The AI investment boom has created enormous growth opportunities, but the scale of spending by major technology companies may also generate a prolonged period of pressure on earnings and capital returns.

Key Highlights

  • Hyperscalers continue committing hundreds of billions of dollars to AI infrastructure.
    • Depreciation expenses will rise as data centres and AI hardware enter service.
    • AI revenue monetisation remains at an earlier stage than infrastructure deployment.
    • Earnings growth could lag investment growth for several years.

The AI Boom Is Also a Spending Shock

The race to dominate artificial intelligence has triggered one of the largest capital expenditure cycles in corporate history. Technology giants are investing aggressively in data centres, networking infrastructure, power capacity, and advanced processors to secure future leadership in AI.

Investors have largely celebrated these commitments, viewing them as evidence of long-term growth potential. Yet the financial consequences of such spending deserve closer attention.

Investment Comes Before Revenue

Large infrastructure programmes create a timing mismatch between investment and returns. Capital is deployed immediately, while revenue often arrives gradually over many years.

During that transition period, earnings can come under pressure even as strategic opportunities expand. For hyperscalers such as Amazon and Microsoft, the issue is becoming increasingly relevant.

As AI facilities move into production, depreciation expenses begin flowing through income statements. These charges represent the accounting cost of infrastructure assets and can become substantial when investment levels reach hundreds of billions of dollars.

Depreciation Does Not Wait for Demand

The challenge is that depreciation does not wait for commercial success. Whether AI applications generate expected revenue or not, infrastructure costs continue to be recognised.

This creates the possibility of a capital allocation recession, a period during which companies continue investing heavily while earnings growth struggles to keep pace.

Historical precedents provide useful perspective. Telecommunications companies during the late 1990s invested aggressively in network infrastructure based on expectations of future demand.

Digital connectivity ultimately transformed the economy, but many operators experienced years of financial strain before investment returns became fully visible.

AI Monetisation Is Still Developing

AI may follow a comparable trajectory. The commercial potential remains significant, but monetisation is still evolving.

Businesses are experimenting with AI tools, pricing models continue to develop, and long-term customer demand remains difficult to forecast with precision. Meanwhile, spending commitments are already locked in.

Data centres cannot be built incrementally once major construction programmes begin. Long-term supply agreements also limit flexibility.

Higher Rates Raise the Hurdle

Higher interest rates further complicate the equation. If borrowing costs remain elevated, the hurdle rate for generating acceptable returns on investment also rises.

Projects that appeared highly attractive in a lower-rate environment may face more scrutiny as financing conditions tighten. This matters because the AI buildout is not only about technology strategy but also about capital discipline.

The Next Phase Is About Returns

None of this suggests AI is commercially insignificant. The technology may ultimately justify the investment being made today.

The concern is that investors may be underestimating the time required for those investments to translate into earnings growth. Markets often focus on revenue potential while paying less attention to the financial mechanics that connect spending to profitability.

For shareholders, the next phase of the AI cycle may be determined less by infrastructure announcements and more by evidence that capital deployment is generating sustainable returns.