Evaluating AI: Market disruption, hype and what next?

The team explore whether AI is even right to be seen as the big disruptor, given how well-integrated it is into existing software providers. And they discuss if companies are able to adapt quickly enough to the implementation of AI tools.

Use the player icons above to listen on your favourite podcast app, or read the full transcript below.

This podcast isn’t personal advice, if unsure of a course of action for your circumstances, please ask for advice. Investments fall as well as rise in value so you could get back less than you invest. Past performance is not a guide to the future.

Full podcast episode transcript

[0:04] Emma Wall: Hello and welcome to the Switch Your Money On podcast from Hargreaves Lansdown. I’m Emma Wall – Chief Investment Strategist.

[0:10] Matt Britzman: And I’m Matt Britzman – Senior Equity Analyst – and this is your Investment Special, and today we’ll be talking everything AI.

[0:17] Emma Wall: Let’s kick off with some AI scene-setting, shall we, Matt?

There’s been quite a lot of AI news around. We’ll start with Result Season, where companies across different sectors have shared how 2025 wrapped, and they’ve given also some indication of the revenue and spending expectations for 2026.

[Pause]

[0:37] Emma Wall: Big Tech have had some big numbers when it comes to AI spend – am I right to say US $650 this year, alone?

[0:48] Matt Britzman: Not far off, depending on which period we include in that – but, clearly, from the results, that was the big takeaway.

It’s good to provide some context to these numbers as well, though, because the headlines sometimes don’t necessarily tell the whole story. So, if we look back to... I’m looking at four companies here – so Meta, Amazon, Alphabet, and Microsoft, typically the largest spenders of AI CapEx out there.

2025 calendar year – just across those four companies – was $375bn. 2026 calendar-year, expectations are, as you say, $611 bn just for those four companies – together, that’s over 60% growth from 2025-2026. Interestingly, 2024-2025 also had a similar 61% growth in CapEx over that period as well. So, this is now two years of extremely high CapEx growth – CapEx being the investment that the companies are spending to build out their AI capacity.

One of the interesting things to talk about first is that there’s a lot of noise going round about how this is gonna impact these businesses, and what it’s gonna do to the free cash flows. So, just wanted to maybe myth-bust slightly on that front because, whilst it’s true that the free cash flows of these businesses are expected to significantly deteriorate next year, it’s important just to understand what free cash flow is.

So, a business generates cash flow from its ongoing operations – we call that ‘Operating cash flow.’ To get to free cash flow, we then deduct the capital expenditure.

So, it’s not that these businesses aren’t expected to generate any cash flow next year from their operations – quite to the contrary. Operating cash flow forecasts for calendar-year 2026 are almost $700bn, just across those four businesses. So, they have the money – or they’re generating the money, the cash flow, to fund these investments, by and large – it’s just a change in the capital allocation decisions the management are making, whereas, previously, a lot of the excess cash that they’ve been generating has either sat on the balance sheets to create these kind of war chests that they’ve got, or they’ve been buying back shares aggressively, to return it to shareholders. So, the difference now is that, pretty much, all of that cash that they’re generating, they’re deciding to invest back into their businesses to try and go after this AI opportunity that’s ahead of them.

The interesting thing is, if we look at results that just came out for Q4 – broadly speaking, they were pretty good across the large tech companies, but the reactions off the back of them were completely different. So, if we look at Amazon, Google, Microsoft... the only one of the four, Meta is the one that had a positive sentiment shift after results – all four of the other ones were negative – and, again, not off of the back of necessarily bad earnings, but off of the back of the uncertainty that comes when you make a serious shift in your capital-allocation profile – which I think is what investors are concerned about, or at least building in and increasing the risk discount that they’re associating with these businesses.

So, mammoth numbers – a lot to think about when you go into them. It really depends on how you feel about the AI spend, as to whether you think this is positive or negative.

[3:57] Emma Wall: As you say, big numbers – $611bn, $650bn... what’s $40bn between friends? Big numbers and a mixed reaction from the share prices, there. As you say, companies that are spending that amount of money, with no guarantee of return on that investment, obviously has made the market nervous.

[4:17] Matt Britzman: Yeah, exactly – and the counterpoint of that is that you never, necessarily, have a guarantee on your investments. So, from an investor standpoint, it’s about trying to assess whether you think that there’s a good enough likelihood that returns are gonna come or not.

So, one of the reasons why I touched on the 2025-2026 CapEx raise is that... if you remember a year ago, we were kind of in a similar situation, where the 2025 numbers – which now looks tiny, at $375bn for those four companies – but, at that time, that was seen as a massive raise over the prior year – and we had all of these conversations at the time – question marks about whether they were gonna generate enough return on the 2025-spend to justify it.

Now, as I said, results have actually been pretty good. So, from the one standpoint, you could say that we are starting to see some proof that the investment is generating returns – it’s just again, obviously, because we’ve taken another step up, all of those concerns have come flooding back in – and, to some extent, rightly so.

But, Emma, big spending hasn’t been the only AI-dominator story in town. We’ve had a lot of stories in the news about AI disruption, and it’s been playing havoc with us on the share prices – so just give us a bit of colour as to what’s been happening there, Emma.

[5:27] Emma Wall: You’re talking about the catchily-named ‘SaaSpoclypse.’ Doesn’t actually work so well when you say it out loud – but, if you see if written down, it’s making the joke that software, as a service, is going through an apocalyptic time – because, you’re right... as well as the massive numbers about CapEx from the big four companies, the other thing that’s been really disrupting the stock market, over the last month, was a little large language model tool, called ‘Cowork,’ from Anthropic’s Claude, which basically has called into question entire industries – and it’s come in waves.

So, first of all, we had software – so companies ranged from 10%-50% share price losses... Oracle was down 50% in trading. So, the aggregate software and services index in the US was down 30%, but we saw companies – Oracle, ServiceNow, Salesforce, Workday – they’re all US companies... On this side of the pond, we saw Sage and RELX particularly hard hit. You know, RELX shares halved from their peak last year – and, one day, they had their biggest intraday loss since 1988.

I know that’s a stock that you own, Matt – and it’s also one of our 5 shares to watch for 2026. I’d be really interested in your views on that – but, before deep-dive on the software sector, it’s worth talking about the fact that software was the first industry taken out. But then, we had another tool that was released from Altruist corp, which then called into question the business model of wealth managers because it had a tax-planning tool. So, a number of our listed competitors – as well as St James’s Place – all fell on the FTSE – and large US firms, like Charles Schwab, lost significant value.

And then, we moved onto the next industry, which was more recently – which was property companies. So, big firms, like Savills, fell 7.5% – Workplace Group, and two big REITs... British Land and Landsec dropped between 2% and 3% – and all of this disruption is really speaking to the fact that AI is a disrupter. We know this, but it’s not potentially properly analysing what the impact would be there, because a number of the companies I’ve mentioned – both the US and the UK ones – actually already utilise AI. So, think about Salesforce – Salesforce has AI capability.

So, it’s some of the analyst rhetoric that has been around when Ecommerce came along – and Amazon came along – and really disrupted the bricks-and-mortar shop model... it’s not quite that comparison, is it?

So, we do think the trend is overdone, and there’s still some pretty good quality companies out there. What’s your Equity-Analyst view on that, Matt?

[8:10] Matt Britzman: Yeah – kind of largely spot on there.

I was looking at some numbers earlier – so was looking at the US software sector relative to the broader market. There’s only been two times in the last 30 years where it’s traded at a discount to the broader market – one of them is now and the other one back in 2008/2009. So, this is an unprecedented sell-off for specific sector – in software – and, as you mentioned, has now bled into other names. I think, broadly speaking, from our standpoint – when we’re looking at companies, and whether they’re gonna be AI winners and AI losers – it’s fair to say that, at the minute, the market is pricing everyone as an AI loser, and they’re then gonna hash it out down the line to see which ones come out on top or not.

Now, from our perspective, that provides an opportunity – if you can get in ahead of that – and there’s three key things that we’re looking at in our companies to whether we can identify they’re likely to be winners or losers. The first one is proprietary data – and you kind of touched on this, Emma – but one of the things that AI tools will be very good at is the application layer. So, the top layer – the interfaces, asking questions, getting responses – but a lot of the data analytics tools, and even some of the larger software companies... the real moats that they have is in underlying proprietary data – you know, data that can’t be easily accessible on the web. Times when companies have got 100 years’ worth of legal documents together, for example – some of which are no longer even in existence... that’s the really value-add element of the service that they’ve provided, not necessarily how you interface with them.

So, the first one we’re looking at is proprietary data – so, ‘Do they have underlying datasets that are really hard to replicate?’ The second one is switching costs – and you mentioned companies like Salesforce, for example, in the US... you know, some of these major companies, it’s open heart surgery to rip them out and replace them. I think that, to some extent, the market is underappreciating how difficult it is for, at least, definitely larger-scale businesses to simply shift all of the tools that they’re using for marketing, for customer service, etc.

So, switching costs and the ability and difficulties of switch is the second element we’re looking at, and the third one is regulated industries. So, we tend to think that the most defensible areas in software and data are gonna be the people that are operating in regulated industries. This is where the vendors have spent years building trust with customers, with the regulators, with auditors, compliance teams, etc – and, again, it’s a very difficult, and it’s a bigger job for a business to gut that and replace it with something new – unproven – than it would be for something else. So, the third element that we’re looking at now is regulated industries.

So, we’re putting those three things together – and, when we’re looking at software and data names, we’re trying to analyse them with an angle of, ‘How do they tie up to those three key elements?’

[11:00] Emma Wall: And would, Matt, RELX, itself... it’s results have come out after the sell-off... Did the financials – did the fundamentals give you some hope, or did they confirm this AI-disruption story?

[11:13] Matt Britzman: Yeah – so this is the interesting thing that’s happening at the moment, where the current financials are strong – and RELX is a perfect example, but you can apply this to the other names that have sold off.

So, the underlying business – again, for RELX, here... they had a strong set of results – and, in fact, it was a strong set of results, where the accelerating growth that they’re seeing is coming from AI. And I think you mentioned this earlier, Emma, that this is an example of a business that is benefitting, right now, from AI – management are calling it out as a key growth driver. We’re seeing it pull through in the numbers – the numbers were good – yet it’s still a name that’s come under serious pressure.

I think part of that is because the worry isn’t that 2026-earnings are necessarily gonna be impacted. The worry is that it’s murky as to what the AI capabilities look like when we go two, three, four years out – and part of that’s because the rate of pace of change of some of these new models is so fast.

So, I think we’re in an interesting kind of scenario now, where near-term performance can and, potentially, will be strong – or continuing – for a lot of these names – although there’s no guarantees. But, even with that, that may not be enough to cause a material shift in the sentiment right now.

So, the difficulty for companies like RELX and others – that find themselves in this conundrum – is that even really positive results may not shift the sentiment in the short term. Now, for investors, we think that it’s a perfect example of why taking a long-term view is important. There’s value on offer, in our opinion, for a selective number of quality names in these areas – but it may take some time for the markets to regain their confidence that they are actually AI winners and not AI losers. So, it’s gonna be a waiting game for good sets of results, to build on good sets of results, and then, over time, the market can reprice – but it is our view that it may take some time for that to actually feed through.

[13:08] Emma Wall: I think that’s fair – and I think, actually, taking time is a good mantra for AI, in general. You know, I was having a great conversation with somebody about the fact that this is an AI Revolution akin to an industrial revolution – or an electrification revolution, or even the Digital Revolution.

Revolutions have false starts – they’re not just linear – and they also benefit more than just the sector that is directly involved in it... by which I mean, industrial revolution and electrification didn’t just help those companies that supply electricity – similarly the Digital Revolution. The internet has helped almost every single sector bar none – and so, really, what we’re witnessing here is AI is disrupting, but, ultimately, as you say, those businesses that adapt – that evolve, iterate – are likely to see long-term benefits. The key is that implementation or adoption gap.

So, ‘Are companies moving fast enough to benefit?’ – ‘Will the adoption be seamless?’ – and the valuations that we’re seeing, and some of the areas of AI that do, despite the recent sell-off, look hot – look fully valued, if not overvalued – ‘Is that pricing – ‘Price-to-perfection,’ as we’re calling it – going to seamlessly roll out into implementation and adoption, or is there likely to be volatility?’

We’re in the camp, thinking there’s likely to be volatility. ‘Is AI going to benefit AI companies – companies that enable themselves with AI?’ Absolutely – ‘But is there likely to be volatility?’ – also, absolutely. Not least because of valuations, as I’ve mentioned, but there’s also quite a lot of economic and macroeconomic uncertainty still kicking about. We’ve still got Trump in the White House – lest we forget! – so our message is, ‘Focus on the fundamentals, tap into some of those key things that you’ve just been talking about – ‘How resilient are these businesses?’ – ‘Are they in a position where they’re adopting AI already?’ – and focus on the long term.’

[15:07] Matt Britzman: Yeah, I think that’s sound advice – and you mentioned about valuation point. So, this is data taken from [s.l. FACSA 15:12] – but S&P 500, at the moment, is at 21.5 x forward earnings. Five-year average is 20, so it’s slightly over where it’s been, historically. And, interestingly, the Tech sector, itself, is actually under where it’s been over the five-year average, and some of the premium that we’re seeing in the market is feeding through into sectors like energy materials, industrials. These are the workhorses of building the AI infrastructure that we’re seeing.

So, it’s interesting that we’re already starting to see – at least, in terms of the markets’ pricing – that some of those benefits of AI are gonna proliferate through, not just to the Tech sectors that you might think are gonna be the immediate and obvious winners, but also, potentially, through to some of these other industries that can benefit in terms of second-order applications.

And, I think, Emma – if we think about the fact that markets are not far off all-time highs right now, but we’re talking here about a massive sector going through a major sell-down... I think it could be a good time to remind listeners, as ever, of the benefits of diversification, and what that can bring to a portfolio.

[16:19] Emma Wall: Absolutely – diversify, diversify, diversity. As you said, yourself, we’re already starting to see second- and third-order benefits, but it also helps build resilience in portfolios.

Our house view for 2026 is that there will be volatility – there will be periods of underperformance – and one way that you can add resilience in a portfolio is to having exposure to different sectors which have different drivers. You know, whether that’s interest-rate sensitivity, macroeconomic-sensitivity, momentum-sensitivity. If you have a globally-diversified portfolio with – for those for whom it is appropriate – different asset classes reflective of your risk tolerance – then you’re likely to have a more resilient portfolio, which participates less when markets go down – and will have a higher likelihood of you reaching your financial outcomes.

[17:07] Matt Britzman: Sound advice – and that wraps us up for this week.

This session was recorded on February 16^th 2026, and all information was correct at the time of recording. Next week, we’ll have Helen Morrissey and Clare Stinton back with your Personal Finance episode.

[17:24] Emma Wall: And a reminder that nothing in this podcast is personal advice, and you should always seek advice if you’re unsure what’s right for you.

[17:29] Matt Britzman: And, as ever, investments rise and fall in value, so you could get back less than you invest – and past performance is not a guide to the future.

[17:35] Emma Wall: This podcast was not a recommendation to buy, sell, or hold any of the investments we’ve discussed today – and no view is given on the present or future value, or price of any investment, and investors should form their own view on any proposed investment.

[17:48] Matt Britzman: So, that’s all that’s left for us, so thank you again to our Producer, Elizabeth Hotson.

[17:52] Emma Wall: And thank you very much to you for listening – we’ll be back again soon. Goodbye!

[17:56] Matt Britzman: Goodbye!