Charlotte de Kerpoisson:

Hello and welcome to this podcast by BNP Paribas Wealth Management.  My name is Charlotte de Kerpoisson.

April and May were incredible months for stock investors – at least, for people with exposure to the global technology sector.

The Nasdaq 100 index is now running at +21% year-to-date, propelled higher; not by the Hyperscaler companies, but rather by semiconductor chip manufacturers, 66% higher since the start of this year. 

AI euphoria is in full swing, not only in the US, but also in other tech-heavy markets, such as South Korea, up 100% for the year so far.

To discuss technology in more detail, I’m joined by Edmund Shing, Global Chief Investment Officer.   

Hello Edmund.

Edmund Shing :

Hello Charlotte

Charlotte de Kerpoisson:

Can this almost parabolic move in IT hardware companies continue? How high is the risk of a market correction?

Edmund Shing :

Quite clearly, when you look at the performance, let's say, of the semiconductor sector over the last couple of months, which is up something like 66% year-to-date, this has been quite an incredible performance, most of which has happened in the last two months. There is, of course, a higher risk of a market correction to the downside. This is typically what happens. No parabolic move of the type that we've seen over the last two months can continue forever. It's just impossible. Therefore, I would believe that the risk of a market correction is higher.

Perhaps we might ask what the trigger would be for such a correction. This could be the upcoming initial public offerings of SpaceX, Anthropic, and OpenAI, these three AI companies. This may trigger a correction. It may be the growing realisation that these AI companies need to reach profitability, and that may be a challenge. And we might see some realism creeping into the earnings projections for all tech companies based on that. But it's very difficult to predict. As with any bubble, bubbles can go on for longer than anyone can believe or predict, and therefore it's a very difficult point.

On the one hand, we don't want to be over-exuberant and too excited, but on the other hand, we still have to participate. We cannot completely avoid the stock market because clearly the gains are being made and they are considerable, not just in the US, but on a global basis.

So I would say the risk of correction is certainly higher, but that doesn't mean that we do not participate in equities. That's why we remain neutral on equities for now.

Google recently announced a USD 80bn stock offering to fund AI investment plans. 

Charlotte de Kerpoisson:

What implications for the Hyperscalers and for the broader tech sector?

Edmund Shing :

Well, I think what is interesting is that Google is making a stock offering, so they're offering new shares. Now, we had known up until this point that the Hyperscalers, that is Google, Meta, Amazon, and Microsoft, would be investing something like $740 billion on AI and data centre investment this year. That is already up from an original estimate of $660 billion. So it's already more than we thought originally. We had thought that they would fund this by issuing a lot of corporate debt, a lot of bonds. And they have indeed been offering a lot of corporate bonds to the market.

But now it seems that Google has had to go one step further. The corporate bond offerings are not enough, and they want to raise even more money. That's why they've issued shares. That suggests to me that perhaps, on the one hand, maybe we're reaching the limits of what the US corporate bond market can absorb in terms of new issuance from US tech companies. On the other hand, it also may indicate that Google simply wants to invest even more than they had originally projected in order to stay in the AI race in an effort to win.

We are seeing somewhat of a "prisoner's dilemma" unfolding here where no major AI company, whether it be Google, Meta, Amazon, can afford not to invest because they cannot be seen to lose the AI race, even if ultimately, the profitability that comes out of their AI investments is lower than is perhaps expected today. So I think some caution should be exercised over the Magnificent Seven and the Hyperscalers' stocks. They are already actually underperforming the US market so far this year. The Magnificent Seven is up 8% where the S&P 500 is up 10% and Nasdaq 100 up 21%. So they are clearly not the obvious place to be within technology right now.

There are other areas, notably the IT hardware and semiconductor names where clearly the momentum is far stronger at the moment.

Charlotte de Kerpoisson:

What should investors holding S&P 500, Nasdaq, Emerging Markets or tech stocks do now?

Edmund Shing :

Now that is a very difficult question. Clearly for now, the argument would be to stay invested. The question is when to take profits. I think the argument would be to adopt something like a trend-following approach where you say: as long as these markets trend higher, you should stay invested. But at the first signs of any correctional weakness, then is the time to start to take profits and rotate into other sectors, maybe to take the money progressively out of the tech sectors, the technology exposure in emerging markets and the US and put them in other sectors.

The sectors, for instance, you might think about, one would be healthcare. Because on the one hand, healthcare is clearly one sector which will be a beneficiary of being an early AI adopter. So we see better profitability, better growth to come from the healthcare sector as they implement AI in their own business models. And yet, at the moment, the valuations of the healthcare sector are still very reasonable. So I think that it is much more defensive. It's not a cyclical sector, but it's much more a defensive sector. So our argument would be to progressively, particularly on signs of any weakness, take profits out of holdings in US and emerging market technology stocks and then reinvest them in more defensive areas that can benefit from AI, such as healthcare.

Charlotte de Kerpoisson:

The European tech sector has also caught AI fever, up 28% this year.  Where can we find cheap tech stocks now?

Edmund Shing :

Yes, it's sort of true. If you look at South Korea and Taiwan, the chipmakers there, the likes of TSMC, have certainly been revalued. The US clearly revalued as well. And now Europe on top of that. Well, there's one country that's left: China. Now there is still a lot of scepticism over the economic growth profile of China, given the property issues they have, and also, I think issues of maintaining 5% GDP growth. But nevertheless, it's quite interesting that of all the major tech sectors, the only one not to perform this year has been the Chinese tech sector, the names such as Alibaba, Tencent, Pinduoduo to name but three. I think there is still enormous value there.

To take one example, Pinduoduo (PDD), this company has half of its entire market capitalisation held today in cash on its balance sheet. It is trading at very low multiples. People are clearly very sceptical that its major operation, Temu, the online shopping network, can continue to grow as it has, can be as profitable as it has been. That, I think, is reflected in a very low valuation. So I would say look at the names, such as Alibaba, Tencent, and others. The Chinese internet giants are perhaps the last area where you can find cheap technology exposure today that has not yet been rerated.

Charlotte de Kerpoisson:

What 3 themes can we invest in to ride the AI investment wave?

Edmund Shing :

I think there are some areas we can still invest in. Clearly, if you think about the bottlenecks, the bottlenecks for AI, the number 1 bottleneck is power. Access to reliable, growing supply of electricity. And I think to resolve that, you want to invest in electricity infrastructure both for generation and transmission. This can include, of course, nuclear energy, renewable energy, natural gas turbines, energy transmission. All of these elements should work. So talking about electricity-related infrastructure in the US, Europe, and Asia, that is, I think, the first place to invest in. This theme has done well, but we still think it holds promise.

Secondly, think about the key raw materials necessary in terms of the investment, in terms of the building of data centres. I'm thinking, for instance, of copper. The amount of copper wiring that is required in a data centre is tremendous. So we do see a lot of demand for raw materials, industrial metals, such as copper and aluminium, even in use in semiconductor manufacture, for instance. So we're thinking about all manner of rare earth metals, which, by the way, 90% come from China today. So again, finding sources outside of China has become a new priority for Europe and for the US. And so I think those are two of the key areas, and I think selectively still investing in semiconductors in the so-called "picks and shovels" underlying the internet can work. But what I would perhaps focus on today would be within the software sector, which has been beaten down pretty hard since November last year, looking at cybersecurity.

Because, again, one thing is clear in this new AI world, data is the new oil, the new valuable commodity, and you need to protect your data. And that's why cybersecurity to protect your personal and your corporate data becomes even more important today than before. So again, investing in cybersecurity, we think, is a great third way to indirectly invest in the AI theme.

Charlotte de Kerpoisson:

Thank you Edmund. And thank you to our audience for following this podcast.  Please like share and subscribe to our weekly podcasts and visit our website for our investment themes for 2026 and strategy research. Goodbye.