AI Capital Expenditure Comes With A Price
AI Underpins U.S. Economic Growth
Capital expenditure expectations for both AI investment and energy usage over the next couple of years exceed $1 trillion.
The hyperscaler infrastructure build-out is contributing 1.5% to the U.S. GDP.
In the first quarter of 2025, the U.S. experienced an average GDP growth rate of 1.6%.
If you remove all the data centre infrastructure build-out spend, GDP growth was basically flat.
The AI trade underpins the entire U.S. economic growth.
Anything related to the AI data centre infrastructure build-out is driving the stocks to elevated market caps.
Morgan Stanley analysts estimate that 75% of all S&P 500 gains, 80% of profits, and 90% of capital expenditures in the last two years were accrued by AI-related stocks.
JP Morgan estimates that 70% of the gain in the S&P 500 can be linked to just 41 AI and AI-related stocks.
The 41 stocks now account for 47% of the S&P 500’s weighting.
What’s unclear is whether the data centre infrastructure build-out is based on forecasts of perceived demand or is simply an “arms race” that hyperscalers don’t want to lose.
If it’s the latter, then it will inevitably lead to overbuilding.
Returns On AI Investment
The recent Bain report indicated that, through 2030, there will be a need for $2 trillion in revenue to finance the big tech spend on AI. They estimate that revenue will be $800 billion short of that.
Internal documents show that Oracle generated approximately $900 million and a 14% gross margin from rentals of servers powered by Nvidia chips in the three months ending in August.
The gross margin figure is lower than many analysts have estimated.
In our 13 January post, we noted:
“Investors will want to see returns on AI investment by the end of 2025 or early 2026.
They will want to see business models and lifestyle changes driven by AI within the next twelve to fifteen months that deliver returns on investment.
If that doesn’t happen, then the momentum of mega-cap stocks may turn.”
Although the timeframe may be extended, our caution regarding the momentum of mega-cap stocks remains.
The capital expenditure involved in the data centre infrastructure build-out is turning once free cash flow large-cap tech stock companies into capital-intensive businesses.
The market continues to value big tech stocks as the asset-light businesses that they were.
However, they will effectively become utilities. That will likely impact their future valuations.
AI Hyperscalers Worth-A-Buy?
AI Returns On Investment In our January 2025 post, we outlined our concerns about the returns on capital expenditures for AI and AI-related stocks. We noted that while AI will be more significant than the Internet, that alone would not drive mega-cap performance in...
AI Market Disruption
AI Related Investments In The Spotlight Much of the focus in the S&P 500 last week was on earnings. More specifically, the earnings of AI-related stocks. Google and Amazon earnings were optically strong, and while there is little concern that they will continue to...
The Business Cycle Explained Simply
Let the stock market spark your curiosity, but let careful understanding guide your journey. The Business Cycle Explained Simply page can equip you with the knowledge and critical thinking to decode the complexities behind the stock market mechanics.