AI earnings are booming, but depreciation assumptions, circular financing and stretched valuations may be flattering the numbers.
**Words by Nigel Green, deVere Group CEO and Founder **
Commentators keep repeating this week that this profit cycle is the strongest the economy has ever produced, and that finding grounds for caution now takes real imagination, yet that framing has it upside down.
The genuine danger is not tucked away somewhere hard to find, it sits inside the same figures being applauded on every earnings call.
The cost of the AI boom #
Look first at how these results are being generated. The largest tech companies are spreading the cost of their AI hardware across five to six years of depreciation. Independent analysis suggests the actual working life of that equipment is closer to two or three years, given how rapidly each generation of chips is replaced by the next.
Lengthen a depreciation schedule and the expense booked today shrinks, which inflates the profit reported today.
If that mismatch is even partly accurate, a sizeable portion of this quarter’s results reflects an accounting choice rather than genuine commercial strength, and the true cost simply arrives later, whether or not this year’s statements show it. Worldwide outlay on AI infrastructure is tracking toward $650bn to $700bn before the year is out, reports say, and spending of that magnitude only makes sense if it delivers a lasting return over many years.
At present, that return horizon remains a working assumption rather than a proven outcome. Staking a full earnings season on an assumption that has not been tested is not conviction, it’s risk wearing the costume of certainty, and investors ought to be able to tell the two apart.
A large share of this outlay also moves in a closed loop. Firms supplying the hardware are taking direct stakes in the companies purchasing it, which then commit fresh spending back to those same suppliers.
None of that is inherently wrong, but it does mean a slice of the income being cheered right now is simply the sector paying itself.
Capital circulating this way is considerably more fragile than headline growth rates suggest, since it relies on every party continuing to spend at an unchanged pace indefinitely.
Doubt is not a bullish signal #
There’s a further claim worth interrogating, namely that stubborn investor doubt is itself a bullish indicator, on the basis that doubters eventually surrender and buy, and that surrender pushes prices higher still.
I doubt that logic holds. Doubt that has not broken is not proof it is wrong, it may simply show that the doubters are noticing something the rally has yet to price in, and writing off that possibility purely because no correction has arrived yet is wishful thinking rather than rigour.
Layer stretched pricing on top of everything above and the argument becomes harder to sustain, not easier.
Revenue multiples sitting between forty and fifty-five times are being defended almost entirely on the strength of the earnings story already thrown into doubt.
Tug at any one strand, the depreciation assumptions, the circular financing, or the unproven return horizon, and the valuation built on top of it looks considerably shakier than the headline figures imply.
Real demand or flattering maths? #
None of this calls for panic or for abandoning AI exposure altogether, since the underlying technology and genuine demand behind it are real, but it does call into question the assertion that this is the safest, strongest earnings environment on record with nothing to worry about, an assertion that simply does not survive scrutiny.
Investors need to pull genuine demand apart from the bookkeeping and funding arrangements currently flattering it, and drawing that line is precisely where independent, professional guidance proves its worth.
These businesses may ultimately vindicate everyone betting on them today. But labelling this the strongest earnings backdrop in history, with hardly any room for doubt, is not analysis so much as a slogan recited so often that nobody bothers checking the maths behind it anymore.
Whether investors come through this cleanly or get caught flat-footed once these assumptions face a real test depends on restraint and spread, not blind faith, and that question is worth sitting with now, long before the market answers it unprompted.
Nigel Green is the group CEO and founder of deVere Group, an independent global financial consultancy.
The views, information or opinions expressed in the commentary in this article are solely those of the author and do not represent the views of Stockhead.
Stockhead does not provide, endorse or otherwise assume responsibility for any financial advice contained in this article.
Related Topics #
UNLOCK INSIGHTS
Discover the untold stories of emerging ASX stocks.
Daily news and expert analysis, it's free to subscribe.
By proceeding, you confirm you understand that we handle personal information in accordance with our Privacy Policy.