Regarding the first point, a subject that naturally dominates today's op-eds and financial fervor, Arvind Krishna warns that if an AI bubble has indeed formed, it is primarily concentrated in the hardware and equipment sector.

Their clients have effectively ordered new data center capacity exceeding 100 GW. This represents investments in the range of 6 to 8 trillion dollars, supposedly amortized over five to seven years; provided, however, that they generate between 1 and 2 trillion in incremental annual revenue, a figure the CEO deems simply unrealistic in the short term.

Arvind Krishna also offers a distinctive take on the future of Large Language Models, which he believes are destined for irreversible commoditization. For the IBM chief, success will hinge primarily on distribution, meaning it will first favor companies with large established user bases.

Distribution would thus outweigh raw technology. In this scenario, the likes of Google, Amazon, Apple and Meta would hold the upper hand in a landscape that ultimately looks quite similar to the previous one. In short, while AI could indeed create enormous global value, only a handful of companies might capture the lion's share of it at the bottom line.

As for the transformation of IBM, whose share price in 2020 was at the same level as during the dot-com bubble at the turn of the century, Arvind Krishna predictably offers an enthusiastic presentation. It is true that in the six years under his tenure, the market capitalization of the Armonk-based group has tripled.

Fundamentals, unfortunately, contrast with this optimistic diagnosis. Over the last fiscal year, IBM's revenue and earnings were lower than they were 10 years ago, while the number of shares outstanding has remained virtually unchanged.

This track record is all the more critical given that IBM spends billions every year on its acquisition strategy. 2025 marked a new peak in this respect, with $8.2bn committed to external growth, a record since the Red Hat acquisition in 2019.

The silver lining is that we are - finally! - seeing positive growth momentum in operating profit following this renewed effort. This is perhaps what is leading investors to take a bullish view and value IBM at earnings multiples near historical highs.

However, as is the case with Hewlett Packard Enterprise, another faalen star of the American tech sector, it may be premature to declare victory. On this subject, see AI boom still very much theoretical at Hewlett Packard Enterprise, published this morning in these same columns.