Daniel Leb is looking for AI stocks outside of the US. What has he recently added to his portfolio?

Daniel Leb's hedge fund Third Point posted modest returns in the third quarter, but increased its bets on the future of technology. In a letter to investors, Leb described new positions in companies benefiting from the AI boom and warned: the scale of investment in AI infrastructure is already comparable to the size of the global debt market - and this could become a new area of risk.

In the third quarter of 2025, Third Point's assets returned 3.2%, slightly outperforming the CS HF Event-Driven Hedge Fund Index (3.1%), but behind broad benchmarks: the S&P 500 added 8.1% over the same period and the MSCI World added 7.4%. Leeb acknowledged that the fund's performance was below expectations due to weakness in several large event-driven positions - investment rates designed for corporate transactions and restructurings. Among them is US pharma maker Kenvue, which has come under fire from the Donald Trump administration for linking its popular painkiller drug Tylenol to the development of autism.

According to Leeb, the quarter's results reflect the structure of the market itself: growth in 2025 was concentrated in a limited number of assets - primarily gold and AI-related stocks. Outside of these segments, movement remained uneven. "We are seeing structural issues in many sectors," Leeb notes in a letter to investors, "and our portfolio of short bets on individual stocks has benefited from localized weakness in consumer discretionary, healthcare, information services and software.

Top contributors to the fund's result were semiconductor manufacturer TSMC, artificial intelligence graphics processor developer Nvidia, building materials supplier CRH, engineering and installation holding company Comfort Systems USA and utility company Pacific Gas and Electric.

The worst performing positions in Third Point's portfolio for the quarter were health products manufacturer Kenvue, transportation and logistics group DSV, bottled water manufacturer Primo Brands, financial services company London Stock Exchange Group and online betting operator Flutter Entertainment.

Artificial intelligence has become a key focus of Third Point's strategy. Leeb describes it as "the main driver of the technology cycle," driving the fund's investment decisions and shaping new growth opportunities.

After Chinese company DeepSeek's breakthrough in training large language models, the fund had concerns that the increased efficiency of the new systems would reduce the need for computation, it said in a letter to investors.

"In reality, this has not turned out to be the case. Despite significant efficiency gains in existing models, new AI capabilities and architectures have more than offset these savings, leading to a significant acceleration in demand for AI computing power," Leeb writes.

Third Point is betting not only on AI developers, but also on the infrastructure that supports technology development - from semiconductors to data center equipment. According to Leeb, the growing demands of new models for computing power strengthen the position of companies supplying this infrastructure.

The beneficiaries of AI computing growth remain TSMC and Nvidia, which Third Point holds in its portfolio as "two key links" in building a new technology architecture.

That said, Third Point is looking for undervalued opportunities outside the US. In the third quarter, the fund added SK Hynix and Ebara to its portfolio, companies that Leb estimates will benefit from the growing demand for AI infrastructure while trading at a significant discount to U.S. peers.

South Korea's SK Hynix is one of the three largest memory chip manufacturers in the world - along with Samsung and Micron, according to a letter to investors. The company leads in the segment of high-speed HBM memory, which is used in graphics processors and gas pedals for artificial intelligence. It accounts for more than half of the global market, and HBM revenue already accounts for about 40% of the semiconductor business.

Despite its technological leadership, SK Hynix securities remain markedly undervalued, with a forecast P/E for 2026 of about 7 versus 10-12 for its competitors, according to Third Point. The industry's move to a new generation of HBM4 memory will make the technology even more challenging. This, according to Third Point, will support product prices and strengthen SK Hynix's position in the industry.

Ebara, a Japanese manufacturer of systems for the semiconductor industry, derives more than half of its profits from supplying equipment for chip manufacturing. The company produces chemical-mechanical polishing (CMP) units needed to process wafers for chip creation. Its equipment is in demand in advanced semiconductor manufacturing, and the market for metal layers, where Ebara competes with Applied Materials, could grow as the industry shifts to finer process standards.

According to Leb, as processor architectures become more complex and demand for precision processing grows, the company will be able to increase its market share. Currently, Ebara shares are trading at a significant discount to its US peers and, according to Third Point, the company remains one of the most undervalued in its segment.

"We are in active dialog with the company's new management team to drive shareholder value," Leeb notes in the letter. - "We are discussing how to increase the company's profitability by more than 50%, narrow the valuation gap compared to competitors, simplify the asset structure, expand the share repurchase program and make management more transparent.

Separately in his letter, Leeb notes investments in Elon Musk's ecosystem. Third Point's loan portfolio includes X (formerly Twitter) and xAI, an artificial intelligence technology company. According to Leeb, under Musk's stewardship, X "has seen an extraordinary turnaround for the better," and xAI remains one of the most interesting ideas at the intersection of technology and capital.

Leeb notes that the field of artificial intelligence is still in its early stages and requires large investments. However, he says X and xAI securities are "extremely cheap" given Musk's resources, backing and access to funding.

On November 6, Tesla shareholders will consider formalizing and expanding its investment relationship with xAI. Third Point expects the resolution to be approved and believes Tesla's investment grade rating and $1.5 trillion capitalization will be a catalyst for the revaluation of xAI's debt instruments.

Leeb views current market conditions as persistently positive. According to him, the growth in shares of leading companies is "largely underpinned by strong earnings performance", the recession remains "well over the horizon", and the Fed's rate cut cycle, together with continued investment in artificial intelligence, form a favorable backdrop for the markets.

The investor notes that the main growth in the market is provided by a few leaders. Third Point believes that in the coming months it will become clear whether this growth will expand to a wider range of companies or remain concentrated - and therefore more risky.

Leeb also draws attention to the scale of what's happening in artificial intelligence infrastructure. "The amount of investment in power capacity and data centers is staggering," he notes in the letter. - But it's important to understand how the market will handle this growth."

Leeb believes investment in energy capacity and data centers has reached unprecedented scale. "The combined market for high-yield bonds and high-risk loans is about $3 trillion, while the capital being discussed for AI infrastructure development exceeds that figure many times over," he writes. Leeb notes that supply and demand are unlikely to develop in sync. He compares the current boom to the Internet boom of the early 2000s, when it seemed that demand for fiber-optic networks would be endless - until a technological breakthrough in data transmission led to overcapacity.

Leeb does not rule out that similar risks exist today: breakthroughs in quantum computing or advances in chip manufacturing could quickly devalue current investments in artificial intelligence. At the same time, widespread substitution of human labor could displace jobs on a scale that risks triggering a deep recession. "However the scenario plays out, the implications - both risks and opportunities - will be enormous," Leeb concludes.

This article was AI-translated and verified by a human editor