Alexandre Zilliox

Alexandre Zilliox

Portfolio Manager
Thematics Asset Management

Carmine de Franco

Head of Research and ESG

Hua Cheng

Portfolio Manager
There’s no shortage of doomsday scenarios associated with rapid advancements in new technologies. They’re often fuelled by popular science fiction and the watchful eye of everyday reality. One day, it’s fears that artificial intelligence (AI) could replace the equivalent of 300 million full-time jobs.1 The next, it’s anxiety about to the coming ‘Q-Day’, when a powerful quantum computer could disrupt everything from secure communications to the underpinnings of our financial system.2

Among investors, concerns about an AI-driven financial crisis have been spelled out by Gary Gensler, the US Securities and Exchange Commission chair, who puts the likelihood of such an event within a decade as “nearly unavoidable” – without regulatory intervention.3 This follows a chorus of calls for more democratic oversight of the development of AI models.

According to British computer scientist Stuart Russell, “there are more regulations on sandwich shops than there are on AI companies”.4 Indeed, top AI academics have recommended that companies and governments should allocate at least one third of their AI research and development funding to ensuring the safety and ethical use of the systems4.

World leaders are stepping in. US President Joe Biden has announced wide ranging action by seeking to increase safety while protecting consumers, workers, and minority groups from the technology's related risks.5 Meanwhile, at a two-day summit at Bletchley Park – where Alan Turing helped crack the Enigma code used by the Nazis during World War II – UK prime Minister Rishi Sunak rallied 28 governments, including China and the US, to sign a declaration agreeing to cooperate on evaluating the risks of AI.6

Yet there’s a flipside to the AI coin – one that embraces the opportunities that technological advancements can bring and sees the future with a more optimistic lens. Take for instance, the news that an AI-powered brain implant helped to capture the electromagnetic signals of a patient paralyzed by stroke and translate them into digital words and facial expressions.7

Consider, too, that AI – including ChatGPT4 – will soon be allowed in all Australian schools after education ministers backed a national framework guiding the use of the technology.8 Or think about the ‘new’ Beatles song, which brings John Lennon’s voice back from the dead, and the fact that this feat of innovation was only made possible with AI.9

In short, whether you are broadly optimistic or pessimistic about an AI-driven future, there’s little doubt that investors will remember 2023 as the year of AI. The growth prospects associated with generative AI like ChatGPT4 – which is trained to create new outputs based on existing datasets – have seen technology stocks dominate equity markets.

We’ve witnessed dizzying returns, unfathomable projections and the acronym becoming a mainstay in seemingly every corporate earnings call, no matter the industry or sector. It’s why myriad companies have been scrambling to leverage the latest AI innovations in a bid to revolutionize their business models – hoping to disrupt before being disrupted.

We spoke with three investment specialists from our Expert Collective to understand how they’ve been investing in AI and where they see the future opportunities.
  • Alexandre Zilliox, Portfolio Manager, AI & Robotics Strategy, Thematics Asset Management
  • Carmine De Franco, Head of Research and ESG, Ossiam
  • Hua Cheng, Portfolio Manager, Global Sustainable Equity Strategy, Mirova
Everyone's been talking about AI this year. What aspects of the technology do you find most compelling?

Alexandre Zilliox, Portfolio Manager, Thematics AM (AZ): “One thing that struck us over the past 12 months is the fact that, given the developments around Large Language Models, anybody can now interact with AI systems. You don't need any kind of specific engineering skills to interact with them, so basically anybody can run a prompt and interact with it, given the ability of such models to understand our human language.

I think that's helped to drive the interest in the likes of ChatGPT. Also, you don't need any specific hardware to use it, it is accessible on your existing device: your smartphone or your computer, it doesn't matter.

Likewise, I would say that I think that people have realised that AI systems are really about enhancing productivity. The interest in the technology is now more obvious and there's a direct impact on productivity. Just take the example of Microsoft Copilot, an AI assistant feature that was introduced earlier this year, used in the Office suite to generate content or summarize, for example, which built on ChatGPT4, OpenAI’s Large Language Model.

We are also closely watching the trend of knowledge distillation where you essentially feed the knowledge of a very large model like GPT4 to a smaller one that will be easier to operate, less costly and more efficient. This process has appeared because for a specific task in a specific domain – for industry specific applications – you don't need a model that has 170 billion parameters. You could scale it down to make it more efficient and less costly to build, train and operate.

But if we go back 12 months, or even further back, the discussions we had with clients and prospects were really around education: what is the potential for the theme of AI, what AI can do today, where it can operate and how we believe AI will penetrate most industries and aspects of our lives, be it personal or professional, in the future.

This year, our discussions are much more focused on where the opportunity lies today – how can we benefit in terms of exposure to the trend in the short term – and what will the next layer of the value chain in AI look like, and therefore who the beneficiaries of tomorrow will be. People are now projecting themselves into the future more and trying to grasp the investment opportunity. We've seen quite an acceleration of interest from smaller investors and institutions this year, so the interest is real and it’s growing.”

Carmine De Franco, Head of Research and ESG, Ossiam (CF): “The speed at which we all started to talk about AI has been pretty impressive. Alex described the way that people interact with the tools. I don’t think we have seen the real applications yet. We’re still assessing if and how they will become popular, and if and how they can really have an impact on our lives.

But many industries are still lagging the application and the use of AI, even if they are sitting on huge amount of data. It’s data about their customers, data about their supply chains, data about their operational processes. Think for example of a chemical company, or a large electric distribution network. Reengineering businesses around data and AI to raise productivity will take time. So, while I agree with Alex on many things, I’m a little more conservative on the productivity aspect.

We’ve heard in the past about salaries and purchasing power increasing with technological advancements, but on few occasions economists were disappointed. We were sold the idea that the internet would make us all more productive: while it has significantly changed the economy, productivity has not risen more than in the pre–internet era.

That being said, the trend is clear and this is probably where we're going to be in the next 20 to 30 years. The thing that will probably have a more direct impact on us, as financial professionals, will be discovering which industry, or which company can emerge or can transform itself around AI and reap the benefit out of it.

How do we identify them now? Can we imagine where AI – and its impact on the economy – will be in 10 years from now? Only then, maybe, AI will lift productivity economy wide. But there will be sectors where rising productivity will be difficult, and I'm not sure that AI necessarily helps. These are big – and growing – sectors in the economy. Think about hospitality. Think about healthcare.”

Hua Cheng, Portfolio Manager, Mirova (HC): “The first thing I would say is that AI is not new. It’s just that the opportunities have only become tangible for most people very recently. So we are seeing rapid AI development driven by computing power, the explosion of data and heavy investment in the industry. Personally, I believe that we will see step–change improvements in technology. I'm really excited about the opportunities of generative AI and its use cases in many sectors – not only IT, but also in manufacturing, healthcare, and so on.

I agree that for several sectors, we still don't know precisely how AI could be applied. But overall, I believe that the potential opportunities are massive. We're talking several trillion dollars.”

ChatGPT and Nvidia’s ability to profit from the growth of AI have upended the markets in 2023. Is there a danger that innovations in AI will be the next big thing for only a limited number of tech companies?

AZ: “I think that in the short term, it's highly probable that larger companies will drive innovation because if you think about foundation models – the models behind the chatbots like ChatGPT – you need a huge amount of investment in hardware. You need a profoundly large data set on which to train this model. ChatGPT4, for example, costs probably between $50 million and $100 million, taking into account the employees, the salaries, the hardware and so on.

But coming back to the idea of knowledge distillation, we are seeing a lot of private companies emerging with a reduced number of parameters in their models that will be very useful and that will address specific use cases. We are seeing an ecosystem where large companies are providing the foundation, but smaller companies are increasingly leveraging their work.”

HC: “On the one hand, I believe with that generative AI can be used in many companies, not only a large cap but also medium and smaller companies. We look at whether a company already has differentiated industry expertise, but also a unique database. AI can leverage the database and enhance existing differentiated industry expertise, and we look for that combination of existing expertise with potential AI functionality, which can be in every segment of the size of the enterprise.”

CF: “I think that as we discover the technology and also its negative effects, I can imagine a future where, for example, in regulated sectors, companies may be required to handle the entire value chain of AI with all the safety checks required by the regulation. For example, think about an utility company trying to manage its network with smart and AI-driven features, or a healthcare company or a bank. I don't see how companies in these ‘sensitive’ sectors could rely solely on off-the-shelf AI models handled by the likes of Alphabet or Microsoft, using them under licence while training their customer data. It could raise issues from regulators around the world.

Therefore, it is likely that large companies will have to get their hands on the technology, maybe using an off-the-shelf product that is tailor-made to their needs, or even developing them internally. The example of this type of development can be seen in the electric-auto business, where large companies have preferred to invest in software and develop it internally rather than buy commercially available tools that could power their cars.”

Nvidia is representative of a sector that's experienced excessive valuations.10 It’s also an example of a stock that's been theoretical overvalued but that’s continued to double its earnings.11 How do you think about current valuations and future growth with these types of stocks?

HC: “We became really excited about Nvidia around three years ago because, before that, the business driver was mainly from GPUs [graphics processing units] in the gaming sector. We realised that the future growth would come more and more from data centres, artificial intelligence, and autonomous driving instead of gaming GPUs.

Another reason is that when we looked at the various entry points, we believed that Nvidia not only has leading technology around the GPU itself, but it has a really integrated platform, with library, software and partnerships with customers. So, the entry points for companies like Nvidia are getting higher and higher because it's more and more difficult for competitors to catch up.

And if we look along the value chain, at software companies, we believe there’s differentiated and meaningfully valuable functionalities. They can be winners over the long term because they can really monetize their products or services. Companies like Microsoft and Adobe. So, Microsoft's investments in ChatGPT and Copilot, as Alex mentioned, are well known by now. But perhaps Adobe is less obvious. Yet when I look at the long term history of innovation from Adobe, it has clearly been investing in AI functionality for many years, and that has been accelerating over the last one to two years.

I remember when I was a student in college and I tried to use some Photoshop functionality for a project. It wasn't easy. Now, with AI functionality, everyone can become a basic graphic designer. The shift from just five to ten years ago is incredible. And similarly, we believe that Adobe has focused on its content creation based tool to make sure there are no issues. This is a really key differentiator against many other companies in my opinion.

Then, if we go all the way through the value chain to users, we see opportunities there too. We look for companies that have already differentiated industry experiences and can leverage AI and data to enhance competitiveness over the long term.

A good example is Intuitive Surgical, a global leader in robotic surgery provision. Over the past 20 years it has had almost zero competition, and it now has a bigger database with more than five million procedure data. I believe that companies like Intuitive Surgical can really combine leading technology on the robotic surgery platform, plus data and functionality to pursue even more exciting growth opportunities over the long term.”

AZ: “We've been owners of Nvidia since December 2018, nearly five years. We’ve always had strong conviction in the company and it has always been in our top ten holdings in the portfolio, sometimes making up 7% or 8% of the portfolio.

How do we manage the position and think about valuation? It's a long-term component. Nvidia has very strong barriers to entry which have even been reinforced over the past year, and they have the fundamentals. It's really more a question of the cycle, and specifically the valuation related to the cycle.

At the end of 2021, the stock had a strong run that was mainly based on the metaverse hype.12 At that time everybody was talking about the metaverse as being the next big thing. Facebook even changed its name to Meta in October of that year. Everybody got ahead of the opportunity but it just never really materialised. At the time, the business mix was roughly 60% graphics applications and less than 40% for data centres.

We had the sense that the valuation was stretched and it was hard to justify the position – or at least justify having 7–8% of the portfolio in Nvidia – especially given that rates were edging higher. We had doubts about the macroeconomic cycle not to mention concerns about inflation and rates. At this point we started to integrate higher discount rates into our models and there just wasn’t that much upside. So, we reduced our position in Nvidia. By 2022, the stock was down probably 50–60%. So, we believe it turned out to be a good call.

Towards the end of 2022, we started to become more constructive. Given all the investment that had happened and the buzz around ChatGPT, we started to hear increasing comments about the capex spending of companies like Meta which were investing a lot in hardware infrastructure and chips. We thought the impact would be felt more in the second half of 2023, so when it showed up in Nvidia’s numbers in the first quarter, we were caught a little by surprise. It all materialised a lot faster than we expected.

Fundamentally, we were looking at the profile of the datacentre business. Profitability is much higher here than in the gaming business. You also have much more visibility on growth, as Nvidia is directly embedded into the R&D road maps of its customers. They have good visibility over what's coming down the pipe. And, to Hua’s point, you also have a software component which brings a bit of recurrence to the revenues and will likely be growing in the future. All of which makes Nvidia an interesting case study at how we review the investment case for AI.

Despite the strong performance since the beginning of the year, the stock is not trading at a material premium relative to history. When you look at the last seven years, on average, it's trading roughly around the average,13 even though you could argue that the growth profile is stronger now than it has been over recent years. What’s more, the technological moat is still there and you now have greater visibility. So you could argue that it should have higher multiples.

We could make the same argument for some other technology stocks. Indeed, looking at the valuation of the Magnificent 7, if you exclude Microsoft and Apple, they are actually trading at a discount to their averages over 10 years.14 As such I wouldn't necessarily say that technology stocks are expensive right now, despite the strong returns and the contribution to the overall stock market. The opportunity over the next few years remains quite attractive. And I don’t think we are in some kind of bubble.”

CF: “From a sector perspective valuations shall be understood differently as one would for stocks. Stocks may experience exponential growth or they can go to default. It depends on many things: their products, the quality of their management, the business environment, customer appeal, brand awareness, and so on. But sectors have a sort of mean reversion because they are related to the ebbs and flows of the economy as well as to investors’ expectations and risk appetite.

For example, many view 2024 as a complex year for equities, for many reasons. If such a scenario turns out to be true, one might wonder which companies might get hit first. Will people sit on the gains they have, for example on Nvidia, Meta or Alphabet, or will they need to liquidate their position to cover losses elsewhere? This is why, from a sector perspective, we tend to be very pragmatic.

We have been investing in tech. We bought the US IT sector around the middle of last year when it was already suffering significant losses. And we only bought back US Communication Services early this year. But we removed IT before summer – not because we don't like it, but because we see more opportunities in other sectors. Those that have been disregarded by the markets, especially at the beginning of the year.

If you think about the S&P 500 and you remove the so-called ‘Magnificent 7’, the rest of the index is almost flat year to date15. That means the entire market, save for these seven companies, is not growing. And of course, this is not sustainable. There will be rotations. Maybe tech will have to readjust in a world with increasing uncertainty – interest rates, inflation, US elections, geopolitical risk. Companies might have to pause their investments because they must roll over their debt at higher rates, constraining their own investments This may have a short-term impact on the IT sector as well. At the same time other sectors might benefit.

So, rotations make us very pragmatic on tech. We still believe that there is potential to grow, but we do not count out sectors that will be users of the technology rather than just the providers of it.”

  • Artificial intelligence (AI) – the simulation of human intelligence processes by machines, especially computer systems. Specific applications of AI include expert systems, natural language processing, speech recognition and machine vision.
  • Chatbot – Short for ‘chatterbot’, these are computer programs that simulate human conversation through voice commands or text chats or both.
  • Deep learning – Also known as ‘deep neural networks’, deep learning is part of a broader family of ‘machine learning’ methods based on learning data representations, as opposed to task-specific algorithms. Neural networks are a series of algorithms that endeavours to recognize underlying relationships in a set of data through a process that mimics the way the human brain operates. Deep learning analyses data in multiple layers of learning (hence ‘deep’) and may start doing so by learning about simpler concepts and combining these simpler concepts to learn about more complex concepts and abstract notions.
  • DLSS – Deep learning super sampling (DLSS) is a family of real-time deep learning image enhancement and upscaling technologies developed by Nvidia that are exclusive to its RTX line of graphics cards, and available in a number of video games.
  • Generative AI – Made famous by the likes of text-generating chatbots such as ChatGPT, generative AI is a conversational technology that can analyse a vast amount of data. Yet it can accomplish essentially only what it was programmed to do – which is where it differs from AGI. However, generative AI’s out-of-the-box accessibility makes it different from all AI that came before it. Users don’t need a degree in machine learning to interact with or derive value from it; nearly anyone who can ask questions can use it. It can enable capabilities across a broad range of content, including images, video, audio, and computer code. And it can perform several functions in organizations, including classifying, editing, summarizing, answering questions, and drafting new content.
  • GPUs – graphics processing units are specialized processors that can process many pieces of data simultaneously, making them useful for machine learning, video editing, and gaming applications.
  • Machine learning – A branch of AI that allows computer systems to learn directly from examples, data and experience. Increasingly used for the processing of ‘big data’, machine learning is the concept that a computer program can learn and adapt to new data without human interference – it keeps a computer’s built-in algorithms current regardless of changes in the worldwide economy.
  • Magnificent 7 – An update on the popular FANG acronym of 2013 that described mega-cap tech growth stocks Facebook, Amazon, Netflix and Google (expanded to FAANG in 2017 to include Apple). The ‘Magnificent 7’ reflects the new group of high-profile technology-centric mega-cap stocks that have emerged in recent years, which are additionally focused largely on secular technology growth trends such as artificial intelligence, cloud computing, online gaming, and cutting edge hardware and software. The Magnificent 7 stocks are: Microsoft, Amazon, Meta (Facebook), Apple, Alphabet (Google), Nvidia, Tesla.
  • Natural language processing – A subfield of computer science, information engineering, and AI concerned with the interactions between computers and human (natural) languages, in particular how to program computers to process and analyse large amounts of natural language data. It is one of the tools used in Siri, the voice-controlled digital assistant. Systems attempt to allow computers to understand human speech in either written or oral form. Initial models were rule or grammar based but couldn’t cope well with unobserved words or errors (typos).
  • Quantum Computing – Quantum computing is a type of computation that harnesses the collective properties of quantum states (based on quantum mechanics), such as superposition, interference, and entanglement, to perform calculations. The devices that perform quantum computations are known as quantum computers – sometimes referred to as supercomputers.
Thematics AM, Mirova and Ossiam are affiliates of Natixis Investment Managers, and form part of our Expert Collective.
1 Source: BBC, 2023,
2 Source: New York Times, 2023,
3 Source, Financial Times, 2023,
4 Source: The Guardian, 2023,
5 Source: Reuters, 2023,
6 Source : New York Times, 2023,
7 Source: Morgan Stanley, 2023,
8 Source: The Guardian, 2023,
9 Source: Fast Company, 2023,
10 Source: Bloomberg, April 2023,
11 Source: Morningstar, August 2023,
12 Source: Investorplace, March 2022
13 Source: Bloomberg and Thematics AM, as at 11 October 2023
14 Source: Bloomberg and Thematics AM, as at 11 October 2023
15 Source: NASDAQ, October 2023,

This communication is for information only and is intended for investment service providers or other Professional Clients. The analyses and opinions referenced herein represent the subjective views of the author as referenced unless stated otherwise and are subject to change. There can be no assurance that developments will transpire as may be forecasted in this material.

The provision of this material and/or reference to specific securities, sectors, or markets within this material does not constitute investment advice, or a recommendation or an offer to buy or to sell any security, or an offer of any regulated financial activity. Investors should consider the investment objectives, risks and expenses of any investment carefully before investing. The analyses, opinions, and certain of the investment themes and processes referenced herein represent the views of the portfolio manager(s) as of the date indicated. These, as well as the portfolio holdings and characteristics shown, are subject to change. There can be no assurance that developments will transpire as may be forecasted in this material. The analyses and opinions expressed by external third parties are independent and does not necessarily reflect those of Natixis Investment Managers. Past performance information presented is not indicative of future performance.