IBM’s profit warning shows tech valuations are all in the timing

{"text":[[{"start":5.35,"text":"IBM’s worst day in 20 years was a pretty reasonable one for the market. The IT giant’s stock fell 25 per cent on Tuesday, but the S&P 500 closed slightly up. In other words, whatever ails “Big Blue” doesn’t seem to be rippling across the financial world. Which is strange, because the message encoded in IBM’s profit warning has implications for other companies too."}],[{"start":32,"text":"Chief executive Arvind Krishna warned that demand for IBM’s mainframe products would shrink faster than previously expected because of a “capex reprioritisation”. In plain English, customers are delaying purchases of big-box computers and related services so they can bring forward purchases of servers, storage and memory, where prices are rising. In plainer English: IT budgets may be ample, but they are finite."}],[{"start":null,"text":"

"}],[{"start":58.05,"text":"If one takes IBM at its word, this is a reshuffle rather than a reset and the long-term prognosis for its products has not changed. But the fact that investors sliced a quarter off its market value offers a lesson about the importance of timing, and one that also applies to providers of AI services themselves. An immutable rule of finance is that money due tomorrow is worth less than the same money arriving today."}],[{"start":84.05,"text":"Imagine a company with a cost of capital of 10 per cent — meaning that’s the rate at which future cash flows must be discounted to give their present value. If a certain level of revenue arrives two years later than expected, its value today falls by a fifth. Modest changes to the plan can thus cause bigger ructions in equity valuations."}],[{"start":105.44999999999999,"text":"There is a case for believing the expansion of AI revenue will happen more slowly than the hyperscalers have baked into their plans. Customers often struggle to calculate the value that an AI project will create. It doesn’t help that the benefit of AI investment made so far isn’t really showing up yet. Economists at Apollo recently showed that profitability for companies other than the biggest tech giants has barely risen in a decade. "}],[{"start":132.79999999999998,"text":"Even where AI has real and measurable benefits, customers may struggle to justify pushing the button as soon as their suppliers would like. Jettisoning staff whose jobs have been automated can take a long time, especially where companies are concerned about hallucinations and accountability for AI mistakes. Since budgets are not limitless, that could prompt more so-called reprioritisation. "}],[{"start":155.7,"text":"Silicon Valley is starting to pay more heed to the potential caesura between supply and demand. Bloomberg has reported that Facebook parent Meta Platforms may rent out surplus computing power, as Google does, squeezing cash today from assets that might otherwise not produce until tomorrow. Elon Musk’s xAI, part of SpaceX, just started doing the same."}],[{"start":179.04999999999998,"text":"That’s good for their investors, but suggests anxiety is mounting. Time really is money. This earnings season, heed not what companies plan to do with AI, but when they plan to do it."}],[{"start":197.74999999999997,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1784181266_5592.mp3"}