Record earnings are masking the market’s biggest risk

Claims this week that this is the strongest corporate earnings environment in history, with barely a case for caution, are being directly challenged by the CEO of one of the world’s largest independent financial advisory organisations.

 
Nigel Green of deVere Group’s comments come as global spending on AI infrastructure heads toward $650bn this year, corporate earnings hit fresh records, and the S&P 500 pushes past 7,700, even as the accounting, financing and valuation assumptions underneath that strength draw mounting scrutiny.

He says: “Calling this the best earnings environment in history, and saying that a bear case takes creativity, gets it backwards.

“The real risk is hiding inside the numbers everyone is celebrating, not outside them.”

He continues: “Start with how these profits are being produced. Hyperscale tech companies are depreciating AI hardware over five to six years.

“Independent estimates put the real replacement cycle closer to two or three years, given how fast this equipment is being superseded.

“Stretch the depreciation schedule and today’s expense shrinks, which flatters today’s profit.

“If that gap is even half right, a meaningful slice of this year’s record earnings is a timing choice, not organic strength, and the bill still comes due.”

Global AI infrastructure spending is on course for somewhere between $600bn and $700bn this year alone, and capital on that scale only makes sense if it produces durable returns for years to come.

“Right now that payback timeline is an assumption rather than a demonstrated fact, and betting a record earnings season on an unproven assumption is not strength, it’s exposure dressed up as confidence,” notes the deVere CEO.

He adds: “Layer on how a growing share of this spending is financed. Suppliers of the infrastructure are investing directly in the companies buying it, who are committing spending straight back into the suppliers in return.

“This circularity doesn’t automatically signal anything improper, but it does mean a portion of the revenue and profit being applauded right now is the industry partly funding itself, which is far more fragile than the headline growth figures suggest.”

There is also a popular argument that persistent scepticism is itself bullish, the theory being that doubters eventually capitulate and buy in, and that capitulation fuels the next leg higher.

“I would treat that theory with real suspicion. Scepticism that refuses to break is not evidence it’s misplaced, it may simply mean the sceptics are seeing something the rally has not priced in yet.”

Stack stretched valuations on top of all of this and the picture gets harder to defend, not easier.

Multiples of 40 to 55 times revenue are being justified almost entirely by the strength of the earnings story just described.

Pull on any one thread, the depreciation assumptions, the financing circularity, the payback timeline, and the valuation sitting on top of it starts to look far less secure than the headline numbers suggest.

He concludes: “None of this means investors should panic or walk away from the AI trade altogether, because the underlying technology and demand are genuinely real, but the claim that this is the safest, strongest earnings environment in history, with no credible case for caution, simply does not hold up.”