Harbor Capital Advisors just filed with the SEC for five new exchange-traded funds, each designed to track the commercial ecosystem surrounding a single private AI laboratory. The funds would trade on NYSE Arca and represent one of the most direct attempts yet to let everyday investors ride the AI wave without needing a venture capital allocation.
The five proposed products are the Anthropic AI Lab ETF, Google DeepMind AI Lab ETF, Meta AI Lab ETF, OpenAI Lab ETF, and xAI AI Lab ETF. Each is structured as an actively managed fund, meaning portfolio managers will pick and weight holdings rather than passively tracking an index.
What the Lab ETFs actually do
Here’s the thing: most of the AI labs driving the current technology cycle are private. OpenAI, Anthropic, and xAI don’t have publicly traded shares. Google DeepMind sits inside Alphabet, and Meta AI is embedded in Meta Platforms, but the pure research arms themselves aren’t separately investable.
Harbor’s workaround is to target each lab’s “ecosystem.” In English: rather than buying shares in the lab itself, the fund would hold publicly traded companies whose products, partnerships, or revenue streams are meaningfully tied to that lab’s technology. Think cloud providers running inference for OpenAI, chipmakers supplying Anthropic’s training clusters, or enterprise software firms embedding a specific lab’s models into their products.
The filing landed on May 22, 2026, under the Harbor ETF Trust. All five funds would list on NYSE Arca, the exchange that already hosts the bulk of US-listed ETFs.
Harbor’s expanding AI playbook
This isn’t Harbor’s first move into AI-themed investing. The firm previously launched the Harbor AI Inflection Strategy ETF, trading under the ticker EPAI, in the late 2025 to early 2026 window. That product took a broader approach to AI exposure across the sector.
The Lab ETFs represent a deliberate narrowing of focus. Instead of one fund covering the entire AI landscape, Harbor is slicing the market into five distinct bets organized around the labs that are actually producing frontier models.
What this means for investors
The actively managed structure means higher fees (though Harbor hasn’t disclosed expense ratios yet) and a reliance on the portfolio manager’s judgment about which ecosystem companies actually matter.
There’s also the concentration question. Each fund orbits a single lab’s technology stack. If that lab stumbles, whether through a failed product launch, a safety controversy, or simply falling behind on model capabilities, the entire fund takes the hit. Diversification across five separate funds is possible, but at that point you might be paying five sets of management fees for something that looks a lot like a broad AI sector fund.