Hundreds of billions of dollars are being spent each year building the infrastructure necessary for the age of artificial intelligence.

The bulk of spending by national governments and global tech giants has so far primarily focused on physical infrastructure, such as production capacity for powerful new chips and vast, power-guzzling data centres.

But companies of all kinds, from retailers to advertising and recruitment firms, are now starting to embrace AI in efforts to slash costs, boost efficiency and ready themselves for the future.

And new companies and products are constantly emerging to meet the demand, creating fresh opportunities for investors hoping to cash in on the AI gold rush.

‘Our bullish view is that investors are still not fully appreciating the tidal wave of growth on the horizon from the $2trillion of spending over the next three years coming from enterprise and government around AI technology and use cases,’ analysts at Wedbush wrote in a recent note.

The broker has backed industry stalwarts Nvidia, Meta, Microsoft, Palantir and Tesla to continue to be the key beneficiaries in the second half of 2025.

Retail investors seem to largely agree, with Tesla, Palantir and Nvidia all among the 10 most-bought stocks held by Robinhood users by the end of the last quarter. AI data centre and cloud firm Applied Digital also featured as the second-most bought.

But Wedbush sees a change of focus ahead.

Companies of all kinds, from retailers to advertising and recruitment firms, are now starting to embrace AI in efforts to slash costs, boost efficiency and ready themselves for the future.

‘Now the time has come for the broader software space to get in on the AI revolution,’ it wrote, citing growing corporate adoption, more proven use cases and new large language models,’ it says.

‘[The] true adoption of generative AI will be a major catalyst for the software sector and key players to benefit from this once-in-a-generation, fourth Industrial Revolution set to benefit the tech space.

‘2025 so far has been an inflection year within enterprise generative AI as true adoption has begun by going from idea to scale as more companies are looking to invest into AI to decrease costs/increase productivity.’

The comments chime with analysis from Goldman Sachs, which estimates the total addressable market for the broader software industry will expand by ‘at least’ 20 per cent over the next four years as so-called hyperscalers prepare to spend $1trillion per year on AI.

Britain’s AI trailblazers

The UK currently ranks as the world’s third-largest AI market, behind the US and China. 

The Government has been keen to talk up Britain’s prospects as an AI destination though spending pledges, new policy initiatives and efforts to held build global regulatory standards.  

Shore Capital’s UK trailblazers 

And Britain is now home to AI unicorns, such as StabilityAI, Tractable and Quantexa, though none of these is currently listed on the stock exchange. 

‘For investors, this creates a compelling backdrop: a mature, innovation-rich environment with global influence, analysts at Shore Capital wrote in a note. 

The broker has drawn up a list of UK ‘trailblazers’ in AI, which includes companies in a variety of sectors from retail to hospitality. 

But the analysts favour companies ‘in the picks and shovels category of AI beneficiaries’ or those where AI offers the potential ‘to enhance an already positive investment’.

‘These include Craneware, Eagle Eye, Kainos, Softcat and Trainline,’ Shore Capital said.

But is it too early for software? 

Research analyst for BlackRock Fundamental Equities Rowan Palmour sees the transition as part of a three-phase progress.

He explains: ‘Phase one is the buildout – the race to build AI’s infrastructure needs. Phase two is adoption – with AI packaged into different apps and software. Phase three is transformation – where AI adoption potentially boosts productivity, creating new business models and industries.’

However, Palmour believes the industry is still in ‘phase one’ with hyperscalers ‘sticking with increased capital spending – and being rewarded for it’ with bumper earnings and share price gains.

He said: ‘We still like the hardware makers and those sectors benefitting from the buildout.

‘Utilities are key AI beneficiaries, with AI mentions in earnings calls far outpacing the S&P 500 average.

‘We see this interest persisting as data centres drive soaring power demand and a greater need for flexible, reliable supply.’

Goldman Sachs estimates the total addressable market for the broader software industry will expand by ‘at least’ 20% over the next four years

Matt Ward, co-manager of the £1.2illion AXA Framlington Global Technology fund, said growing numbers of companies are starting to embrace AI to ensure their systems are prepared to handle and process ‘an explosion of complexity’ and growing volumes of data.

However, the AI software solutions bought by firms remain largely ‘niche and focused on solving specific problems at this stage’, according to Ward, and there is little evidence of ‘mature, horizontally applicable software that can be taken on board by large customers and deliver consistent return on investment’.

He said: ‘Frankly we haven’t seen a singular product that is scalable in the billions of dollars of revenues. We’re some way away from that.’

AXA Framlington Global Technology currently allocates 32.44 per cent of its assets to semiconductor and semiconductor equipment, while software makes up 25.23 per cent, according to FE Fundinfo data.

US-based cloud companies Snowflake and Datadog, and supply chain platform JFrog are among the firm’s software holdings.

Ward added: ‘If you look at Salesforce, for example, pure AI revenues are a drop in the ocean in terms of their overall business.

‘But you look at Microsoft, it’s a material number now – it’s not a drop in the ocean.

‘It isn’t a concern for us – we just think we’re very early here.

‘We see the money coming through and being spent, and the forecast only going up.’

Gerrit Smit, head of global equity management at Stonehage Fleming Investment Management UK, says it is difficult to have ‘full confidence’ in much of the sector, which remains volatile and often unprofitable.

His firm’s focus has been on physical infrastructure names like Amphenol, which cables and connectors for data centres.

Smit said: ‘The benefits of AI are starting to be realised in different forms. Corporate AI uptake is very strong, so there’s no reservation about that.

‘We see it in many company presentations now, they will make a point of the difference that AI is going to make for them.’

DIY INVESTING PLATFORMS

AJ Bell

AJ Bell

Easy investing and ready-made portfolios

Hargreaves Lansdown

Hargreaves Lansdown

Free fund dealing and investment ideas

interactive investor

interactive investor

Flat-fee investing from £4.99 per month

InvestEngine

InvestEngine

Account and trading fee-free ETF investing

Trading 212

Trading 212

Free share dealing and no account fee

Affiliate links: If you take out a product This is Money may earn a commission. These deals are chosen by our editorial team, as we think they are worth highlighting. This does not affect our editorial independence.

Compare the best investing account for you