Why Google is getting the cut from Australia's first AI-driven ETF

AI remains the all-conquering market thematic, even if concerns are starting to bubble to the surface. But beyond the obvious investment opportunity, AI technology is also working its way into investing research and strategy itself.

And nowhere is that more evident than the VanEck Dynamic International Equity ETF (ASX: GOAT) - the first AI-powered ETF on the ASX that offers exposure to 150 global stocks selected by a proprietary AI engine. 

Here, VanEck's Head of Investments and Capital Markets Russel Chesler talks through the recent moves in the portfolio, including one big sell and two key overweights, and why AI inevitably remains the biggest topic of discussion.

1. What was the most notable addition to the portfolio this month and why?

Lam Research (NYSE: LRCX) was the largest addition at the latest monthly rebalance, entering at 1.5% of the portfolio. Profitability and quality were among the leading model signals supporting its inclusion.

The broader pattern was also notable. Allstate (NYSE: ALL), 3M (NYSE: MMM), Monster Beverage (NASDAQ: MNST) and Booking Holdings (NASDAQ: BKNG) joined the portfolio, with profitability and quality among the two strongest signal categories for each. For Allstate, it was the strongest.

The common thread across these different businesses was the importance of profitability and quality in their selection. GOAT has no permanent commitment to a particular investment style. Its exposure to quality can increase when those characteristics receive stronger support from the model and change again as other opportunities emerge.

2. What was the most notable sell or downsize in the portfolio this month and why?

Alphabet (NYSE: GOOG) was the largest sale, exiting from a 5% position. Its overall model score fell from the 92nd percentile to the 60th over the month, making it substantially less competitive against other eligible companies. Support had weakened across most signal categories, leaving macroeconomic indicators as the only positive contributor to its score.

Energy and materials allocations also fell. Energy declined from 13.5% to 9.3%, with thirteen mostly smaller, non-US energy companies removed, while materials declined from 6.2% to 4.5%.

Alphabet illustrates an important part of the discipline: a large holding must continue to earn its place at each rebalance. When its relative ranking deteriorates sufficiently, the process reallocates to companies with stronger support.

3. What’s your most notable overweight and why?

Micron Technology (NYSE: MU) is the largest overweight stock, with a portfolio weight of 4.1%, up from 2.8% at the latest rebalance and close to three times its benchmark weight.

Its overall model score ranks it 77th among roughly 1,250 eligible companies. The strongest support comes from macroeconomic signals, with share-price trends also contributing positively.

Micron was among September’s strongest contributors. Its record fiscal 2026 revenue of US$133.2 billion, announced at month end, also illustrates the scale of demand supporting the memory industry.

Alongside ASML (NYSE: ASML), the second-largest overweight, Micron gives the portfolio meaningful exposure to the infrastructure behind AI. That is a useful distinction for investors as the portfolio’s AI exposure extends beyond the major platforms to the businesses supplying the technology they depend on.

4. What’s your most notable underweight and why?

Communication services is the largest sector underweight following the latest rebalance. The portfolio holds no companies in the sector, leaving it 7.7 percentage points below the benchmark. Alphabet and Fox (NASDAQ: FOX) exited, while Meta (NASDAQ: META) remained outside the portfolio.

That positioning results from individual company assessments within the portfolio construction rules. It also carries an opportunity cost.

A systematic process will sometimes miss strong performers, and this was one example. The aim is to apply the selection discipline consistently over time. At each rebalance, sector weights are constrained to within 10 percentage points of the benchmark, limiting the size of sector deviations.

5. What’s been one of your most notable performers over the month?

Bloom Energy (NYSE: BE) was September’s largest contributor to relative performance, adding around 28 basis points as its shares rose approximately 38%. The portfolio held around 0.9%, compared with roughly 0.1% in the benchmark.

Bloom makes solid oxide fuel cell systems that generate electricity on site, giving it a clear connection to the power requirements of expanding AI infrastructure.

The model’s positive assessment drew on company fundamentals, including changes in long-term operating assets, alongside balance sheet and valuation signals. Its overall score rose from the 96th to the 98th percentile, leaving it among the highest-ranked companies in the portfolio.

The notable point is that Bloom continued to rank strongly after the rally. A rising share price alone does not determine whether the process retains a holding.

6. What themes and trends are dominating discussions right now?

The two main themes dominating discussions at the moment are the cost of capital and the broadening of AI investment.

Higher borrowing costs put more emphasis on whether companies can finance investment from their own earnings and cash flows. That makes profitability, balance sheet strength and refinancing needs important considerations. Within GOAT, profitability and quality now provide the strongest overall model support for the portfolio.

The AI discussion is also extending across to the infrastructure required to support it. September’s leading contributors included businesses exposed to memory, storage, servers and on-site power.

Looking ahead, the question is how well companies can convert investment and demand into sustainable earnings while financing remains expensive.

GOAT’s positioning can change as the model reassesses those opportunities at each monthly rebalance.