What Is BAI? iShares A.I. Innovation and Tech Active ETF

Add up the memory and storage names and the character of the portfolio becomes clear. SK Hynix at 8.0%, Micron at 6.6%, SanDisk at 3.3% and Western Digital at 3.2% come to 21.1% of the fund in companies that make chips or drives that store data. Add Lam Research at 5.2%, which makes the etch and deposition tools those chips are built on, and Taiwan Semiconductor at 4.5%, which fabricates for everyone. NVIDIA, the company most people associate with the phrase artificial intelligence, sits at 3.8%, less than half the weight of SK Hynix.

There is a coherent thesis behind that. Training and serving large models consumes high-bandwidth memory in quantities that were not planned for, and memory is a supply-constrained, capital-intensive business with only a handful of credible producers and real pricing leverage when supply is tight. A manager who believes the bottleneck in AI infrastructure sits at memory and manufacturing equipment rather than at the processor designers would build something close to this portfolio. Whether that view is right is exactly the question a buyer is taking a position on.

It also means the fund carries a risk profile most AI buyers are not expecting. Memory is the most cyclical corner of semiconductors, with a long history of violent boom and bust as capacity arrives late and all at once. A fund weighted this way tracks the memory pricing cycle closely. That cycle has historically had little to do with the pace of software adoption.

There is no index here. The manager selects holdings, and the portfolio can look entirely different in a year. That flexibility is the argument for the 0.55% fee, which is several times what a broad technology index fund costs and represents a real hurdle: the manager must add that much value every year simply to match a cheaper passive alternative. It is a defensible fee for a genuinely differentiated portfolio, and this one is differentiated, but the arithmetic does not go away.

The 2024 launch matters more than it looks. The fund has not existed through a full semiconductor cycle, let alone a broad technology drawdown. Any judgement about how the strategy behaves in stress has to be made from the composition rather than from its record, because the record is too short to carry weight. Reaching $15.7B in that time is a comment on demand for AI-labelled products rather than on the strategy itself.

One consequence of active management deserves flagging. Because holdings turn over at the manager's discretion, the exposure you researched is not necessarily the exposure you own six months later. That is the point of paying for active management, but it does mean the position needs periodic checking rather than the set-and-forget treatment an index fund allows.

The overlap question is unusually important for this fund. Anyone holding a total US market fund or an S&P 500 fund already owns NVIDIA, Broadcom, AMD, Intel and Micron at meaningful weights, because those companies have grown into some of the largest listed businesses in the world. Adding BAI does not introduce new companies so much as it multiplies the exposure to a handful of names already sitting near the top of the core portfolio.

What it does add that a US index fund does not is SK Hynix, the Korean memory maker at the top of the fund, and direct exposure to Taiwan Semiconductor at 4.5%. Those are foreign-listed businesses that a domestic index fund misses entirely. If the reason for buying is access to the parts of the AI supply chain that sit outside the United States, that is the honest version of the argument.

Sector data confirms there is nothing else in here. Technology is 91%, with token weights in industrials at 5%, communication services at 3% and consumer discretionary at 2%. This is a concentrated sector bet, and it should be sized as one rather than treated as a diversified growth holding.

Approximate weights as of August 2026. Each ticker links to its individual stock guide in Walnut.

There are three common ways to get BAI exposure. Buy shares (or fractional shares) of BAI directly at any major broker that lists it. Hold it as a core position and layer more concentrated ideas on top. Or build it into a thematic portfolio in Walnut, so BAI sits alongside other holdings that express the same thesis, with target weights you can rebalance toward. BAI trades like a stock during market hours, so you buy it the same way you would any listed share.

New to buying funds? See how to buy an ETF, step by step.

Whether BAI is a good buy depends less on any single call and more on your time horizon and what you already hold: it tracks Actively managed, no tracked index, so the real question is whether you want that exposure in your mix and at what weight. We walk through valuation, concentration, and what would have to be true for it to outperform from here in is BAI a buy?

BAI gives you Actively managed, no tracked index exposure in one ticker at a 0.55% expense ratio. Most investors use it as a core holding and layer more concentrated thematic portfolios on top.

Whether BAI is worth buying today depends more on your time horizon and what you already hold than on any single call. We walk through valuation, concentration, and what would have to be true for it to outperform from here in is BAI a buy?

BAI yields 1.13% as of August 2026, paid by passing through the dividends of its underlying holdings. For the payout schedule, history, and how the distributions are taxed, see BAI dividend: yield and schedule.

New to funds like BAI? Start with what an ETF is, then how to buy an ETF, or browse the full guide to ETF investing.