27 February 2026

Jonathan Sparks

Chief Investment Officer, UK, HSBC Private Bank and Premier Wealth

This year has delivered yet another twist in the tale of AI, drifting increasingly towards science fiction. Last year the narrative was all about the here and now; will there be enough datacentres? Will there be enough energy to power them? What about commodities, labour, planning approvals? While undoubtedly the demand was there, and the aptitude of models was clear to see, there was speculation on how far the models could develop. Yet any stock downside was limited to companies in the immediate firing line and broader downsides were longer-term and more academic – therefore out of reach for typical investors.

Back to the present day, and that narrative has shifted towards ‘what-iffery’. What if these ‘agentic’ AI models (advanced, autonomous systems using large language models to effectively act independently with limited supervision) are so good, they can wipe out whole business models? In that case, is anyone safe? Asset managers, lawyers, games companies – anything that monetised a software platform or an ‘intelligence premium’ - foresaw stripping of costs and slashing of jobs. A doom loop of intense job culls, mass unemployment, and further inevitable AI investment would surely follow.

This narrative shift had a spark: several genuinely jaw-dropping advancements from Anthropic’s Claude model and Google’s Genie application. However, investors appear to have embraced what began as an engaging science fiction story as a reasoned academic forecast.

The impact on the markets has been striking. The US S&P 500 software index has fallen by 20% this year as investors questioned the whole Saas (Software as a Service) sector, with some companies falling much further. Yet the market hasn’t just punished software: any business or sector that could be disrupted, as mentioned above, has been impacted.

Frankly, we can brush these predictions aside and see an opportunity. There are plenty of reasons why the attack on the Saas sector has gone too far. An AI model may well be capable of replacing an existing embedded software solution, but does it have access to the right data? Is it secure enough for critical systems? Is a company ready and willing to retrain all its staff? How easy is a highly embedded system to replace? Yes, a wave of AI applications could force Saas company fees lower, but then again, AI could also help those companies cut costs.

As for the doom loop of unemployment, that’s certainly a risk. But as history teaches, when something gets cheaper, demand rises, greasing the wheels of the economy and driving growth of other jobs – more likely focused on a human’s ‘personality premium’ than the ‘intelligence premium’.

With all the market volatility around software, the impact of the US supreme Court decision on tariffs has almost taken a backseat. The hard truth is that markets have become much more desensitised to the ebb and flow of tariff risks. Businesses have, to a greater or lesser extent, adapted. Earnings will be affected, but perspective is key: total tariffs collected under the ruled illegal tariff amounted to less than 1% of revenue of the top Fortune 500 US companies. In other words, this is an ongoing risk needing to be managed, but not enough to shift our leaning towards US markets.

Closer to home, next week the Chancellor will deliver her Spring Statement, which usually serves as an opportunity to adjust spending plans in light of updated economic forecasts. So far, these have been a mixed bag: green shoots in some survey data and a boost to government coffers from increased capital gains receipts to the tune of £7bn, but an anaemic growth outlook and troubling employment figures. That said, barring a big shock, Reeve’s goal will surely be to make as little market impact as possible, and deliver a statement that is, at best, boring.