Is AI Creating Winners and Losers in Tech? What Investors Need to Know Now

AI is reshaping the technology sector at speed, and not everyone is winning. While semiconductor companies and AI infrastructure stocks have surged, software companies are facing an existential threat as AI tools undercut their traditional business models. For investors, the days of buying a broad technology fund and sitting back may be over. Here's what's happening, who's at risk, and where the opportunities might still lie.

In the past, a diversified

technology portfolio has served investors well. They had access to a range of segments, including cloud computing, software, cybersecurity and social media. However, AI is disrupting the entire sector, creating a gulf between winners and losers. Investors may need to be more discerning in their technology selection from here.

The performance of the Nasdaq Composite gives every indication of a sector in rude health – up over 13% for the year to date and 40% over one year[1]. However, look beneath the surface and there has been a vast gap between sectors within technology. The S&P Software & Services Select Industry Index is down 12% for the year to date[2], while the S&P Semiconductor Select Industry index is up 87.8%. High profile names such as Amazon and Alphabet are somewhere in between. Amazon is up 17.6%, while Alphabet is up 20%, for example.

The following shows returns on a £1000 investment over a one year period.

Which parts of the technology sector are performing, and which aren't?

Within funds, the top performers over one year are those that have been in the right parts of the market, including Polar Capital Global Technology, Amundi MSCI Semiconductors. Liontrust Global Technology, AB International Technology Portfolio and WisdomTree Artificial Intelligence. Wisdom Tree Cloud Computing ETF is at the bottom of the heap, alongside State Street SPDR MSCI Europe Communication Services, iShares Digitalisation, WisdomTree Cybersecurity UCITS ETF and Invesco Cybersecurity UCITS ETF.

Storm Uru, manager on the Liontrust Technology fund, says: “The winners of the last decade will not be the winners of this decade. The opportunity set is in a different part of the market.” He says this tends to happen in every technology cycle “and if you’re invested in the incumbents from the last cycle and not invested in the new companies. It’s very hard to beat the market.”

The assumption from markets has been that the gains from the vast spending programme from AI will accrue to the AI infrastructure companies – the chip makers and their supply chain. Big tech companies are slated to spend around $725bn in 2026 alone on building out AI capabilities[3] (#_ftn3). This build-out is created unprecedented demand for data centres, semiconductors and energy generation. These companies have done extremely well, not just Nvidia and other US manufacturers, but also Asia companies such as TSMC or Samsung.

Is AI infrastructure spending creating a bubble?

This is not a re-run of the technology bubble of the late 1990s. Uri says that companies are building into existing demand. For example, Nvidia’s latest set of results showed an 85% increase in year on year revenues.

Software companies are on the opposite side of the trade. The worry for many investors is that new model software groups such as Anthropic will eat their lunch. It is no coincidence that the software sell-off started immediately after the release of Anthropic latest Claude tool.

Why are software companies under threat from AI?

Uru says there has been a paradigm shift from software 1.0 to software 2.0. Many software companies will become commoditised and the value will sit elsewhere. “Share prices for the enterprise software companies have depreciated in a meaningful way as products have come to market from software 2.0 companies. The barriers to competition to creating software are weakening.”

David Coombs, head of the multi-asset team at Rathbones, says: “These software companies were long the most cherished part of the stock market, which meant many of them were highly valued relative to their prices… they have few physical assets that require upfront investment (like factories) or tie up cash (like raw materials that go into their products), so they don’t need big upfront loans or a big chunk of ongoing expenses to finance them. And that means virtually every extra customer or contract is pure profit, so they can grow rapidly and without constraint, since all they sell are intangible computer services. AI changes the game.”

Coombs says that software companies face another challenge: the standard business model for these companies is to sell their access to their programs by ‘seats’, i.e. the more users, the more revenue. He believes this will come unstuck if AI means fewer workers doing more with existing tools. In that situation, software companies could lose sales even as they take market share.”

The question from here is whether markets may have gone too far. Coombs believes that some businesses are structurally under threat: “We sold Booking.com, Salesforce and Adobe. As AI becomes more effective, companies will need fewer people to do the same work. Salesforce and Adobe make most of their money from per-seat revenue, so under their current business models, better AI would lead to less revenue. As for Booking.com, it aggregates prices and filters features, which AI can do for people without them having to scroll through endless pages of reviews.”

Have some tech stocks sold off too much?

But some businesses may have sold off too much. For example, UK data groups RELX and Stock Exchange Group have been caught up in the fear over software, but are still backed by many high profile UK fund managers such as Nick Train of the Finsbury Growth trust.

On the other side of the equation, there is the question over whether the share price appreciation for some of the AI infrastructure companies has gone too far. Nvidia’s share price is up 1,223% in five years, Advanced Micro Devices up 484%, while TSMC is up 292%. These companies are undoubtedly seeing strong demand, but a great company can be a bad investment if expectations are too high. Nvidia’s recent results were greeted with a shrug by investors, suggesting that the share price may have over-reached.

Could new IPOs disrupt the technology sector?

Victoria Scholar, head of investment, interactive investor says: “The bar is very high for the artificial intelligence bellwether which has made a habit out of delivering incredibly impressive results. Investors ‘bought the rumour, sold the fact’ as shares had already rallied ahead of earnings. There are some concerns among Nvidia investors about the growing threat of competition as the data center landscape shifts and hyperscalers develop their own chips.” She points out that Nvidia is still seeing price target upgrades on the back of earnings including from JP Morgan, Deutsche Bank, Morgan Stanley and Jefferies.

There is another problem within the technology sector. This year is likely to see a raft of new IPOs, including SpaceX, Anthropic and OpenAI. These are likely to disrupt existing indices, and passive investors will need to sell existing index heavyweights to accommodate these new entrants. This could weigh on Nvidia’s share price in the short-term, along with the other large technology companies – Apple, Meta, Alphabet and Amazon.

Should investors still hold passive technology funds?

This is an odd time for technology investors. The sector appears polarised into expensive but fast-growing AI infrastructure groups, or cheap but threatened software companies, with areas such as the hyperscalers or cybersecurity somewhere in between. While passive technology exposure used to be a reasonable solution to the complexities inherent in the sector, this period of disruption may require more careful selection between technology sub-sectors.

______

[1] MarketWatch

[2] S&P Global

[3] Financial Times