How to Invest in ChatGPT in Australia (Best ETFs 2026)
OpenAI is private and ChatGPT will never have a ticker. But there are real, measurable ways to gain exposure to the AI revolution from Australia: from owning Microsoft directly to ASX ETFs with 8-10% MSFT weightings. Here's a complete breakdown.
You can't buy ChatGPT stock directly, but a core-satellite mix of VGS or IVV with a 10–20% allocation to NDQ or FANG gives you meaningful AI exposure without betting the house.
- 01OpenAI is private with no ticker: Microsoft owns ~49% of its profits and makes up 8–10% of NDQ on the ASX.
- 02If you already hold VGS, you've got ~4.9% Microsoft and ~3.8% NVIDIA baked in without doing a thing.
- 03Fees climb with concentration: IVV costs 0.04%, NDQ 0.22%, FANG 0.35%, SEMI 0.57%, and so does the drawdown risk (SEMI and FANG can drop 50%+).
- 04The rational play is 80–90% in a broad core like VGS or IVV, with 10–20% in a concentrated satellite like NDQ or FANG for extra AI upside.
who this is for: Australian investors who keep hearing about ChatGPT and want to know exactly which ASX-listed ETFs actually give them exposure, without FOMO-buying a single meme stock.
When ChatGPT launched in November 2022, it became the fastest product in history to reach 100 million users in just two months. When GPT-4 launched in March 2023, it crossed a threshold that even AI researchers found surprising: genuine professional-level performance on legal, medical, and financial reasoning tasks. The question that followed immediately in investment circles was simple: how do I own a piece of this?
The answer is more complicated than it should be. OpenAI is private, worth approximately $157 billion as of its last funding round, and has no current plans to go public. But there is a real and significant way to gain exposure, and there are ASX-listed ETFs that give you measurable access to the AI revolution without needing an international brokerage account. This article gives you the full picture.
You cannot buy ChatGPT stock: here's why
OpenAI was founded in 2015 as a non-profit by Elon Musk, Sam Altman, Greg Brockman, and others. In 2019 it restructured into a 'capped profit' model to attract the capital needed to train frontier AI models. Under this structure, investors can earn a return, but returns are capped at 100x their investment; any profits beyond that threshold revert to the original non-profit entity.
OpenAI announced a further corporate restructuring in late 2024, converting to a Delaware public benefit corporation (PBC), partly to facilitate future capital raises. But this is not the same as an IPO. Sam Altman has said publicly that an IPO is 'not a priority right now', and even if it were, it would likely be years away. Pre-IPO shares do exist and occasionally trade on secondary markets like Forge Global, but these are illiquid, expensive, and entirely inaccessible to most retail investors.
Multiple fraudulent investment products have attempted to capitalise on the ChatGPT brand. There is no legitimate 'ChatGPT stock', 'OpenAI ETF', or 'official AI token' available for retail purchase. Any product claiming direct ownership of ChatGPT or OpenAI equity is almost certainly fraudulent. ASIC has issued warnings about AI-themed investment scams targeting Australian retail investors.
The Microsoft-OpenAI deal: what you actually own through MSFT
Microsoft's relationship with OpenAI is uniquely structured and deserves careful reading. The deal has gone through three phases:
- ✓2019: Microsoft's initial $1 billion investment gave it non-voting equity and the exclusive right to deploy OpenAI's technology commercially.
- ✓2021: A further undisclosed investment extended the commercial partnership and deepened Azure's role as OpenAI's exclusive cloud provider.
- ✓2023: A reported $10 billion 'multiyear investment', now commonly cited as $13 billion total, giving Microsoft 49% of OpenAI's profits until it recovers its investment. After recovery, Microsoft's share drops to approximately 49% of remaining profits, with OpenAI's non-profit entity receiving the rest.
Three critical structural advantages come with this deal. First, Azure is OpenAI's exclusive cloud provider. Every ChatGPT query, every API call, every enterprise deployment runs on Microsoft's infrastructure. Azure AI services revenue was growing at over 50% year-on-year as of late 2024. Second, Microsoft has exclusive rights to productise OpenAI models: the foundation of GitHub Copilot (1.3+ million paid subscribers), Microsoft 365 Copilot (deployed across enterprises paying ~$30/user/month), Azure OpenAI Service (the enterprise API), and Bing Chat. Third, Microsoft sits first in the capital recovery waterfall: it must recoup its $13B+ investment from OpenAI's profits before any other equity holder receives returns.
Microsoft's AI-driven revenue lines as of FY2025: Azure AI (>50% YoY growth), GitHub Copilot ($19-$39/user/month, 1.3M+ subscribers), Microsoft 365 Copilot ($30/user/month enterprise), Azure OpenAI Service (API access billed per token). Collectively, Microsoft CEO Satya Nadella guided that AI was on a path to a '$10 billion annual revenue run rate' as of mid-2024. This is not speculative future revenue, it is recurring, subscription-based income already being invoiced.
Microsoft as a standalone investment for Australians
For Australian investors with access to international markets, buying Microsoft (MSFT) directly is the most concentrated bet on the ChatGPT ecosystem available. MSFT trades on the NASDAQ (ticker: MSFT) and can be purchased through brokers like CommSec International, Stake, Interactive Brokers Australia, or Superhero.
Microsoft's investment case rests on more than just OpenAI. It is one of three companies globally that operates hyperscale cloud infrastructure at massive scale (alongside Amazon AWS and Google Cloud), has a >$100 billion annual revenue franchise in enterprise software (Office 365, Teams, Dynamics), and has a >$50 billion annual run rate in Azure cloud. The investor relations page confirms Azure as the primary growth driver, with AI as the accelerant. MSFT's forward P/E of approximately 32-35x as at March 2026 reflects the market pricing in significant continued AI revenue growth.
When you receive dividends from MSFT (currently ~0.75% yield), they are subject to 15% US withholding tax under the Australia-US tax treaty. You must complete a W-8BEN form through your broker to qualify for this reduced rate (the default rate is 30%). The withheld amount is generally creditable against your Australian income tax liability. Capital gains on MSFT held for more than 12 months still qualify for Australia's 50% CGT discount, the foreign-sourced capital gain is included in assessable income and the discount applied.
The practical limitation for most Australians is access. Not all brokers offer international trading, and those that do typically charge higher brokerage ($5-$20 per trade versus $6-$10 for ASX ETFs). Currency conversion also applies, AUD must be converted to USD to purchase MSFT, introducing transaction costs and ongoing currency exposure. For investors who prefer to stay on the ASX, ETFs are the more practical vehicle.
ASX ETFs with Microsoft exposure, ranked by weight
Five mainstream ASX ETFs hold Microsoft as a top holding. The chart below shows the percentage of each fund's portfolio that is allocated to MSFT. This represents your indirect exposure to the ChatGPT/OpenAI ecosystem through each fund.
FANG+ ETF (ASX: FANG), 0.35% MER
The BetaShares FANG+ ETF tracks the NYSE FANG+ Index, a concentrated basket of approximately 10 leading tech and tech-adjacent companies. Microsoft, Amazon, Apple, Meta, Alphabet, Netflix, NVIDIA, Tesla, Snowflake, and one other rotating name. Because the index is equal-weighted and rebalanced quarterly, each position is roughly 10% of the fund. This makes FANG the highest-MSFT-weight ETF available on the ASX, and also the most concentrated.
The concentration is a double-edged sword. In 2023 and 2024, when the Magnificent Seven drove virtually all S&P 500 returns, FANG dramatically outperformed. In any period where large-cap tech underperforms (as it did in 2022), FANG suffers more than a diversified alternative. With only ~10 holdings, a single company disappointing on earnings can visibly move the fund.
Nasdaq 100 ETF (ASX: NDQ), 0.22% MER
The BetaShares Nasdaq 100 ETF tracks the 100 largest non-financial companies listed on the NASDAQ exchange. Microsoft is the second-largest holding at approximately 8.1%, behind only Apple. The top five holdings, Apple, Microsoft, NVIDIA, Amazon, Alphabet, represent roughly 40% of the fund. NDQ offers broad tech sector exposure with meaningful MSFT weight, at a lower MER than FANG and with significantly more diversification (100 holdings vs ~10).
For most Australians wanting AI exposure via ETFs, NDQ represents the sweet spot between concentration and diversification. You get meaningful Microsoft weight (~8%), plus direct NVIDIA exposure (~5%), the chip company whose GPUs power every ChatGPT training run and inference call, plus Meta (which has built its own LLaMA models), Alphabet (Gemini), and Amazon (Bedrock). Every major player in the generative AI race is represented.
S&P 500 ETF (ASX: IVV), 0.04% MER
The iShares Core S&P 500 ETF is the cheapest way to hold MSFT on the ASX. At 0.04% MER, 85% cheaper than FANG, IVV gives you Microsoft as its largest holding at ~6.7%, plus NVIDIA (~6%), Apple (~6.5%), and Alphabet (~4%). The tradeoff versus NDQ is broader diversification (500 stocks including healthcare, financials, energy) and lower fees, but slightly less concentrated tech exposure.
For investors who want AI exposure without overweighting tech, IVV is the most cost-efficient option. The fee difference between IVV (0.04%) and NDQ (0.22%) compounds meaningfully over decades, on a $100,000 investment over 20 years, NDQ's additional 0.18% annual fee costs approximately $38,000 in foregone compounding. That is a real cost that needs to be weighed against the higher tech concentration.
Global Robotics & AI ETF (ASX: RBTZ), 0.57% MER
The BetaShares Global Robotics and Artificial Intelligence ETF tracks an index specifically designed to capture companies deriving significant revenue from AI and robotics. Holdings include NVIDIA, Microsoft, Intuitive Surgical, Keyence, and various industrial automation companies. MSFT's position (~6%) is meaningful, but RBTZ also captures mid-cap AI specialists that don't appear in the Nasdaq 100 or S&P 500.
RBTZ is the most thematically 'pure' AI ETF on the ASX in that it explicitly screens for AI and robotics revenue. The cost is the highest MER in this comparison (0.57%) and a more limited track record than NDQ or IVV. Performance in 2023-2024 was strong, driven by NVIDIA, but the fund's thematic construction also means it can diverge significantly from broad market returns in both directions.
Picks and shovels: the semiconductor angle
The 19th century California Gold Rush produced very few wealthy miners, but enormous profits for the people who sold the picks, shovels, and denim jeans. The AI boom has an equivalent: the semiconductor companies that make the chips required to train and run large language models. ChatGPT, Claude, Gemini, and every other frontier AI model runs on data centre GPUs made almost exclusively by NVIDIA.
The BetaShares Global Semiconductors ETF (ASX: SEMI) tracks the top global semiconductor companies: NVIDIA (~20% weight), TSMC (Taiwan Semiconductor, which manufactures chips for Apple, NVIDIA, and AMD), Broadcom, ASML (the Dutch company that makes the extreme ultraviolet lithography machines that enable the chips), AMD, and others. SEMI holds zero Microsoft, it is purely chip-focused.
This is a different bet from MSFT or NDQ. SEMI is a bet on AI infrastructure demand, that regardless of which AI model wins the consumer or enterprise market, the demand for chips to run those models continues to grow. NVIDIA's data centre revenue grew from $3.6B to $47.5B between FY2023 and FY2024, a 13x increase in a single year. The semiconductor supply chain (TSMC manufacturing, ASML lithography, Broadcom networking) all benefit from this secular demand.
NVIDIA's H100 and A100 GPUs are not interchangeable with other hardware for AI training at scale. OpenAI, Google, Meta, and Microsoft all purchase NVIDIA GPUs, there is currently no viable alternative for cutting-edge model training. This near-monopoly on the AI training compute layer is why NVIDIA's gross margins expanded to over 74% in FY2024 and why its market cap surpassed $3 trillion in 2024. SEMI investors get approximately 20% NVIDIA weight, the single most direct chip-level exposure to the AI boom available on the ASX.
Performance: how have these options actually performed?
The chart below compares approximate 1-year and 3-year annualised returns (in AUD) for MSFT directly and each of the main ASX ETF options. These figures include currency movements, a significant factor given the AUD/USD rate has fluctuated between 0.62 and 0.70 over the past 3 years.
Several patterns stand out. First, SEMI has delivered the strongest returns in both the 1-year and 3-year periods, driven almost entirely by NVIDIA's extraordinary performance. SEMI is up roughly 38% over the past 12 months (AUD) and approximately 21% per year annualised over three years. Second, FANG has outperformed NDQ over both periods, reflecting the concentration effect, in bull markets, FANG's mega-cap tech concentration amplifies returns. Third, Microsoft's direct return (14% over 1 year) has lagged the broad Nasdaq 100, partially because MSFT was already priced for AI success in early 2023 and has re-rated less dramatically than NVIDIA.
The semiconductor and tech sector's extraordinary 2023-2025 returns have been unusual by historical standards, driven by a market re-rating following unexpected AI capability jumps. The same thematic concentration that produced SEMI's ~38% 1-year return means it is capable of -50%+ drawdowns if the AI investment cycle disappoints, chip demand weakens, or new competitors emerge. NVIDIA traded at a P/E of over 60x as at early 2026. At those multiples, the market is pricing in sustained hyper-growth, any growth deceleration carries significant downside.
Fee comparison: what you pay for AI focus
A clear pattern emerges when you compare the annual fees of each option. More specialised AI focus generally costs more. The cheapest broad exposure (IVV at 0.04%) gives you MSFT, NVIDIA, and all other major AI players via market-cap weighting. The most AI-focused options (SEMI, RBTZ at 0.57%) charge 14x more annually.
On a $100,000 portfolio, the difference between IVV (0.04%) and SEMI (0.57%) is $530 per year, every year. Over 20 years at 10% gross returns, that fee gap compounds to roughly $95,000 in foregone wealth. The question is not whether that gap is large (it is), but whether the more specialised AI exposure justifies the cost. That depends on how right you believe your AI thesis to be and your willingness to accept the sector concentration risk that comes with SEMI and RBTZ.
IVV: $40/yr per $100k | NDQ: $220/yr | FANG: $350/yr | SEMI & RBTZ: $570/yr. These fees are deducted daily from the fund's net asset value, so you never see them as a line item. Over 30 years Over 30 years at 9% gross returns, the gap between IVV ($40/yr) and SEMI ($570/yr) compounds to approximately $160,000 on a $100,000 starting balance. That is the price of thematic specialisation.
Building a ChatGPT portfolio as an Australian investor
How you combine these options depends on three questions: how concentrated do you want to be in AI/tech, what level of single-stock (or single-sector) risk are you comfortable with, and what role does this play in your broader portfolio?
Option A: Diversified core with AI tilt (lower risk)
Hold VGS or IVV as the core position (70-80% of the portfolio allocation) and add NDQ as a tech tilt (20-30%). This gives you approximately 6-8% Microsoft weight and broad NVIDIA/tech exposure via NDQ, while limiting the concentration risk of a pure tech bet. The blended MER is approximately 0.06-0.10% depending on the split. A very cheap blended cost..
Option B: Concentrated AI bet (higher risk, higher potential return)
Hold NDQ (50%) + SEMI (25%) + FANG (25%). This gives concentrated exposure across the software/platform layer (via MSFT, NVDA, Meta in NDQ and FANG) and the hardware/infrastructure layer (via NVDA, TSMC, AMD in SEMI). Blended MER is approximately 0.34%. This portfolio will significantly outperform or underperform a broad global index depending on how the AI theme plays out. Not appropriate as a total portfolio, only as a satellite position.
Option C: MSFT direct + NDQ (for international brokerage users)
Buy Microsoft directly on NASDAQ via Stake or Interactive Brokers (2-5% of portfolio), complemented by NDQ on the ASX for the broader AI ecosystem. This gives you the highest MSFT concentration of any approach while diversifying the remainder across 100 companies. Requires an international brokerage account and comfort with W-8BEN tax forms.
For long-term investors, AI/tech exposure is already present in any VGS, IVV, or NDQ holding at meaningful weight. If you already hold VGS, you own approximately 4.9% Microsoft, 3.8% NVIDIA, and 3.5% Apple, you already have AI exposure. The question is whether to overweight it. For most investors, a satellite allocation of 10-20% of the total portfolio in a more concentrated AI vehicle (NDQ, FANG, or SEMI) is a reasonable way to express a higher conviction view without abandoning diversification entirely.
Risks to the AI investment thesis
The AI thesis is compelling but not without significant risks. Honest analysis requires confronting them directly.
- ✓Valuation risk: NVIDIA at 60x earnings and Microsoft at 35x earnings assumes sustained hyper-growth. If AI adoption in enterprise software slows, or if cloud AI infrastructure spending decelerates, these multiples compress. A P/E de-rating from 60x to 35x is a ~40% price decline even with no earnings disappointment.
- ✓Competition risk: The AI model landscape is intensifying rapidly. China's DeepSeek R1 demonstrated in January 2025 that frontier AI models could be trained at dramatically lower cost than US competitors assumed, causing a significant single-day decline in NVIDIA and AI-linked stocks. If the 'compute moat' erodes faster than expected, the chips-and-infrastructure thesis weakens.
- ✓Regulation risk: The EU AI Act became law in 2024, with phased implementation through 2026-2027. US AI regulation is evolving. OpenAI itself faces regulatory scrutiny on data practices, copyright, and safety. Additional regulatory burden could slow AI adoption curves in key enterprise markets.
- ✓OpenAI-specific risk: Microsoft's AI advantage is structurally tied to its OpenAI relationship. If OpenAI encounters major technical setbacks, loses key personnel, or faces litigation that limits its ability to operate, Microsoft's AI advantage could narrow. OpenAI has already faced significant internal instability (the Sam Altman board removal and reinstatement in 2023).
- ✓Currency risk: All ASX ETFs holding US stocks are subject to AUD/USD movements unless hedged. In periods when the AUD appreciates (typically when global risk appetite is high), unhedged ETF returns in AUD terms are reduced. Over the past 10 years the AUD has broadly depreciated, providing a tailwind, but this is not guaranteed to continue.
Methodology
ETF data (MER, holdings, weights) is sourced from each fund provider's current Product Disclosure Statement and monthly fund update reports: BetaShares (NDQ, FANG, SEMI, RBTZ) and iShares/BlackRock (IVV). Microsoft's weight in each ETF reflects the most recent available monthly holdings data as at March 2026. Holdings weights can shift between monthly publications based on index rebalancing and market price movements.
Performance figures are approximate and sourced from fund provider fact sheets and publicly available market data. All returns are expressed in Australian dollars (AUD) and include the impact of currency movements from AUD/USD fluctuations. Past performance figures used in the chart are intended to be illustrative and directionally accurate; they should not be treated as precisely audited figures. For current performance data, refer directly to the fund providers: BetaShares, BlackRock iShares. The Microsoft-OpenAI investment terms referenced in this article are drawn from public regulatory filings, including Microsoft's 10-K, and contemporaneous reporting by the Financial Times, Wall Street Journal, and Bloomberg. This article is general information only and does not constitute financial advice. All investments carry risk.