It’s payback time: Big Tech’s $700bn AI bill

The value of your investments and the income you receive from them can go up and down, and you may get back less than you invest. Any examples are for illustration purposes only.

Big leaps in technology tend to seem quick and clean when you look back on them. In the middle of them it can all feel a bit blurrier. Take mobile phones for example. We didn’t kick off with the iPhone moment and over $1tn in annual App Store sales [1] - my Philips Savvy (complete with aerial) is testament to that. More recently, the AI buildout has moved at such a pace that it’s becoming more and more tempting to see it all as one inflection point - pre and post-GPT - but the blurriness this time round is becoming less about the tech itself and much more about the price tag that comes with it.

In February, Amazon, Alphabet, Meta and Microsoft collectively announced spending plans of around $650bn [2] with AI, cloud and data centre projects all the clear beneficiaries. Last week, that figure rose to roughly $725bn [3] after the hyperscalers upped their budgets during Q1 earnings season. For a bit of perspective, the same four companies reported a combined capital expenditure figure of $151bn in 2023 rising to $246bn in 2024 and, having forecast at least $320bn this time last year, ultimately hit $410bn in 2025. [4] That uplift of around $575bn since 2023 is greater than the entire market capitalisation of Intel. In three years.

Amazon, Alphabet, Meta and Microsoft’s collective annual capital expenditure

Some estimates see Big tech shelling out $1tn in AI spending in 2027 [5]. Source: Financial Times.

Those upgrades have offered a double-edged sword to AI sentiment for at least a year, with investors often stuck between championing the technology and wondering how soon the headline firms will recoup those chunky outlays. And you don’t even need to be a tech fan to feel uncomfortable. Given how much these companies contribute to overall S&P 500 earnings (Alphabet, Amazon and Meta were the largest contributors to the rise in the S&P's overall earnings growth rate for the end-March period, according to FactSet [6]) and how much of the index they occupy (the Magnificent 7 made up 30.2% of the S&P 500 as of the end of March [7]), there’s a lot riding on these plans bearing fruit.

And that’s where the recent flurry of tech results provides a bit of direction for investors.

Looking at the latest updates from Silicon Valley, on the surface earnings are strong. Revenue is growing, cloud businesses are accelerating and AI is offering green shoots of monetisation - particularly at Alphabet, with cloud revenue jumping 63% and profits surging. Amazon Web Services (AWS) and Microsoft’s Azure are seeing demand strong enough to strain capacity and across the sector earnings are ahead of expectations. With that context in mind, it’s maybe not a surprise to see such huge spending plans - it all becomes a bit of a land grab with the companies that build the biggest, cheapest AI infrastructure winning out. That also presents its own problem though - if you aren’t shelling out hundreds of billions of dollars, what are you doing?

That train of thought will draw a few winces from those able to remember the late-1990s fibre optic boom. Enormous upfront investment preceded years of excess capacity and weak returns before the eventual winners emerged.

The recent spate of results helps separate those comparisons. While Alphabet shares got a bump after they reported, Microsoft stepped down and Meta fell off a cliff. The difference isn’t in the spending - all are clearly pumping cash into AI - rather it’s that Microsoft’s Q3 Azure growth of 40% just pipped expectations of 39% and was always going to have to smash estimates to justify a $190bn capex plan, and Meta’s path to monetisation just isn’t clear at the moment.

It’s another sign that the market is moving beyond being dazzled by AI innovation and is now expecting it to translate into tangible revenue. And when you’re one of the only ones not able to evidence where that’s going to come from, the market has its way of telling you to hurry up.

1. Clarity is being rewarded. Payback worries will likely rise as the staggering sums being spent balloon - markets are already discriminating between those with a plan and those lagging behind. The dot.com fibre optic memories won’t let even the snazziest plans cloud the eventual revenue picture so the key risk now isn’t that AI fails, it’s that returns take longer than expected, or don’t come at all. Look for clear routes to monetisation - that might be cloud growth, product pricing, utilisation rates - whatever the measure, this earnings season showed that the cleanest evidence gets the most confidence.
2. Valuations matter. Whichever way you want to slice it, hopes really are high. Mag-7 forward price-to-sales (P/S) ratios sit at 7.2, down from peak payback fear in late 2025 but still far above the c4.4 level when ChatGPT was released in 2022. [8] If ratios start to look even more stretched, the fall from grace could be significant, should the market not like what it sees. That matters even more now as higher inflation threatens to stop the interest rate-cutting trajectory. Those long-dated growth stories could start to look even more vulnerable without a path to profit.

The positive in all this is that the market seems to be rewarding genuine routes to monetisation. The broader danger is the asset-light, efficient profit centres in the tech space become the bulky, capital-intensive, asset-heavy firms that tech investors have loved to hate. If the same spending that protects long-term growth makes the same companies look less financially firm in the meantime, payback plans mean everything.

Sources: [1] Apple [2] Silicon Republic [3] Financial times [4] Financial Times [5] Barron's [6] FactSet, Investopedia [7] Vanguard [8] Yardeni

Important information

When investing, your capital is at risk. The value of your investments, and the income you receive from them, can go down as well as up and you may get back less than you invest. Forecasts aren’t a reliable guide to future results or returns.

Make sure to do your own research on what investments are right for you before investing or consider seeking expert financial advice. Please note that this article is meant for information and does not constitute any financial advice. This is not an offer, recommendation, inducement or invitation to buy, sell, or hold any securities, or to engage in any investment activity or strategy.

Robinhood doesn’t provide tax advice. You should seek advice if you have any questions regarding the impact your investments will have on your income tax and tax filing requirements.