ETFtracker · Institutional
ETF-Implied Demand · Part 2 of 2 · Offshore
Ownership & concentration · August 2026

You own more Nvidia than CBA.

Not you personally, but the Australian ETF investor, in aggregate. Through their funds, Australians hold A$7.3bn of Nvidia against A$7.1bn of Commonwealth Bank, and a bigger equity book offshore (A$148.5bn) than at home (A$90.0bn). Part 1 measured the passive bid on the home market. Part 2 turns the same engine on the global look-through, and finds a portfolio most investors would not recognise as their own.

01

The bigger book is offshore

The mental model of an Australian ETF portfolio is banks and miners with a global fund bolted on the side. The look-through says the opposite. Resolve every ETF’s holdings to the underlying companies and the offshore book, at A$148.5bn, is 65% larger than the entire ASX book at A$90.0bn. The average passive investor in this country is now more exposed to global mega-cap technology than to the Big Four combined.

Nothing illustrates the shift better than a single ranking. Sort every company, Australian or foreign, by the dollars the ETF complex holds in it, and Nvidia lands at number two, wedged between BHP and Commonwealth Bank. The largest holdings Australians have quietly stopped being only their own champions.

The ten largest single-stock ETF positions in Australia
ASX-listedGlobal
BHP8.9 NVDA7.3 CBA7.1 AAPL6.5 MSFT5.4 NAB4.3 WBC3.6 ANZ3.4 AMZN3.0 GOOGL3.0 A$ billions held across all ASX-domiciled ETFs
A$148.5bn
offshore ETF-held, vs A$90.0bn at home
A$7.3bn
Nvidia held, the #2 position in the country
21%
of the offshore book sits in the Magnificent Seven
+A$3.01bn
net mechanical demand offshore, August
02

A concentration nobody chose

“Diversified” global exposure is, underneath, a single trade.

The offshore book is not a broad basket of the world. It is a wager on eight American companies. The Magnificent Seven plus Broadcom account for A$31bn, or 21% of the entire A$148.5bn, and once the rest of the semiconductor complex is added, Taiwan Semiconductor, ASML, Micron, AMD, roughly a third of every offshore dollar Australians hold sits in AI and the chips that feed it. This is not a criticism of the funds; it is what capitalisation-weighting a concentrated market produces. But it means the words “globally diversified”, on the tin of most international ETFs, describe a portfolio whose fate is largely a single theme’s.

Largest offshore holdings of the Australian ETF complex
StockCompanyETF-heldMech. demandIn funds
NVDANvidiaA$7.31bn+A$140.2m59
AAPLAppleA$6.49bn+A$119.1m46
MSFTMicrosoftA$5.38bn+A$103.9m52
AMZNAmazonA$3.01bn+A$55.1m38
GOOGLAlphabet (A)A$2.99bn+A$58.2m63
AVGOBroadcomA$2.86bn+A$47.9m49
METAMeta PlatformsA$2.28bn+A$43.7m40
2330Taiwan SemiconductorTWA$1.71bn+A$33.1m15
LLYEli LillyA$1.67bn+A$30.8m47
ASMLASMLNLA$1.25bn+A$23.8m53

Alphabet appears twice across the complex (A and C shares); combined it would rank fourth. Held figures are floors across covered ETFs.

03

The same tilt, a world away

The rotation Part 1 found at home is running offshore too.

Apply the intensity lens, mechanical demand as a share of each name’s ETF base, and August’s offshore cross-section rhymes with the home market to a striking degree. The names the passive bid pressed hardest were gold and defensives: precious-metals streamers Wheaton and royalty miner Agnico Eagle at the very top, alongside regulated utilities (Duke, Southern, Enbridge) and a Canadian bank. The names it pressed coolest were the crowded technology winners, Korean memory makers and the cybersecurity complex, running at a third of the median pace.

Gold and defensives led the passive bid on both sides of the Pacific. This is not an Australian quirk; it is what August’s flows did to the index everywhere Australians invest.

Offshore, pressed hardest
StockIntensityHeld
WPMCA6.00%A$209m
AEMCA5.33%A$244m
RYCA2.90%A$221m
DUK2.74%A$222m
SO2.72%A$228m
ENBCA2.71%A$210m

Gold streamers (Wheaton, Agnico), utilities and energy infrastructure (Duke, Southern, Enbridge), a Canadian bank.

Offshore, pressed coolest
StockIntensityHeld
0006600.36%A$882m
0059300.57%A$1.02bn
FTNT0.65%A$359m
CRWD0.84%A$519m
PANW0.97%A$675m

Korean memory (SK Hynix, Samsung) and US cybersecurity (Fortinet, CrowdStrike, Palo Alto), owned but drawing a thinner slice of the flow.

The median offshore name ran at 1.92% intensity, well above the home market’s 1.40%, a quiet confirmation that in August Australians pushed harder into global funds than domestic ones. The tilt within that push, though, was the same rotation seen at home: toward gold and the defensives, away from the crowded growth names that already dominate the book.

04

What this lens is for

A word on how to read Part 2 against Part 1, because they are not the same instrument. At home, mechanical demand is a genuine market force: local ETFs are a meaningful marginal buyer of ASX stocks, so the bid moves prices and the signal is tradeable. Offshore it is not. A$140m of monthly Nvidia demand is a rounding error against a company worth trillions; Australian flows do not move Nvidia. So Part 2 is not a demand signal. It is an exposure and crowding lens, and on that basis it earns its place:

  • Know the real portfolio. The look-through is the only way to see what an Australian’s diversified ETFs actually own. The honest answer is: a large, concentrated, single-currency bet on US technology.
  • Measure the crowding you inherit. Breadth offshore is extreme, Alphabet and Nvidia sit in 60-plus funds each. The same name arrives through dozens of “different” products, so apparent diversification hides a single concentrated position.
  • Watch the theme, not the ticker. With a third of the book in AI and semiconductors, the risk that matters to an Australian passive investor is not a stock, it is a theme derating. This lens sizes that exposure precisely.
05

Method & limits, up front

  • Large-cap by design. The global master is bootstrapped from broad developed, US and emerging-market ETFs, so it resolves the mega-caps that hold the overwhelming majority of the book cleanly, and deliberately leaves a long tail of small global names unresolved. The mega-cap figures are robust; the deep tail is indicative.
  • Identifier collisions in the tail. Offshore, a bare ticker is not unique across exchanges (a numeric Tokyo code or a single letter can mean two different companies). We exclude collisions we detect, but some in the long tail remain; treat individual small-cap lines with care, not the aggregates.
  • An exposure lens, not a signal. Australian ETF flow is immaterial to global prices, so mechanical demand here measures what Australians are accumulating, not a force acting on the stock.
  • Floors, dollars, and a monthly snapshot. Same basis as Part 1: figures are floors across covered ETFs; dollars, not percent of float; weights current, flows and FUM as at the latest August disclosure. Geared funds excluded.
Part 1 of 2
The stocks the index is quietly buying, at home
ETFtracker · Institutional Research · ETF-Implied Demand, Part 2 (Offshore) · August 2026 data. A research and exposure lens, not investment advice. Global figures are floors across covered ETFs and focus on large-caps; intensity = mechanical demand ÷ ETF-held.
Explore the data on ETFtracker

Research from ETFtracker. A research signal, not investment advice.