Is U.S. Coal The Key To AI Expansion?

AI Capital Expenditure A Drag On Confidence

As predicted last week, the response to Nvidia AI related earnings on Wednesday determined the direction of the S&P 500.

Nvidia’s earnings were optically strong, and while there is little concern that it will continue to deliver on earnings, questions about the long-term sustainability of AI-related investment are gaining momentum.

In our recent posts, we have highlighted concerns about return on investment and the recent trend of circular financing in the AI industry as a drag on investor confidence.

Capital expenditure related to data centre build-outs has drawn particular attention.

Until recently, that capital expenditure was funded mainly from cash flow. However, over the last three months, hyperscalers’ bond issuance has totalled nearly $100 billion.  

The capital expenditure involved in the data centre infrastructure build-out is turning once-free-cash-flow large-cap tech companies into capital-intensive businesses.

AI Data Centre Construction Delay

To add to the mix, both Microsoft and Amazon Web Services (AWS) have paused or slowed down data centre construction projects.

Analysts note that hyperscalers are recalibrating data centre expansion plans due to AI “demand signals”, grid capacity constraints, and rising construction costs.

Grid capacity constraints and, more specifically, the cost of energy warrant particular attention. 

AI Cost Of Energy 

The cost of energy in the U.S. has been relatively cheap for the past 15 years due, in part, to the shale revolution.

However, U.S. oil production is peaking, and investment in energy infrastructure has been limited, in part due to regulations (ESG) that have made capital investment in the energy space more challenging.

Energy demand associated with data centre build-outs will be significant.

If the U.S. government is going to facilitate AI expansion and avoid higher electricity bills for consumers, then it needs to agree to outsource data centre build-outs to countries with low energy costs (for example, the Middle East) or find a way to produce significantly more electricity, and cheaply, within the next couple of years.

Given the timeframe, natural gas and gas turbine generators could be the answer.

That said, coal currently accounts for 18% of U.S. electricity production (in 2005 it was 50%) and, if the coal price becomes more competitive, a greater role can’t be ruled out.

That may explain the climate change perspective outlined by Trump in his September address to the United Nations.

S&P 500 Key Levels

The S&P 500 closed higher on Friday; however, it gave up gains on the day and, overall, corrected 2% on the previous week.

AI-related stocks experienced significant corrections to the downside and are now sitting at key support levels. Expect some buy-the-dip activity.

From a technical perspective, 6840 remains a key level. The S&P 500 needs to close and confirm above this level to signal that upside momentum is back in play.

Buy-the -dip activity could push price through 6650. However, momentum remains weak, and the balance of probability suggests choppy price action.

A further correction to the downside and a test of 6500 can’t be ruled out.