Is 17% US GDP Growth in 2030 Realistic? Why the Tech Experts Don’t Know What Happens Next

If you want a helpful answer about what’s coming, ask an expert, right? 

Maybe, but what if the experts are as bad at predicting the future as the rest of us? 

Examples proving this point abound. 

In October 2022, US inflation had topped 9%, the S&P 500 was down 25% and a Bloomberg Economics model put the odds of an American recession within a year at 100%.  

The following month, OpenAI released a ‘low-key research preview’ called ChatGPT. Staff had run a betting pool on first-week users. The most optimistic guess was 100,000. It passed one million in five days.  

As we all know, the US recession never came, and AI became the engine of this global bull market. 

Four years on, the people building AI are making bold forecasts about jobs, growth and the global economy of 2030.  

Australian investors with global ETFs and funds already hold a large bet on those forecasts coming true.  

Before relying on tech insiders, though, it’s worth asking how well they have predicted what’s coming next within their own industry. 

Key Takeaways 

• Even the builders of breakthrough technology routinely misjudge how fast it will spread and what it will do. 
• Anthropic staff’s survey answers, run through the company’s economic model, imply 17.1% US GDP growth in 2030, a pace never seen in the post-war era; those of US adults imply 5.4%. 
• History shows that transformative technology and strong shareholder returns are not the same thing. 
• With the tech sector representing only 2% of the ASX 200, Australians get most of their direct AI exposure offshore, much of it through super, so sizing and diversifying it matters more than any forecast. 
  
  

The Insiders Didn’t See It Coming Either 

Given AI is the main game in town, let’s step back to 2021, before it had emerged as a distinct investment theme across global markets. 

Silicon Valley spent 2021 selling the metaverse, Web3 and NFTs.  

Facebook even renamed itself Meta. Its metaverse division, Reality Labs, has racked up more than US$80 billion in operating losses since late 2020.  

Almost nobody was pitching chatbots at the time. 

Then ChatGPT arrived and surprised the people who built it. OpenAI staff told MIT Technology Review they had been caught off guard by the strength of its reception. 

Fast Forward to the Forecasts Being Made Today 

With that context, let’s turn to the forecasts tech insiders are making today. 

This month, Anthropic published ‘The AI expectations gap’, comparing the expectations of 1,334 of its staff with 10,980 US adults.  

Anthropic staff expect AI to handle 75% of their own work by 2030, against 28% for US workers, and 91% expect a Nobel-level AI discovery by then, against 39% of the public.  

Run through the company’s economic model, the median staff answers imply US GDP growth of 17.1% in 2030. You read that right. Taken at face value, staff answers point to an unprecedented period of US GDP growth. The public’s expectations are still bullish at 5.4% versus actual US economic growth last year of 2.2%. 

Anthropic stresses these are model outputs from survey answers, not company forecasts.  

There are other reasons for caution.  

The staff poll was run internally in August and was unweighted, and its 1,334 respondents work for a company whose fortunes depend on AI adoption. By contrast, a repeat survey of the public in September produced a near-identical 5.6% 2030 US GDP growth expectation. 

Moreover, economist Ernie Tedeschi observed that the US has never recorded annual growth anywhere near 17% since the Second World War, even in the post-Covid rebound. The post-war high was 8.7%, in 1950. 

None of this means the optimists are directionally wrong. AI may well lead to higher US and global economic growth. 

But investors need to understand that tech expertise is quite different from the expertise needed to forecast the economic consequences.  

Case in point: Philip Tetlock’s landmark study of 300 experts found they barely beat chance on long-range predictions, and the most confident and famous often did worst.  

And in 2007, Microsoft’s Steve Ballmer famously said there was ‘no chance that the iPhone is going to get any significant market share’. 

Real-World Friction Rarely Aligns with the Forecasts 

Forecasts of an AI utopia tend to assume a frictionless world.  

Yet we all know the real world pushes back, sometimes hard.  

The Financial Times recently described how AI agents could move customers’ idle cash into higher-yielding accounts, squeezing bank margins. Around the same time, JPMorgan said it was blocking outside AI agents from acting in customer accounts. The real world pushed back on the theory. 

Australian portfolios are exposed to a similar theme.  

Financials are the largest sector of the ASX 200 and the big four banks are core holdings for most local investors, so any AI-driven squeeze on cheap deposit funding would flow straight through to Australian shareholders. 

Even the present moment is hard to read.  

The New York Times ran stories 12 days apart arguing first that higher rates weren’t impacting the AI boom and then that they were. Both cited well-informed experts. 

Australia faces its own examples of the real world pushing back.  

CommBank estimates the local data centre pipeline at around six gigawatts, with a central build estimate of A$150 billion by 2030.  

However, that depends on a lot of ducks lining up.  

For example, executing on this plan depends on power, planning approvals and adequate financing. Yet the RBA lifted the cash rate from 3.60% to 4.35% in the first half of this year, then to 4.60% in September, its fourth rise of 2026 and the highest cash rate since November 2011.  

The pipeline is already thinning in the real world.  

The Australian Energy Market Operator (AEMO) reports that developers have cancelled more than a third of the data centre projects it listed a year earlier.  

And CommBank observed the boost to GDP is far smaller than the investment headline, at only 0.2 percentage points p.a. in 2026 and 2027. 

Great Technologies, Mixed Outcomes 

History is even blunter on the subject.  

Railways once transformed the world. Yet Britain’s 1840s railway mania ended in a crash that wiped out many of the investors who funded it.  

Cisco, the Nvidia of the dot-com era, needed more than 25 years to close above its March 2000 peak on a price basis, excluding dividends, finally doing so in December 2025.  

In 1999, Warren Buffett reminded investors that the US had produced 2,000 car makers, of which only three survived.  

Economist Jay Ritter found a negative correlation of minus 0.37 between economic growth and real equity returns across 16 countries from 1900 to 2002.  

The lesson was growth that everyone expects tends to be priced in. 

Disruption also cuts the other way.  

In FY26, fears that AI would undermine software businesses helped drive the ASX 200 technology sector down 37%. 

That sell-off was hard to call — as was the subsequent rebound. ASX 200 tech shares rebounded by 20% between 31 March and early July. 

In short, things rarely play out as expected. 

Investment Implications 

Let’s start with the elephant in the room for Australian investors. Technology makes up just 2.1% of the ASX 200, compared with 40% of the S&P 500, where the ten largest companies alone make up about 39% of the index.  

Most Australian investors’ AI exposure therefore comes via their broad global holdings, such as market-cap-weighted world ex-Australia ETFs, where the largest US technology companies dominate the top holdings.  

Factor-based funds tilt that mix. The Global X S&P World ex Australia GARP (Currency Hedged) ETF (ASX:GHRP), launched in May 2026, holds 22.6% in information technology, well below the S&P 500’s weighting. 

An unhedged global tech holding is also a bet on the Australian dollar against the US dollar. A falling Australian dollar has historically cushioned local investors during global sell-offs, while currency-hedged funds remove that effect. Many investors split their global allocation between hedged and unhedged versions rather than backing one currency view. 

If you want targeted exposure to the tech theme, thematic ETFs and Technology & AI funds will enable you to further concentrate your exposure.  

Position size matters more than product choice. Make sure you do your due diligence. 

And remember to diversify. Balance your tech exposure with assets driven by different forces, such as infrastructure, bond funds and Australian equity funds.  

Our analysis of the AI capex boom sets out the warning signs worth watching. 

Listen to the Builders, Then Diversify 

The people building AI deserve a hearing on what the technology can do. They have a weaker record on how fast it will spread, who will profit and what regulators, banks and voters will allow. In 2022 the confident call was for a US recession, while the real story was a chatbot almost nobody predicted as the runaway success it became. 

Nobody knows how this will end. Not the venture capitalists, the engineers, the financiers, the politicians or you. A portfolio built for a range of futures doesn’t need anyone to. 

Frequently Asked Questions 

Can Technology Experts Predict the Sharemarket’s Outlook? 

Rarely. They understand what a technology can do, but adoption speed, regulation, competition and valuation all shape returns. OpenAI’s own staff underestimated ChatGPT’s first-week uptake at least tenfold. 

What Did Anthropic’s Staff Survey Find? 

Anthropic’s ‘The AI expectations gap’ report, published on 16 October 2026, found staff expect far faster AI progress than the public. Their median answers, run through Anthropic’s economic model, imply 17.1% US GDP growth in 2030, against 5.4% for US adults. Anthropic notes these are model outputs, not company forecasts. 

How Much AI Exposure Does an Australian ETF Investor Have? 

Usually more than they think. Technology is only 2% of the ASX 200, but information technology is around 40% of the S&P 500 and the ten largest S&P 500 companies, most of them AI-linked, make up about 39% of that index. Most Australians also hold global ETFs through their super. 

Does a World-Changing Technology Guarantee Good Returns? 

No. Railways, cars and the internet all changed the world, yet many investors lost money. Ritter’s research found faster economic growth has not translated into higher equity returns across countries. 

Where Do Most Australians Get Their AI Exposure? 

Mainly offshore. Because technology is only 2% of the ASX 200, most exposure comes through the international shares held in superannuation funds and global ETFs. Local exposure comes indirectly through data centre, property and energy businesses. 

How Are Higher Interest Rates Affecting the AI Boom in Australia? 

They raise the cost of financing Australia’s data centre build-out, which CommBank estimates at $150 billion by 2030. The RBA has lifted the cash rate four times in 2026, to 4.60%, and AEMO reports that more than a third of previously listed data centre projects have been cancelled. 

Why Did ASX Tech Shares Fall in FY26? 

Fears that AI would disrupt software businesses, along with company-specific problems such as WiseTech’s governance issues, drove the S&P/ASX 200 Information Technology index down 37% in the year to 30 June 2026. WiseTech fell 70% and Xero 60%. 

Should Australian Investors Hedge Their Global Tech Exposure? 

It depends on your view of the Australian dollar and your need for diversification. Unhedged exposure adds currency risk but has historically cushioned sell-offs, while hedged funds track the underlying shares more closely. Splitting between the two is a common middle path. 

Disclaimer: This article is prepared by Simon Turner. It is for educational purposes only. While all reasonable care has been taken by the author in the preparation of this information, the author and InvestmentMarkets (Aust) Pty. Ltd. as publisher take no responsibility for any actions taken based on information contained herein or for any errors or omissions within it. Interested parties should seek independent professional advice prior to acting on any information presented. Please note past performance is not a reliable indicator of future performance.