The math behind artificial intelligence growth

Wall Street is locked in a debate over whether the tech market is a sustainable growth cycle or a bubble. Daily price fluctuations often look driven by sentiment and AI hype. Stripping away corporate marketing reveals a different picture on actual balance sheets.

As highlighted in the Columbia Threadneedle Midyear Equity Outlook¹alongside insights from the Ameriprise Investment Research Group (IRG), this is a historic capital markets event. Tech giants are executing an unprecedented infrastructure buildout. Tracking where that capital flows makes the math behind today's market valuations highly visible.

Tracking the physical infrastructure layer

Evaluating the tech sector requires moving past consumer apps to focus entirely on the physical infrastructure layer. Trillions of dollars are buying tangible, capital-intensive assets:

• Data centers and hardware: Facilities are expanding globally to house advanced processing units. Goldman Sachs Research²projects global data center power capacity will scale up to 165% by 2030, requiring roughly $720 billion in grid upgrades along the way.
• The utility and power constraint: Data centers have an immense appetite for electricity. Projections from the Electric Power Research Institute (EPRI) Powering Intelligence Report³indicate they could consume between 9% and 17% of total U.S. electricity generation by 2030. While the national average rises, localized tech hubs face extreme concentration; EPRI notes Virginia’s data center share could hit up to 57% of the state's entire grid capacity by 2030, while seven other states—including Oregon, Iowa, Arizona and Nevada—are projected to exceed a 20% share.
• The grid impact: An updated study from the Lawrence Berkeley National Laboratory⁴aligns with these targets, calculating that U.S. data center power use will climb to between 9.5% and 15.3% of all U.S. electricity by 2030, up from 4.7% in 2024. This structural demand shifts the investment landscape, turning unglamorous sectors like regulated utilities, electrical equipment manufacturers and industrial infrastructure suppliers into critical bottlenecks with pricing power.

This physical spending is why tech infrastructure capital expenditure contributes heavily to broader macroeconomic data.

Understanding market dispersion and the software trap

For investors reviewing portfolios, a key theme noted by Columbia Threadneedle is dispersion. High dispersion means a handful of dominant tech infrastructure enablers achieve record earnings, while a wide performance gap opens between the winners and the rest of the market.

Certain tech segments face distinct headwinds. Traditional software and data-led companies are increasingly viewed as vulnerable to disruption from automated AI agents, causing a clear divergence in long-term earnings power within the sector itself.

Consequently, a broad, passive allocation to a standard tech index carries hidden concentration risks. You cannot simply buy the entire sector and assume everything will rise together. Selectivity and individual stock selection have become structural requirements.

Strategic patience outside of technology

While the infrastructure arms race dominates headlines, large-cap corporate earnings outside of tech remain resilient, growing at a comfortable low- to mid-teen pace. This underlying profitability can provide a strong stabilizing force for the broader economy, even as the Federal Reserve maintains an extended period of policy restraint than originally anticipated.

For clients across Hawaii, Seattle and California, a balanced wealth strategy means looking past immediate tech noise to find quality entry points in capital-intensive, non-AI sectors:

• Industrials and machinery: Companies building high-barrier physical equipment see downstream demand from manufacturing and grid buildouts.
• Financials and energy: These sectors offer attractive valuations and consistent cash flows, providing an essential defensive counterbalance to high-flying tech positions.

A sound retirement strategy does not rely on chasing record highs or reacting to headlines. By maintaining a programmatic, diversified portfolio built on hard metrics and corporate profitability, we work to help you plan for long-term growth while navigating short-term market volatility.

Sources cited:

1. https://www.columbiathreadneedleus.com/insights/latest-insights/midyear-equity-outlook-earnings-strength-fuels-optimism

2. https://www.goldmansachs.com/insights/articles/ai-to-drive-165-increase-in-data-center-power-demand-by-2030?ref=blog.brainpool.ai

3. https://powering-intelligence.epri.com/

4. https://eta.lbl.gov/publications/united-states-data-center-energy-2025