Has AI's biggest spending boom just become software's biggest headache?
Watching $IBM(IBM)$ suffer its worst one-day share price collapse since records began in 1968 was rather like watching a seasoned marathon runner trip over the finishing tape. Few expected perfection, but almost nobody expected a 25%-plus plunge that erased roughly US$67 billion in market value before the company had even reported quarterly earnings.
Yet I think the market may be asking the wrong question.
The debate isn't simply whether IBM's AI turnaround has failed.
It's whether the AI investment cycle itself has temporarily turned against the very software companies expected to benefit from it.
That sounds absurd until you consider IBM's explanation.
Enterprise customers haven't cancelled AI projects. They've deferred software and consulting purchases while diverting budgets towards AI infrastructure—servers, storage and memory—amid an ongoing shortage of high-bandwidth memory.
In other words, the AI revolution may have temporarily robbed Peter to pay Jensen.
AI money isn't disappearing—it's simply taking a different exit
That distinction could prove enormously important.
Wall Street can't even agree on what just happened
Perhaps the strangest aspect of IBM's collapse isn't the share price at all.
It's that Wall Street has become more divided after the warning than before it.
HSBC responded by cutting IBM to ‘Reduce’ with a US$191 price target.
Morgan Stanley went in precisely the opposite direction, lifting its target to US$293.
That's more than US$100 of disagreement over exactly the same business only days after its worst trading session in almost six decades.
Normally, profit warnings force analysts towards consensus.
IBM has somehow produced the opposite.
The disconnect becomes even more fascinating once you leave Wall Street altogether.
Prediction markets are pricing roughly a 94% probability that IBM disappoints next quarter.
Meanwhile, the average analyst target still sits near US$274, implying almost 29% upside from Friday's closing price. Depending on which post-warning price you measure from, that implied upside has been quoted even higher.
Either the betting markets are far too pessimistic, or equity analysts are still giving IBM considerably more credit than investors currently are.
Both cannot be right for very long.
Sometimes volatility rewrites the investment narrative overnight
Following the money instead of the headlines
Strip away the drama and IBM's underlying financial picture hardly resembles a company in freefall.
Trailing twelve-month revenue has climbed to US$68.9 billion.
Net income has almost doubled to US$10.8 billion.
Diluted earnings per share have risen to US$11.31, while free cash flow stands at US$12.9 billion.
Margins have quietly strengthened too. Gross margin exceeds 58%, operating margin sits around 18%, and free cash flow margins remain close to 19%.
Those aren't the financial fingerprints of a business sliding into structural decline.
One underappreciated feature of IBM's recovery is that research and development spending has continued increasing while profitability has improved. Mature technology companies often sacrifice one for the other.
IBM has quietly managed to improve both.
Why investors felt blindsided
The violence of the sell-off wasn't really about one disappointing quarter.
It was about credibility.
Only three months ago IBM reported software revenue growing 11% to US$7.05 billion while comfortably beating earnings expectations.
That strong quarter reinforced the market's confidence that IBM's AI strategy was gaining traction.
Which is precisely why the warning landed like a punch to the jaw.
Investors weren't merely disappointed by weaker guidance.
They suddenly wondered whether management had recognised the slowdown developing sooner than they admitted.
Once confidence in guidance begins to crack, markets tend to mark management lower before they mark earnings lower.
The bull and bear cases are surprisingly balanced
The bullish argument remains straightforward.
IBM insists the deals haven't disappeared—they've simply slipped.
Customers are building AI infrastructure first before returning to software deployment, consulting and implementation later.
Importantly, management still expects roughly US$1 billion of year-on-year improvement in free cash flow.
If that proves correct, today's revenue disappointment may represent delayed timing rather than deteriorating demand.
The bearish case is equally compelling.
Perhaps this isn't a temporary pause at all.
Perhaps hyperscaler infrastructure spending has permanently become the first call on enterprise technology budgets, leaving software vendors competing for whatever remains.
If AI hardware continues consuming capital at today's pace, IBM may discover that software spending doesn't rebound nearly as quickly as investors hope.
That's a structural challenge rather than a cyclical one.
Competition has quietly changed
IBM no longer competes solely with $Microsoft(MSFT)$, $Oracle(ORCL)$, $Accenture PLC(ACN)$ or other enterprise technology companies.
Increasingly, it competes with the AI infrastructure spending cycle itself.
Every additional dollar allocated towards accelerated computing, memory and data-centre expansion becomes a dollar unavailable for software implementation today.
That's an unusually awkward competitive position.
IBM benefits enormously from enterprise AI adoption over the long term.
But in the short term, the hardware phase of that same investment cycle may be cannibalising its own revenue growth.
Few companies find themselves caught on both sides of the same megatrend.
Conviction usually returns before confidence catches up
The market may be overlooking one crucial distinction
Here's what I believe many investors haven't fully appreciated.
Cancelled software projects and delayed software projects are fundamentally different.
A cancellation reflects weaker demand.
A delay often reflects sequencing.
If IBM is telling the truth—and that remains the key question—customers haven't decided against buying software. They've simply decided to finish building the engine room before decorating the bridge.
That makes forecasting next quarter exceptionally difficult.
It also makes forecasting 2027 potentially far more interesting than forecasting July.
Broken story or simply awkward timing?
I don't believe IBM's investment case has suddenly collapsed.
Nor do I think investors should dismiss the warning as nothing more than a temporary hiccup.
Both conclusions feel too simplistic.
IBM remains a highly profitable business generating robust cash flow, improving margins and growing earnings. Those fundamentals didn't disappear in a single trading session.
At the same time, this episode exposed a genuine vulnerability in IBM's AI narrative. Success depends not only on enterprises embracing artificial intelligence, but on them reaching the software stage of adoption quickly enough to sustain IBM's growth.
On 22 July, management doesn't simply need to beat expectations.
It needs to restore credibility by proving delayed deals remain firmly within the 2026 pipeline and demonstrating that software, Red Hat and consulting can continue growing once the industry's hardware spending frenzy begins to ease.
Sometimes foundations deserve more patience than headlines allow
If IBM succeeds, history may remember this as the quarter when investors confused sequencing with weakness.
If it doesn't, this won't merely be the worst day since 1968.
It will be remembered as the moment the market realised that, in the AI gold rush, selling the maps can be surprisingly difficult while everyone is still buying the shovels.
@TigerStars @Daily_Discussion @Tiger_comments @Tiger_SG @Tiger_Earnings @TigerClub @TigerWire
Comments