Nvidia, Microsoft, and Other AI Winners Will Drive a New Tech-Spending Cycle

Franklin Technology, which trades in Luxembourg, returned more than 54% in 2023, thanks to substantial bets on most of the so-called Magnificent Seven, a group of top-performing tech companies. Shares of chip maker
Nvidia,
the fund’s single largest position, are up another 25% so far in 2024. Curtis sees more gains ahead for both the stock and the sector, driven by his aggressively bullish view on artificial intelligence.

Barron’s recently caught up with Curtis to discuss his favorite stocks and the future of AI. An edited version of the conversation follows. 

Barron’s: Tech stocks soared in 2023 and are off to a strong start in 2024. How does the rest of the year look to you?

Jonathan Curtis: Certainly, 2023 was a magnificent year, led by the Magnificent Seven [
Alphabet,

Amazon.com,

Apple,

Meta Platforms,

Microsoft,
Nvidia, and
Tesla
]. We own six of the seven. They are high-quality businesses, with good balance sheets and high levels of profitability, and have been right-sizing costs. Most have a compelling AI [story]. We expect the strength we have seen in the Magnificent Seven to start broadening out this year, as we believe this is the start of a new business cycle in tech. 

And AI is at the heart of it.

All of the companies we invest in either build on the cloud or are suppliers to cloud businesses, and for the most part have reported either stabilization or acceleration in their business. We went to the Amazon AWS Re:invent conference late last year, and boy, everybody was bullish. There is a good reason to invest in 2024: the promise of AI. It is a board-level priority. That says to us that we’re at the start of a business cycle in tech, and we think it is going to be a profound one, because the opportunity in AI is so significant that it touches every bit of creative work. It will start slow and then broaden out. 

From a stock-performance perspective, the Magnificent Seven will do well, but we are underweight them by about 20 percentage points in our strategy. We are overweight everything else. We are putting our money on this broadening-out thesis, and tilting more into enterprise stocks and less toward consumers.

When will we see AI-related revenue show up in the numbers for companies like Microsoft,
Adobe,
and
Salesforce
? There are some worries it could take longer than investors might like.

That’s correct. Microsoft CEO Satya Nadella, CFO Amy Hood, and others are trying to temper our enthusiasm, and appropriately so. But we’re already seeing AI in the build-and-experimentation phase. The clue that spending is broadening out will be Microsoft Copilot for Office. We will start seeing that in the second half of 2024. Then we will see if knowledge workers take to this.

Will they?

I’ll use my own team as an example. I’m all over this stuff. I use it all the time; I love it. There are others on my team who think it is weird. They don’t know how to use it. Once Copilot gets rolling, I expect knowledge workers to say, “Wait a minute. I need these capabilities in every enterprise task I engage in.”

That’s a lot of pressure on Microsoft.

It is, and on that one number, Copilot adoption. Microsoft has provided some numbers, such as how many people are doing trials. But I doubt that Microsoft will give us data on measures like acceleration in average revenue per user for Office until the calendar second half.

Nvidia was your largest position at year end and has rallied sharply this year. What now?

If we see adoption of Office Copilot, and then it broadens out, the productivity gains for knowledge workers will be profound and the demands on the underlying infrastructure will blow the doors off. 

We’ve seen two important data points on Nvidia recently. Meta CEO Mark Zuckerberg said his company will own 350,000 H100 GPUs [graphics processing units] by the end of 2024, and
Taiwan Semiconductor Manufacturing
has released some important numbers. We know mobile isn’t great, and mobile is a big part of what TSMC does. So, if mobile is underperforming, and TSMC is doing well, then what is outperforming is high-performance computing—and that relates to Nvidia.

The Nvidia move has been substantial year to date, but not crazy. If half a billion knowledge workers use these tools as much as I use them every day, we don’t have enough Nvidia GPUs out there, period. 

But isn’t it possible that we could see a pause if it takes time for people to adopt the use of Copilot and other software?

Sure, that’s possible. But everyone reading this could get value from these tools right away. ChatGPT got to 100 million users in the blink of an eye. Why? Because we all immediately know how to use it. These GPUs are out there, and it isn’t hard to see what the business model is: Sell Copilot licenses to get the best intern you’ve ever had sitting alongside you for $360 a year. Bargain.

Nvidia and Microsoft are the largest positions in the Franklin Tech fund. Amazon was third-largest at year end. While the stock has done well, there is some suspicion the company is behind Microsoft and Alphabet in AI.

There are five things you need to play in the market for the cloud and AI. You need data, and the ability to get more unique data. You need AI engineers. You need the financial model that allows you to wrangle all the computing power you need to do this. And you need a place to deliver the value back to get paid for it. Microsoft was missing one element for a while—AI talent—and it got it through its investment in OpenAI. Amazon was in the same situation, but was also missing the AI engineers—and so it agreed to invest in Anthropic. All of the world’s most advanced software applications are in Amazon’s cloud. New data are showing up there all the time, and the company has an enormous opportunity around shopping. 

Your chatbot personal shopper?

This past Christmas, I went to ChatGPT, explained who my son was and what he liked, and asked for some book recommendations. It isn’t a stretch to think about doing the same thing and including a purchase button. I’ll be shocked if ChatGPT doesn’t add that kind of functionality. Amazon has a number of ways to play things here. It will be just fine. And now it has an advertising engine inside its business, which is highly profitable. Generative AI will allow companies to talk to their customers in a personalized way.

If AI is such a lucrative business for Nvidia, are there other chip stocks that you would buy, as well?

We own
Advanced Micro Devices
for a few reasons. One, it has the Intel architecture on the world’s best process nodes, with Taiwan Semiconductor as a manufacturing partner. And that has made it a market-share gainer. Two, we are at the beginning of another PC cycle, because it has been four years since the start of Covid, and maybe there are reasons to buy AI PCs. And three, AMD has a story around GPUs. OpenAI CEO Sam Altman is trying to build chips everywhere. Hardware companies aren’t trying to get away from Nvidia, but they want to have options. AMD can satisfy some of that demand. 

You have big positions in chip-related names—
ASML
in the semiconductor equipment sector, and
Synopsys
in chip-design software. Why?

ASML is unlike any semiconductor supplier on the planet. It has cracked EUV. 

You mean extreme ultraviolet lithography, a technology used to make some of the most advanced semiconductors.

It is the only player. It effectively has no competition. Some people are exploring alternatives, but today, it’s ASML and EUV.

We have owned
Synopsys
for many years. We got it right that there was going to be an acceleration in demand for design tools, as Moore’s Law slowed. We didn’t understand it, but have benefited from the need for all these specialty chips around AI. As the world gets more complicated, much more simulation is going to be needed.

You mentioned earlier that you own six of the Magnificent Seven. Which one have you avoided?

We don’t own Meta. Although the stock has had an enormous run, the company is facing challenges with regard to the types of content on its platform. Look at things like the impact we’ve seen on emerging democracies, and some of the company’s own data on what has happened to young girls on Instagram. Meta has a safety issue to deal with. It is investing so heavily in AI and large-language models because it sees this technology as a pathway to solving the problem. But it isn’t an expensive stock, and if it can solve the problems and calm regulators, the risk profile for the business could be different. We’re keeping an open mind.

What is your thinking on Apple, which was a top-five holding at year end?

I see two risks and one enormous opportunity. One risk is what’s going on in China right now—a weak macroeconomy and a government that is sending a signal to government employees that they should look at the Huawei phone as an alternative to the iPhone. The other risk is the Department of Justice lawsuit against Google.

The government doesn’t like Google paying Apple to be the default search engine in the iPhone—they see it as anticompetitive. But in some ways, you think, a Google loss could be a win for Google financially and a loss for Apple.

Google would love to stop paying Apple roughly $15 billion a year to be the preferred search engine on the iPhone. There is earnings risk to Apple if that goes away. The opportunity is with large-language models, and fixing Siri, finally. 

Siri looks decidedly underpowered as a chatbot compared with ChatGPT and other AI software. But you think Apple could catch up, potentially, with a little dose of AI?

Not only could it drive a powerful upgrade cycle, but it also might give Apple a way to start charging for it.

You talked about broadening out your holdings. What are some examples of small-cap and mid-cap names that fit the bill?

The next step after buying Copilot licenses is getting your data in order. That leads to
Snowflake,
and we own Databricks on the private side. Another use case is making developers more productive. Software developers have totally embraced this stuff. In addition to owning Microsoft, which owns GitHub, we own
GitLab,
 which provides tools for coders. It has a huge opportunity to raise prices and drive greater productivity.

Thanks, Jonathan.

Write to Eric J. Savitz at [email protected]