By Reshma Kapadia

Sonu Kalra has managed the Fidelity Blue Chip Growth fund since 2009. Today, the $88 billion fund is filled with the bull market's winners, including Magnificent Seven stocks such as Amazon.com, Alphabet, Meta Platforms, Microsoft, and Nvidia. Although Kalra is constantly on the lookout for developments that could interrupt the growth of artificial intelligence, which has propelled many of these companies and their shares, he believes that it's early days for the technology, and the stocks have much further to run.

Kalra often scouts for investment ideas and battle-tests them by monitoring the habits of his teenage children. "I'm a big believer that the future is being shaped by younger consumers who are early adopters of technology and trends that can drive sustained growth," he says.

The strategy has enabled Kalra to spot companies with the potential to become much bigger, often by expanding into adjacent markets. He likes to invest in companies as they begin to show improving returns on equity and invested capital.

Fidelity Blue Chip Growth has benefited from this approach; it returned more than 18%, on average, over the past 15 years, outperforming 96% of its peers, according to Morningstar.

Barron's spoke with Kalra in early October and again via email at the end of the month about ChatGPT's "iPhone" moment, Nvidia's winning streak, Amazon's e-commerce opportunities, and more. An edited version of these conversations follows.

Barron's : Many investors, like you, have been captivated by AI-related stocks, although there are now concerns about the amount of AI spending. How do you envision the technology's trajectory?

Sonu Kalra: AI is the fourth tech wave I have seen in my 25-plus-year career: There was the internet in the late 1990s, the smartphone in the 2000s with the iPhone's launch in 2007, and then cloud infrastructure in the 2010s. This latest wave, generative AI, really started with ChatGPT's launch in November 2022. That is when AI came to the masses and had its iPhone moment.

I think of generative AI as converting investment dollars and energy into intelligence. The goal is artificial general intelligence. There are about 100 million knowledge workers in the U.S. and about one billion globally. If you assign some sort of value to each worker -- let's say $100,000 to make the math simple -- and you believe that AI can drive 10% productivity, that implies $1 trillion of value in the U.S. and $10 trillion globally. That is why I'm a believer.

What parallels, if any, do you see to the dot-com bubble of the late 1990s?

In the late 1990s and early 2000s, dot-com-related stocks went through a correction and digestion phase. But in the next 20 years, the internet became embedded in everything we do, creating value over time. That is the road map I'm using to think about generative AI.

I am a believer in [futurist Roy] Amara's Law: We tend to overestimate the impact of technologies in the short term but underestimate them in the long term. I use Netscape, the internet-browser company, as an example. It went public in 1995 and brought the internet to the masses -- similar to ChatGPT. That was the starting line [of the internet cycle].

Back then, it was difficult to see the second- and third-order effects of the internet -- that the value of taxi medallions would drop 90%, for instance, or that we would feel comfortable sleeping in strangers' homes. But that is what happened with the rise of Airbnb and Uber, which raised its Series A funding in 2010, 15 years after Netscape marked the beginning of the cycle.

Some of today's concerns center on circular funding, or companies funding one another to grow. Does this remind you of the 1990s?

Yeah. We have to understand how the dollars are flowing from one entity to another. It doesn't mean these [funding] deals will end badly, but it is something you have to watch. The market will determine whether the investments made by the biggest tech companies are justified, either by the revenue generated or the productivity that results. One way or another, we need to see some tangible return on that investment. That is the concern in the marketplace, and it is healthy.

Which companies are winning so far?

Right now, I would put Nvidia at the top of the list. It is the leader in providing the infrastructure, or the picks and shovels for the AI buildout. In terms of use cases, Meta Platforms has been vocal about the benefits of AI in improving the efficiency of its ad platform and driving engagement. Meta's revenue growth has accelerated in the past couple of years.

Alphabet has the largest employee base of AI researchers. A lot of technologies were invented at Google. The majority of its revenue has come from search, and it is trying to bridge the gap to the world of generative AI. Alphabet has its own AI models. Gemini is most prominent, but it also has other AI models for video.

What would make you reassess the prospects for these and other AI-related companies? Market-share declines would be a warning sign. These companies are also investing a lot not just in capex but also in hiring, so we need to see profitability come through. I consider them well positioned. One of their advantages is incumbency. A large installed base of users and data is the currency needed for success in generative AI, which is different from previous tech cycles.

Nvidia has been affected by U.S. restrictions on China's access to advanced technology. Do you see other geopolitical risks?

For Nvidia, it can't get worse: its sales to China went to zero in the past six to 12 months. As an investor, I have already taken the pain. If the administration can find a way that enables Nvidia to have a healthy business in China, that would be great. But it's a "nice to have," not a "must have."

Nvidia has been investing in several "neocloud" companies such as CoreWeave, which in turn have been buying Nvidia chips. Is this a smart approach?

These specialized cloud providers are focused on building AI infrastructure and providing cloud-based AI services. Google, Amazon, and other tech companies have chips that compete with Nvidia's, so Nvidia is making its products available through as many distributors as possible. But we need to make sure the neocloud companies are generating the revenue they expected.

Neocloud companies spend billions to build out data centers, and then the customers come in. But neoclouds get paid monthly over several years. Cash flow is positive over the life of the contract, but it takes time to reach the break-even point. What is your framework for assessing Nvidia's valuation?

Nvidia is the leading semiconductor and infrastructure provider for training large language models and inferencing. It offers a complete stack of solutions including chips, systems, software, and networking, along with best-in-class performance at the leading edge. We are still early in the transition from data centers running on traditional CPU [central processing unit]--based architecture to an AI-based architecture running on GPUs [graphics processing units].

Nvidia's stock is trading at less than 30 times 2026 consensus estimated earnings per share, below the 35 times at which it traded in late 2022 when ChatGPT was launched. The company has grown revenue more than 60% in 2025.

What are the potential constraints on AI?

Electricity can be a constraint to growth and is something our team is monitoring. There have been a lot of deals around power and access to power over the past five months. I watch the power announcements being made, how long it takes to bring the power on-line, and whether it is enough to power these data centers. There are a lot of power investments that need to take place.

What I am monitoring most closely are use cases. I want to see this technology become embedded in our everyday lives. In what use cases can AI deliver experiences that we couldn't experience before? OpenAI has recently signed partnerships with applications such as Expedia and Booking.com for travel-related tasks; Zillow for searching real estate listings; Canva and Figma for creating content and collaboration; Walmart, Shopify, and Etsy for shopping-related tasks. It has also announced it will be integrating DoorDash, Uber, PayPal, and OpenTable in the future.

But we want to see if AI can complete tasks on our behalf. Coding is a successful use case. AI could also be used to fundamentally transform customer service. As technology and AI mature, the customer experience may improve, with AI delivering faster and more accurate resolutions while simultaneously driving down operational costs for companies.

What is one of your biggest holdings that has momentum outside of the AI theme?

Amazon is still the leader in e-commerce, and e-commerce is still less than 20% of U.S. retail sales. Incremental penetration of the market provides a secular tailwind for Amazon. Half of every incremental retail dollar is now going into e-commerce.

Amazon trades at a discount to retailing peers Walmart and Costco; it has a price/earnings ratio of 32 times 2026 consensus earnings-per-share estimates, compared with Costco at 42 times and Walmart at 35 times. Amazon has been growing revenue by 12% in 2025, compared with 8% revenue growth for Costco and 4% for Walmart.

Where are you finding innovation outside of tech?

SharkNinja is a home-appliance company that has been hurt by tariff concerns because much of its supply chain is in China. But it is bringing innovation into a category that hasn't seen much of it. People may know about the SharkNinja blender or vacuums. The company is introducing new products such as the Creamy Ice Cream machine. My daughter got one, and now we make protein ice cream at home.

(MORE TO FOLLOW) Dow Jones Newswires

November 06, 2025 00:30 ET (05:30 GMT)

Copyright (c) 2025 Dow Jones & Company, Inc.