By Reshma Kapadia
Brian Barbetta, co-head of technology investing at Wellington Management, oversees a team managing $40 billion. The job routinely puts him in contact with the leaders of OpenAI, Anthropic, SpaceX, and other public and private companies leading the artificial-intelligence revolution. To call him bullish about the technology's potential would be an understatement.
Barbetta co-manages the Vanguard U.S. Growth exchange-traded fund and the $2.7 billion Wellington Global Technology Opportunities strategy, geared to institutional investors. It has returned an average of 37% annually in the past three years, after fees. The analyst and investor sees some of the best investment opportunities in companies central to AI: The actively managed ETF holds heftier stakes than peers in companies such as Nvidia, Taiwan Semiconductor, and ASML Holding.
Barron's spoke with Barbetta in May and via email on June 8 about the likely market impact of SpaceX's pending initial public offering, companies wrongly deemed AI losers, and what might cause him to rethink his sunny stance. An edited version of the discussion follows.
Barron's: Investors are growing more skeptical about the AI trade. What is most misunderstood about AI-focused companies?
Brian Barbetta: People expect the law of large numbers to catch up with companies whose revenue is growing by 15%-plus a year. But they fail to appreciate that companies innovating and creating new markets can grow faster for longer and in a more durable, repeatable manner than the market generally expects.
[ Alphabet's] Google is growing almost as fast today as it was more than 10 years ago when I started covering it because the company continues to innovate, create new markets, and build new products. Revenue grew by more than 21% in the first quarter, compared with 12% in the year-ago quarter. As the company brings AI to its users "to organize the world's information and make it universally accessible and useful," as Google's mission states, we continue to expect strong growth.
In every other technology shift -- mainframe, mobile, internet -- you needed people and companies to adopt the technology. I needed to shop online, play a mobile game, watch Netflix, or use software at work to drive usage. The limit to growth was usage. Now, internet access is nearly ubiquitous globally, and this technology can use itself.
In the past, we ran into overcapacity issues because usage didn't catch up with the buildout. With AI, usage is growing faster than capacity. Capex [capital expenditure] investments are generating strong returns, and the companies are supply-constrained relative to the revenue and profit-producing activity they could otherwise deliver.
Give me an example.
Uber Technologies CEO Dara Khosrowshahi recently said AI is central to product design and engagement. Three-quarters of rides are correctly predicted by Uber's AI systems. Verizon Communications CEO Dan Schulman cited double-digit improvements in customer satisfaction scores and $200 million in energy cost savings from using AI [to optimize its infrastructure] . Companies that use AI are talking about tangible returns on investment.
It's impossible to escape the buzz about SpaceX's IPO. Anthropic and OpenAI have also filed with regulators to sell shares to the public. What impact will these so-called mega IPOs have on the broader stock market?
These companies are building entirely new industries, an area where the market has historically underestimated the durability of growth and the ultimate market opportunity. While some public companies that investors held as proxies for private companies could come under selling pressure, I don't expect a material negative impact on other stocks, partly because the stocks that come public typically have a limited float.
We seek to initiate positions and add to them as IPO-ed stocks trade down, which typically happens when the lockup periods ease [allowing insiders to sell more shares]. More broadly, publicly held technology investments have done significantly better over any measurable time period than the vast majority of private funds focused on technology.
What did you learn from investing in private companies that informs your investments in public companies?
The prices of public securities reflect the known demand for existing AI services but underestimate the potential for what the AI labs are trying to build. Strength in AI stocks and the consistent positive revisions in earnings estimates have been driven by the success of AI thus far. If the private labs achieve just part of what they are aiming to build, we believe significant upside remains in many public stocks.
The other learning: Companies are building new products that will disrupt existing, publicly traded companies that we seek to avoid.
Investors are concerned that many software companies will be rendered obsolete by AI. What is your view?
Some software and service companies will see their role become somewhat marginalized -- and in some cases eliminated. When Verizon says its customer service is getting so much better because of the use of AI, it's reasonable to expect that software companies that help people in customer service could be under pressure. The same holds for consulting firms that offer outsourced coding.
Yet, some companies whose stocks are under pressure due to AI use are going to be beneficiaries of AI. Samsara sells hardware with software embedded. Its customers see a return on investment of more than 800% from its products, which help digitize physical infrastructure. For example, Samsara's cameras used in delivery trucks can help determine when a driver is distracted or drowsy, and deliver real-time alerts. The company's sensors detect fuel usage and aggressive braking, preventing accidents and fraudulent transactions.
The CEO is a tremendous technologist, and is increasing the pace of new-product development to bring more AI solutions into the installed hardware base. Samsara can grow revenue at a compounded annual growth rate of 20% through 2030, resulting in an earnings estimate well ahead of current expectations.
What other companies have been misclassified as AI losers?
Unity Software built the world's leading game engine for mobile gaming and monetizes that through subscriptions and in-game advertising services. CEO Matt Bromberg, whom we have known for a long time from [his former role as chief operating officer at] Zynga, and a new chief financial officer, Jarrod Yahes, whom we knew from prior roles, have repositioned the company to benefit from everything happening with AI.
They are allowing developers to build new content based on existing digital assets, reducing the friction in bringing new games or features, and bringing AI into the part of the business that helps gaming companies with monetization. The company's outlook for the core advertising business for the second quarter included a 50% gain in revenue year over year, excluding some discontinued parts of the business.
The market has said the explosion in creative tools is going to make Unity's game engine less relevant. That misses the difference between using generative AI to play on your computer and building a mobile gaming business.
Broadcom's earnings report sparked a recent selloff in AI-related stocks. Did anything in the report give you pause?
With AI infrastructure stocks up so much since the advent of generative AI, we have frequently seen marked selloffs around specific events as investors wonder if the "cycle is about to roll over." We didn't see Broadcom's earnings as particularly indicative of any shift in the continued strength in AI infrastructure demand.
What is your most contrarian view about the AI cycle?
Shifts in algorithmic architecture can lead to a material shift in the hardware and semiconductors necessary to deliver great results and products. While we remain positive on continued growth in AI demand and are positioned accordingly, we recognize that a scientific breakthrough could cause a shift in the infrastructure necessary to deliver AI services.
What is an example of a possible breakthrough?
Today's models are extremely compute-intensive, particularly as they process more data; handle larger amounts of text, code, images, or other information in a single request; and take on more complex reasoning tasks. That has created extraordinary demand for advanced semiconductors, networking, memory, power, and data-center capacity. Over time, researchers may find ways to make models more efficient, reduce the amount of computation required, or change model architectures in ways that bend the cost curve meaningfully lower.
What would cause you to reassess your bull stance on Nvidia?
Competition. Do we see others sufficiently meeting demand in a way that impacts Nvidia's growth and profitability? We feel confident in the company's position today, but the world changes quickly.
Also, we would reassess if we see demand for AI services taper off. Enterprise adoption of AI services is a critical leading indicator of continued demand for AI services. Public commentary and our real-time data continue to show incredible strength.
Is it time to start investing in the companies that are going to benefit from AI?
There is still such tremendous growth in the core that it appears to be the best part of the investment opportunity. ASML Holding is a critical supplier of the most advanced tools for semiconductor fabrication. It is effectively the only company that can make extreme ultraviolet, or EUV, lithography systems required for leading-edge chips today. The market continues to underestimate the size and duration of the AI investment cycle, and ASML will continue to benefit.
How do you think about the regulatory and geopolitical risk?
It's top of mind. Given the power of some AI models, there's a question of whether governments regulate how this technology is rolled out, if it hurts job creation, and how that is mitigated.
Different administrations will take different views on the degree to which technology needs to be protected. But demand for these companies' services is so strong that the geopolitical [risk] concerning where chips sit may become less relevant over time. For example, many companies in China train their models via clouds in other countries, which isn't something the U.S. has disallowed, so there is a way for China to get access to leading technology even with technology restrictions.
Several members of my team recently visited the fabs [fabrication plants] that Taiwan Semiconductor is building in Phoenix. The geographic diversification of the company's footprint and [other steps taken] to alleviate geopolitical concerns should allow Taiwan Semi to reduce the discount in the stock. The company continues to show it is the manufacturing backbone of the AI era. In the latest quarter, revenue was roughly $36 billion, gross margin was above 66%, and management guided to continued strong growth.
What should policymakers be thinking about in terms of AI?
They haven't fully appreciated the potential risks. I expect there will be more job creation than disruption, but groups have been left behind in the past and I'd want a working group to create programs for those under- or unemployed. From my conversations with world leaders, I don't think their heads are in the right place right now.
The other is safety. These models can do things from a fraud and crime perspective that is fundamentally different from anything in the past. For example, they can be trained to impersonate someone. That is going to accelerate meaningfully. In my family, we have a password to be used to identify ourselves. We need to use the models to prevent others from using them for harm. In terms of defense, the U.S. is already leveraging [AI] to protect our assets around the world, but the threat vectors are increasing, and we need to increase awareness and spending.
Thanks, Brian.
Write to Reshma Kapadia at reshma.kapadia@barrons.com
This content was created by Barron's, which is operated by Dow Jones & Co. Barron's is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
June 11, 2026 01:30 ET (05:30 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments