AI IPO mania: are investors paying for innovation or imagination?

Let’s begin with a proposition. Eye-catching businesses don’t necessarily make great investments, especially at IPO, when feverish narratives can create towering valuations.

Sticking by this conviction has required intestinal fortitude from institutional investors who resisted being swept along by the excitement associated with the recent SpaceX IPO.

With OpenAI and Anthropic, the companies behind ChatGPT and Claude, respectively, anticipated to list publicly over the next year or so, investors will need to remain clear-headed or risk buying attention-getting but overvalued businesses.

Mindful of the considerable adviser interest in the AI investment theme and the recent SpaceX public listing, we want to share the thinking gained from global investment managers we engage with who have brought forensic minds to assessing links between innovation, valuation and prospective returns.

A scan of our global shareholdings reveals investments in the likes of Microsoft, Apple, Nvidia and Alphabet, great companies, in our view, with visible earnings. Each has earned a place in our portfolios by delivering customer value profitably.

By contrast, it’s debatable whether some of their publicly traded or privately owned counterparts will meet that test any time soon.

SpaceX and demanding assumptions behind some great assets

We would suggest that SpaceX illustrates the distinction between a potentially great company and an IPO offered at a debatable price.

Market participants generally agree on certain things. The company’s rocket launch operations and Starlink are large, high-margin businesses. However, the IPO valuation required investors to attribute substantial future value to AI and platform opportunities beyond these two profitable cores.

A sum-of-the-parts analysis suggests that launch and Starlink, despite their quality and importance, accounted for around US$300 billion of present value to those businesses, leaving most of the roughly US$1.75 trillion valuation dependent on AI-related activities and other longer-dated opportunities whose earnings power remains difficult to assess.¹

In other words, a substantial leap of faith is required to accept the US$1.75 trillion valuation. SpaceX has a long history of turning ambitious ideas into commercial realities, but investors participating at the IPO price appear to have paid upfront for businesses that are either in their infancy, highly capital-intensive, or yet to prove their economics.

Even under optimistic assumptions, the implied valuation appears to rely heavily on businesses beyond launch and Starlink. The central question, therefore, is how much value investors should attribute to SpaceX's emerging AI activities.

Opportunities lie in terrestrial AI infrastructure and compute capacity. More ambitious concepts, such as orbital data centres, may eventually prove valuable, but they face significant technical and economic hurdles and are better viewed as long-term optionality rather than value that can be confidently underwritten today.

SpaceX may be uniquely positioned to address these challenges, but the investment case still depends on outcomes that appear to remain many years away.

Governance is not a footnote

Governance is often overlooked when excitement around an IPO runs high. Yet rights and protections afforded to minority shareholders can materially influence long-term investment outcomes.

In SpaceX's case, investors bought into a structure that leaves Elon Musk, who holds around 42% of the company’s shares but with about 80% voting rights, seemingly indefinitely in control, alongside limited board independence and ongoing questions around related-party dealings.²

From our perspective, founder control is not inherently negative. Some of the world's most successful companies have benefited from leaders with the freedom to think long term. However, when outside shareholders have little ability to influence governance or challenge key decisions, that limitation should be reflected in the valuation they are willing to pay.

Elon Musk is a remarkable entrepreneur but, in our view, also mercurial. We place great value on strong governance and so believe it’s sensible to question the absence of stronger guardrails in the running of SpaceX.

OpenAI and Anthropic show that popularity is no guarantee of profitability

OpenAI and Anthropic are expected to come to market with compelling growth stories and enormous investor interest. Yet investors need to be careful not to conflate rapid adoption with sustainable profitability. High valuations ultimately need to be supported not just by user growth, but by a business model capable of generating durable cash flows.

That may be easier said than done. While both companies have attracted extraordinary private-market valuations with both OpenAI and Anthropic targeting IPO valuations in the US$1 trillion ballpark^3&4, questions remain about how effectively they can convert widespread use of AI tools into long-term sustainable earnings.

The suggested US$1 trillion valuations would place both companies among the most highly valued technology businesses in the world, despite operating in a sector where long-term profitability remains difficult to assess.

The path from AI adoption to shareholder returns remains uncertain. OpenAI does not expect to become cash-flow positive until 2029 and burned US$3.7 billion in the first quarter of 2026 despite generating billions of dollars in revenue.⁵

Anthropic appears further advanced on profitability, but profitable frontier AI businesses remain rare because of the enormous costs associated with developing and serving advanced models.

Some industry observers have also noted that enterprise customers are becoming more cost-conscious, with some exploring cheaper or open-source alternatives as AI spending rises.⁶

The economics of frontier AI differ markedly from those of traditional software businesses. Developing and operating advanced models requires significant investment in specialised chips, data centres, power, networking infrastructure and highly skilled talent. The need for continual model training and upgrades means costs may remain elevated even as revenues grow.

For investors, the key issue is capital intensity. Rapid revenue growth can still produce disappointing investment outcomes if maintaining it requires substantial ongoing expenditure.

Consequently, we think investors should be cautious about relying solely on revenue multiples when assessing prospective AI IPOs. The more important question is whether these businesses can ultimately translate technological leadership into attractive and sustainable returns on capital.

One report noted that OpenAI was generating around US$20 billion in annual revenue while its infrastructure commitments ran into the hundreds of billions of dollars.⁷

Capital intensity is at the centre of the debate

The longer-term AI investment thesis may hold, but valuation and future expected returns are paramount, rather than simply following short-term market momentum. Current market conditions around the likes of SpaceX, OpenAI and Anthropic resemble earlier periods of extreme optimism and abundant liquidity.

Rapid revenue growth alone is not enough. If maintaining a competitive edge demands continuous spending on compute infrastructure, data centres and specialised talent, meaningful free cash flow may take much longer to emerge than headline revenue figures imply.

Investors can simultaneously believe AI will be transformative and remain disciplined about valuation. The real test is whether the future cash flows justify the capital required to generate them, and whether those expectations are already reflected in today's share price.

Scepticism doesn’t mean cynicism

There is an important distinction between scepticism and dismissal. Sceptical investors do not question whether companies such as SpaceX, OpenAI or Anthropic are innovative or influential. Rather, they question whether the valuation, governance arrangements and path to sustainable cash generation justify the price being paid.

The recent SpaceX IPO illustrates the point. While few would dispute the quality of its core businesses, a significant portion of the proposed valuation appears to depend on AI-related opportunities and other future ventures whose economics remain uncertain. Similar questions are likely to arise for OpenAI and Anthropic.

Great companies can still be risky IPOs. The discipline lies in knowing the difference between admiring a company and underwriting it at the price being offered.

More news & insights

Important information

This communication is issued by MLC Asset Management Pty Ltd (ABN 44 106 427 472, AFSL 308953), part of the Insignia Financial group of companies comprising Insignia Financial Ltd ABN 49 100 103 722 and its related bodies corporate (Insignia Financial Group). The Insignia Financial Group is ultimately controlled by CC Capital Partners LLC and its affiliates (a New York based private equity firm) and OneIM GP LLC and its affiliates (a London based private equity firm).

This communication and the information in it is intended for licensed financial advisers and wholesale clients in Australia. It provides a general outline only and is not intended to be a definitive statement on the subject matter. It has been prepared without taking into account individual objectives, financial situation or needs and because of that investors should, before acting on the information in this communication, consider the appropriateness of the information having regard to their personal objectives, financial situation and needs.

Any reference in this communication to a specific company is for illustrative purposes only and should not be taken as a recommendation to buy, sell or hold securities or any other investment in that company.

Any opinions expressed in this communication constitute MLC Asset Management’s judgement at the time of issue and are subject to change. Past performance is not an indicator of future performance. Forward-looking statements should not be relied upon. To the maximum extent permitted by law, neither MLC Asset Management, Insignia Financial Group nor their directors, employees or agents accept any liability for any loss arising in relation to this communication.