GOAT: Can AI Pick the Next Market Winners?
Written by: Liam Balintec, Aerina Lee, Martin Kim
Published: 03/08/2026
An index fund managed by artificial intelligence, rebalancing itself every month without a human ever touching it. That’s the pitch behind VanEck’s newest ETF, GOAT (ASX:GOAT). It's a bold claim in an industry historically built around the trade-off between passive investing and the judgement of a fund manager. So where does an AI stock-picker fit within this landscape?
The ETF Landscape
To understand where GOAT lies, it helps to understand how ETFs have traditionally worked. An ETF, or exchange-traded fund, is a basket of securities such as stocks, bonds, or other assets, that trades on an exchange just like an individual share. Instead of selecting individual companies, investors gain exposure to an entire portfolio through a single investment, providing diversification at a relatively low cost.
Today, the ETF landscape is largely dominated by three approaches: passive index tracking, active management and smart-beta strategies.
Passive ETFs remain the foundation of the industry, tracking existing indexes and offering transparency and low fees. However, these funds are limited by the composition of the index they follow.
Active funds take a different approach, with portfolio managers using research and judgement to select investments, but often charge higher fees and rely heavily on human decision-making. Between these two approaches sit smart-beta ETFs, which use predefined rules to select companies based on specific factors such as value, growth or quality. While these strategies attempt to outperform traditional indexes, they are often dependent on the strength of a single investment signal.
How GOAT Breaks the ETF Mould
With the rise of AI, GOAT is engineered to sit outside the conventional spectrum entirely. Whilst it tracks a rules-based index like a passive ETF, the index itself is built and rebalanced monthly by AI.
Each month, its proprietary model evaluates around 1,200 developed-market companies using thousands of data points spanning company fundamentals, market signals and macroeconomic trends, before selecting the 150 companies with the strongest outperformance probability. By combining multiple signals into a single composite score, GOAT reduces reliance on any individual factor - addressing a key limitation of traditional smart-beta ETFs, which typically rank companies based on a single factor such as value or growth.
This data-driven approach also transforms how investment decisions are made. Unlike passive ETFs that typically adjust only when their underlying index changes, or active managers who are constrained by the limits of human decision-making, GOAT’s index is rebalanced monthly. In effect, this allows the fund to respond more dynamically to changing market conditions, rotating towards emerging tailwinds ahead of traditional review cycles.
Having established what sets GOAT's methodology apart, the next question is whether this actually gives VanEck a genuine edge over ETFs already occupying this space.
Competitive Landscape and VanEck’s Value Proposition
The presence of AI-adjacent ETFs is not new to Australia with the likes of Global X ROBO Global Robotics & Automation ETF (ROBO) and BetaShare’s Global Robotics and Artificial Intelligence ETF (RBTZ). Yet while these ETFs invest in companies developing AI technology, GOAT takes a fundamentally different approach: using AI itself as the stock-picker. When the Australian Financial Review (AFR) asked rivals about the launch, Global X had no immediate plans for a similar product, while BetaShares declined to comment, suggesting that GOAT currently faces limited direct competition in Australia’s AI-powered ETF space.
The technology is nothing new either. GOAT is built with Akros Technologies, a Seoul-based quantitative investment platform whose technology already underpins more than 75 ETFs globally and approximately US$10 billion in assets. VanEck’s proposition is shaped around being the first to market this institutional-style quantitative investing, once reserved for elite quant funds, in a low-cost ETF wrapper, accessible to everyday investors.
However, being first does not necessarily mean being proven. The biggest challenge facing GOAT is that its strongest evidence remains based on simulated historical performance. Since the index’s base date in July 2005, VanEck reports a simulated annual return of 12.49%, versus 9.33% p.a. for the MSCI World ex Australia Index - an annual outperformance of 3.16%. That figure is exactly that: a backtested number that lacks a live trading record.
There are also concerns around investor trust given GOAT is not an entirely new ETF listing, but a rebrand of a previously underperforming fund, the VanEck Vectors Morningstar World ex Australia Wide Moat ETF. While VanEck has changed the strategy behind the ticker, this still raises the question of whether GOAT represents genuine innovation or an underperforming product repositioned to ride the AI hype cycle, which will only become clear once real trading data is available.
Can GOAT deliver?
GOAT’s pitch is simple: let the algorithm do the stock-picking and rebalance faster than any human could. With the first-mover edge and backtest numbers that look compelling on paper, VanEck has built a strong case for itself before a single trade has even gone live.
However, the real test begins now. Entering the market as the first Australian AI stock picking ETF, does GOAT distinguish itself as a potential market pioneer or an AI-driven market experiment? This all depends on how the algorithm performs when faced with real markets, not just historical data.
