Beyond Big Tech: The Increasing Specialization of AI ETFs

Harbor Capital's AI Labs Ecosystem ETFs

What if you wanted to bet on one particular AI model outperforming its competitors? Companies behind models such as Claude, Gemini, Llama and ChatGPT are constantly releasing new versions, publishing benchmark results and competing to develop increasingly capable systems, with artificial general intelligence (AGI) representing a longer-term ambition for several of them.

The challenge for investors is that not all these companies are publicly traded. Even when they are, the AI model may represent only one part of a much larger business. But investors don't necessarily need to own the company developing the model to gain exposure to its potential success.

Each AI lab operates within a broader ecosystem of companies supplying the infrastructure, technology and services needed to develop, deploy and commercialize its models. That's the opportunity Harbor Capital is attempting to capture with its AI Labs Ecosystem ETF suite.

Investors currently have five options: Harbor Anthropic AI Lab Ecosystem ETF

The easiest way to understand the concept is to picture each AI lab as the hub of a wheel, with the spokes representing the publicly traded businesses supporting its operations.

Those spokes can include semiconductor designers and hardware manufacturers, cloud computing providers, networking equipment suppliers, data services companies, enterprise software platforms, application developers, consulting and systems integration firms, and businesses supplying power and cooling infrastructure. The relationships can also extend to companies distributing AI models, incorporating them into commercial products or helping enterprise customers deploy them.

Each ETF attempts to map these relationships into an investable portfolio centered on a particular AI ecosystem. This gives investors considerable flexibility. Someone particularly bullish on Anthropic could concentrate on ANTW, while an investor expecting Google DeepMind or OpenAI to capture a larger share of the market could choose DEPW or OAIW. Alternatively, investors could combine several funds and assign their own weights to different AI ecosystems.

There will inevitably be overlap. Major AI developers often rely on many of the same semiconductor manufacturers, cloud infrastructure providers and electrical equipment suppliers. A company supplying several competing AI labs could therefore appear in multiple portfolios.

Nevertheless, the underlying approach offers a degree of differentiation from conventional AI ETFs. Rather than treating AI as one broad investment theme, Harbor's suite allows investors to express views on the competitive positioning of individual AI labs through the surrounding publicly traded businesses.

The important caveat is that success for a particular AI model won't necessarily translate into superior returns for its ecosystem ETF. Suppliers may serve competing labs, their AI-related revenues may represent only a fraction of their businesses, and valuations can already reflect growth expectations.

KraneShares' Asian AI ETFs

Most of the world's largest AI hyperscalers are based in the United States, but investors shouldn't overlook the other side of the Pacific. Asia plays an important role in the AI supply chain, from semiconductor fabrication and memory manufacturing to the development of competing models.

China is particularly interesting. Despite U.S. sanctions and export controls restricting access to advanced chips and semiconductor manufacturing equipment, Chinese AI developers have produced models that perform competitively on several industry benchmarks. Companies such as DeepSeek and Alibaba have demonstrated that sophisticated AI systems can be developed under tighter hardware constraints, even if questions remain about their capabilities, costs and commercial adoption.

China also increasingly views AI development as a strategic priority, with implications for technological competitiveness, economic growth and national security. That has encouraged domestic investment in semiconductors, computing infrastructure and alternatives to American technology. Of course, export restrictions, geopolitical tensions and the possibility of further sanctions remain significant risks.

For investors seeking dedicated Asian AI exposure, KraneShares is worth examining given its longstanding specialization in Chinese and broader Asian equity markets. Those interested specifically in China's domestic technology industry can consider the KraneShares SSE STAR Market 50 Index ETF (KSTR).

It tracks the SSE Science and Technology Innovation Board 50 Index, which selects 50 large and liquid companies listed on the Shanghai Stock Exchange's STAR Market. The STAR Market was established to support innovative Chinese businesses, including companies operating in semiconductors, biotechnology, advanced manufacturing and information technology.

If the Hang Seng TECH Index is loosely comparable to the Nasdaq-100, the STAR 50 can be thought of as an attempt to capture the next generation of Chinese technology leaders. It's a concentrated portfolio, but one that provides access to domestic Chinese A-shares that many U.S. investors would struggle to purchase directly. Many are relatively young, research-intensive companies that would otherwise receive little representation in conventional Chinese equity benchmarks.

One of its largest holdings is Cambricon Technologies, at approximately 10% of the portfolio. Cambricon develops AI processors and related computing technology, making it a prominent domestic contender as China attempts to reduce its dependence on foreign chip suppliers. There's also an approximately 2% allocation to CXMT reflecting China's efforts to develop its domestic memory industry.

But Asia's semiconductor expertise extends well beyond China. Taiwan dominates advanced contract chip manufacturing, South Korea is a major supplier of high-bandwidth memory, and Japan remains important in semiconductor materials, equipment and precision manufacturing.

Investors seeking exposure to that broader regional supply chain can consider the KraneShares Asia AI Technology ETF (KAIT). Its holdings look substantially different from KSTR's. Among its three largest positions are Taiwan Semiconductor Manufacturing, SK Hynix and Samsung Electronics, companies occupying critical positions in the production of advanced processors and memory.

That makes the distinction between the two ETFs fairly straightforward. KSTR offers a concentrated bet on China's domestic technology development and semiconductor self-sufficiency, while KAIT provides broader exposure to Asia's established AI hardware and technology supply chain.

Both offer diversification away from the familiar U.S. hyperscalers, although their risks differ considerably. KSTR is more directly exposed to Chinese policy, sanctions and domestic technology commercialization, whereas KAIT is more sensitive to global semiconductor demand, the capital spending cycle and the competitive dynamics of chip manufacturing.

Please note that this article reflects the author’s personal views and does not represent the opinions of the publication or its affiliates. It is for informational purposes only and does not constitute investment advice. It is essential to seek guidance from a registered financial professional before making any investment decisions.