creators of the largest, most expensive models, but rather those who focus on model efficiency and targeted AI applications. The deployment of AI must transition from generalised, bloated novelties to focused, highly efficient, domain-specific applications that can generate a tangible return on investment.

The Future of Work: Trust, Agency, and the AI-Augmented Human

The pervasive theme that AI will wholesale replace human capital fundamentally misunderstands the nature of value creation. High-value work will continue to be, at the very least, an AI-augmented human endeavour.

Investing and business leadership are never purely financial or computational endeavours. Complex decision-making, strategic capital allocation, and relationship management are driven by integrity, intelligence and incentives none of which AI possesses.

As AI rapidly advances and borders on the illusion of sentience, it raises profound commercial and societal issues around trust, agency, and the very real risks posed by bad actors. The ability to generate deepfakes, automate fraud, and manipulate data means that corporate governance and strict risk management will become more critical than ever. Customers will demand that humans remain firmly "in the loop" to maintain agency and mitigate these potential high-impact liabilities. AI will provide the analysis, but humans will bear the ultimate fiduciary and moral responsibility.

The True Moat: High-Quality Franchises and Proprietary Data Streams

If the foundational AI models themselves are destined to become commoditised—and they will be, given the rapid proliferation of open-source alternatives—where does the true investment value lie? 

The answer is unambiguous in that value lies in contemporary, proprietary data streams and the high-quality business franchises that control them.

Technology remains a core pillar for investment portfolios because high-quality franchises stand to benefit disproportionately from AI adoption. In any industry, competitive advantage essentially means scarcity, possessing an asset that cannot be merely bought off the shelf. As AI processing power becomes a ubiquitous commodity, data becomes the ultimate scarcity.

Businesses that capture deep, real-time, proprietary data flows will be the outsized winners of the AI revolution. Consider platforms that embed themselves into the daily transactional lives of consumers; they possess vital financial and behavioural data that credit agencies and generic AI models completely lack, transforming their closed databases into highly valuable, monetizable assets.

Crucially, as the public internet becomes increasingly polluted with AI-generated content, existing, human-generated data flows become exponentially more valuable because they are impervious to "synthetic data." Models trained on synthetic, AI-generated data eventually degrade and collapse. Therefore, contemporary, proprietary, ‘at the coal face’ data is the only fuel that can sustain accurate commercial AI applications going forward. Companies with genuine monopoly economics and structural data advantages will utilise commoditised AI to drastically expand their margins, cementing their status as the blue chips of tomorrow.

Emerging Value: Capitalising on the Tech Sell-Off

Market volatility is the greatest friend of the absolute return investor. Recently, the technology sector has been indiscriminately sold off by the broader market. While the herd often chases the latest macro thematic, certain sectors are inevitably forgotten, leading to significant valuation differentials and capturable alpha.

“Small cap earnings multiples remain at a material discount to large caps, something highly unusual historically,” Emanuel explains.

When the market panics, high-quality businesses frequently get marked down alongside fundamentally flawed enterprises. Following this recent tech sell-off, we see enormous, emerging value in tech market leaders. Many of these technology companies have spent the last two years cutting costs, optimising their balance sheets, and focusing on bottom-line growth; they are now significantly leaner and more profitable than they were during the peak of the previous bull market.

By deploying capital when others are forced sellers, investors can acquire stakes in fundamentally dominant technology franchises at a steep discount to their intrinsic value. These market leaders armed with their captive customer bases, unpolluted proprietary data streams, and the financial discipline to invest in efficient AI applications, are structurally positioned to generate enduring, outsized shareholder wealth in the coming decade. The AI revolution is not an existential threat to these businesses, indeed it is the very catalyst that will propel their next phase of transformational growth.