One of the biggest spending sprees in corporate history is in full swing. Companies and investors are betting huge sums on generative artificial intelligence (AI) – the technology that promises to transform whole industries, economies and even societies.
Tech giants like Amazon, Google and Meta are leading the way on a splurge set to exceed $1 trillion in the coming years.
This colossal outlay includes major investments in the data centres, microchips and power grids that underpin AI.
It comes as BlackRock, the world’s largest investment manager, prepares to launch an AI fund of at least $30billion (£23billion) with technology titan Microsoft on projects to meet growing demands stemming from AI.
Excitement over AI’s potential has already propelled Nvidia, which dominates advanced microchip design, to become – briefly – the world’s biggest company by stock market valuation.
While it is the biggest winner so far, Nvidia’s shares have fallen back amid growing doubts about just how sustainable the AI boom is.
Questions are being asked about when, or even if, these massive investments will ever pay off or reap the benefits claimed.
So is it all just hype? And if it’s not, how are investors playing the AI game? Which stocks and sectors are being backed, and which are they wary of?
The answers may surprise you. We have been here before. At the turn of the millennium the stock market was gripped by dotcom fever.
Investors piled into companies in the hope that they would be among the winners of the internet revolution. Many were no more than a website without even a business plan. Few made any money – or ever would.
When the bubble burst, some of the biggest casualties were telecoms firms who had bet the farm on buying and building mobile networks. Among them was Vodafone, which landed Germany’s Mannesmann in a hostile takeover at the top of the dotcom boom.
Its shares now trade at levels last seen in 1997. By contrast some of the biggest winners of the digital dawn, like Meta-owned social networking site Facebook, did not even exist when the dotcom bubble burst.
Others such as Netflix, Amazon and Nvidia only really broke through after fundamentally re-inventing themselves.
A big difference this time is that the biggest AI winners so far – Nvidia, chip manufacturers such as Taiwan’s TSMC and Dutch-owned ASML, which produces the machines that make the chips – already make tons of money.
The question is not whether these ‘picks and shovels’ pioneers will be around in five or ten years’ time but whether they are overvalued now after enjoying spectacular share price rises in recent years on the back of the AI frenzy.
Like any new technology, a major attraction of AI for companies is the prospect of lower costs – and therefore higher profits.
For example, two thirds of professionals working in law, tax, accounting and fraud prevention think that AI could save them hundreds of hours a year by automating the most menial or unpopular tasks, according to a survey by data group Thomson Reuters.
While some of that saving could be used to rest or enjoy hobbies, it could also translate into $100,000 (£75,000) in new billable time per US lawyer each year, the survey said.
The prospect of enormous efficiency and productivity gains is fuelling an AI ‘arms race’ as tech giants pump record sums into the hardware needed to develop and run their AI models.
‘The risk of underinvesting is dramatically greater than the risk of overinvesting,’ says Sundar Pichai, boss of Google parent Alphabet.
The most tangible sign of all this spending is the depressing sight of joyless data centres springing up everywhere.
There are more than 500 of them in the UK, mostly in London and the South East. The US has more than 5,000. But the Government explicitly wants to encourage further construction as part of its proposed changes to planning rules.
Amazon recently pledged to invest £8billion in data centres in the UK, creating 14,000 jobs, and it is set to spend £110billion on them worldwide over the next 15 years.
Data centres – once known as ‘the cloud’ – used to store stuff like emails but are being adapted to house the specialised chips – mostly designed by Nvidia – that are needed to develop and run the next generation of AI applications.
AI data centres are far more power hungry than their predecessors because they use advanced chips that need a constant and reliable source of energy to operate.
Any dip in power supply could harm the so-called ‘training runs’ that are used to analyse, store and retrieve vast volumes of data to improve AI models. The financial stakes are high as each training run can cost tens of millions of pounds and take weeks.
How to navigate risks and prosper
By PATRICK TOOHER and ANNE ASHWORTH
The soaraway performance of the AI stocks may make you nervous about putting money into their shares in case there is sudden reverse in their fortunes.
One way to lessen the gamble is to invest in the funds and trusts that back pioneering AI or the businesses set to benefit from this extraordinary innovation.
You can hold these investments through an Individual Savings Account (ISA) or a Self Invested Personal Pension (SIPP).
Both these schemes can be set up on online platforms such as AJ Bell, Bestinvest, Hargreaves Lansdown and Interactive Investor.
Making monthly contributions, rather than committing a lump sum, lessens your exposure and allows you to benefit from any fall in the share price since your contribution buys a bigger slice of the fund or trust.
You can research the stocks that the managers of the fund or trust have selected by checking the fund factsheet which is available online. Check the factsheets of funds or trusts into which you have already put money.
For example, the popular F&C trust invests in tech giants Nvidia, Microsoft and Meta.
Scottish Mortgage, an investment trust which is a member of the FTSE 100, has stakes in Amazon, ASML, Meta, Nvidia and TSMC.
The trust (it does not invest in lenders, despite its name) is regarded as a high-risk proposition given that it also holds companies that are not yet listed, but may be among the future winners of the AI revolution.
Other options include Allianz Technology, which has stakes in Nvidia and TSMC and Polar Capital Technology. The portfolio of this trust includes Nvidia, Microsoft, Meta, TSMC and semiconductor company Advanced Micro Devices (AMD).
The Blue Whale Fund does not invest solely in technology but such AI names as Applied Materials, Broadcom, Meta and Nvidia are among its most substantial holdings.
Top holdings at Sanlam Global Artificial Intelligence encompass Microsoft, Nvidia and biotechnology company Intuitive Surgical.
In a country still wedded to fossil fuels, there are no fewer than 50 listed US utility or energy stocks, so investors are spoilt for choice.
That also makes picking individual winners harder so a better option might be iShares US utilities Exchange Traded Fund (ETF), which tracks an index of US equities in that sector and therefore spreads the investment risk.
British investors can buy American stocks through an online broker, but may pay a higher trading or administration fee.
Also UK-based shareholders of US companies must fill in a W-8 BEN form which allows them to pay a reduced tax rate on the investment.
Mark Zuckerberg, boss of Facebook owner Meta, thinks energy constraints are the biggest bottleneck to building AI data centres.
In 2022, data centres gobbled up 460 terawatt hours of electricity but the International Energy Agency reckons consumption will more than double to 1,000 terawatt hours by 2026.
That’s equivalent to the annual electricity consumption of Japan, a country of 124m people, the IEA notes. Power grids are already creaking.
There is a moratorium on new data centres being built in Dublin because they use nearly a fifth of Ireland’s electricity – a figure which is expected to grow significantly in the next few years.
So could it be that it is boring old utility companies, rather the tech giants, who will ultimately benefit from the AI revolution? Investors increasingly seem to think so.
They have punished shares in the likes of Microsoft and Meta for splashing out on AI without generating revenue fast enough.
David Cahn, a partner at tech investor Sequoia Capital, believes AI firms will need to generate £450billion in revenue to justify this year’s investment in data centres and chips alone.
Most firms don’t disclose how much they earn annually from AI but experts think it is at most in the tens of billions. That suggests the AI bubble may be at a tipping point – and could be about to burst.
Meanwhile, utilities, which include water and gas companies, is the best-performing sector in the benchmark S&P 500 index this year, returning 27 per cent, just beating information technology and communications services.
The utilities sector has not performed as strongly in the UK, but is still up nearly 4 per cent in the last six months.
Analysts at US investment bank Goldman Sachs think utilities have further to go because they offer investors two benefits: AI exposure and ‘defensiveness’.
Building grid capacity to meet the increased demand for AI will take time – and could cause power shortages in the meantime, the bank argues.
But the ‘significant’ economic benefits from data centres in terms of construction jobs and tax revenues make it worthwhile for power companies to invest in their infrastructure and use energy more efficiently, it adds.
‘No utility wants to turn away customers because they couldn’t provide enough power,’ says Brian Janous, co-founder of Cloverleaf Infrastructure, which helps data centres access renewable power such as wind and solar.
The next decade, though, is likely to be ‘painful’ as power demands outpaces available supply, he adds.
Goldman Sachs also likes the sector because utilities would continue to provide a steady stream of earnings – and dividends – in the event of an economic slowdown.
But Jim Covello, its global head of equity research, thinks the AI story will not end well. The technology is ‘exceptionally expensive’ and must be able to solve complex problems ‘which it isn’t designed to do’ if it is to justify those costs, he argues in a recent report.
‘What $1 trillion problem will AI solve?’ Covello asks.
Replacing low-wage jobs with costly AI is basically ‘the polar opposite of the prior technology transitions’ like the internet, he adds.
‘Even in its infancy, the internet was a low-cost technology solution that enabled e-commerce to replace costly incumbent solutions – Amazon could sell books at a lower cost than Barnes & Noble because it didn’t have to maintain costly brick-and-mortar locations,’ says Covello.
But ‘over-building things the world doesn’t have use for, or is not ready for, typically ends badly,’ he adds.
Daron Acemoglu, a professor at Massachusetts Institute of Technology, believes that only a quarter of
AI-exposed tasks will be cost-effective to automate within the next 10 years, implying that AI will impact less than 5pc of all tasks.
Enthusiasm for AI may begin to fade in the next 12 to 18 months unless important applications for the technology are found, Covello adds.
In the meantime, investors should stick with the AI ‘picks and shovels’ companies like Nvidia because they continue to benefit directly from massive AI infrastructure spend.
‘History suggests that an expensive valuation alone won’t stop a company’s stock price from rising further if the fundamentals that made the company expensive in the first place remain intact,’ says Covello.
Compare the best DIY investing platforms and stocks & shares Isas
Investing online is simple, cheap and can be done from your computer, tablet or phone at a time and place that suits you.
When it comes to choosing a DIY investing platform, stocks & shares Isa or a general investing account, the range of options might seem overwhelming.
Every provider has a slightly different offering, charging more or less for trading or holding shares and giving access to a different range of stocks, funds and investment trusts.
When weighing up the right one for you, it's important to to look at the service that it offers, along with administration charges and dealing fees, plus any other extra costs.
To help you compare the best investment accounts, we've crunched the facts and pulled together a comprehensive guide to choosing the best and cheapest investing account for you.
We highlight the main players in the table below but would advise doing your own research and considering the points in our full guide linked here.
>> This is Money's full guide to the best investing platforms and Isas
Platforms featured below are independently selected by This is Money’s specialist journalists. If you open an account using links which have an asterisk, This is Money will earn an affiliate commission. We do not allow this to affect our editorial independence.
| Admin charge | Charges notes | Fund dealing | Standard share, trust, ETF dealing | Regular investing | Dividend reinvestment | ||
|---|---|---|---|---|---|---|---|
| AJ Bell* | 0.25% | Max £3.50 per month for shares, trusts, ETFs. | £1.50 | £5 | £1.50 | £1.50 per deal | More details |
| Bestinvest* | 0.40% (0.2% for ready made portfolios) | Account fee cut to 0.2% for ready made investments | Free | £4.95 | Free for funds | Free for income funds | More details |
| Charles Stanley Direct* | 0.35% | No platform fee on shares if a trade in that month and annual max of £240 | Free | £11.50 | n/a | n/a | More details |
| Fidelity* | 0.35% on funds | £7.50 per month up to £25,000 or 0.35% with regular savings plan. | Free | £7.50 | Free funds £1.50 shares, trusts ETFs | £1.50 | More details |
| Hargreaves Lansdown* | 0.45% | Capped at £45 for shares, trusts, ETFs | Free | £11.95 | £1.50 | 1% (£1 min, £10 max) | More details |
| Interactive Investor* | £4.99 per month under £50k, £11.99 above, £10 extra for Sipp | Free trade worth £3.99 per month (does not apply to £4.99 plan) | £3.99 | £3.99 | Free | £0.99 | More details |
| iWeb | £100 one-off fee (waived until Dec 2024) | £5 | £5 | n/a | 2%, max £5 | More details | |
| Accounts that have some limits but attractive offers | |||||||
| Etoro* No investment funds or Sipp | Free | Investment account offers stocks and ETFs. Beware high risk CFDs. | Not available | Free | n/a | n/a | More details |
| Trading 212* | Free | Investment account offers stocks and ETFs. Beware high risk CFDs. | Not available | Free | n/a | Free | More details |
| Freetrade* No investment funds | Basic account free, Standard with Isa £5.99, Plus £11.99 | Freetrade Plus with more investments and Sipp is £9.99/month inc. Isa fee | No funds | Free | n/a | n/a | More details |
| Vanguard Only Vanguard's own products | 0.15% | Only Vanguard funds | Free | Free only Vanguard ETFs | Free | n/a | More details |
| (Source: ThisisMoney.co.uk July 2024. Admin % charge may be levied monthly or quarterly | |||||||
