Author: Guy Foster, Chief Strategist

It’s an extraordinary time to be an investor.

We’re witnessing a moment in which the boundaries of technology are expanding faster than at any other point in human history. In recent weeks, the financial world has been transfixed by a slate of blockbuster public listings that feel less like corporate milestones and more like cultural phenomena.

Elon Musk’s SpaceX shattered records with a monumental initial public offering (IPO) that raised over $75 billion, valuing the aerospace and satellite pioneer at an astonishing $1.75 trillion.¹ Meanwhile, AI titans Anthropic and OpenAI have both confidentially filed paperwork for their own public listings. Anthropic recently closed a substantial funding round valuing it at $965 billion, while OpenAI is actively targeting the $1 trillion mark.²

An IPO is the first opportunity for investors to buy shares in a company on the public stock markets –meaning an investment that they can buy or sell at will. Earlier investors had to pay for shares in the company without knowing when, or even if, they would ever recoup their money. The SpaceX IPO enabled SpaceX to raise further capital to pursue its own investment plans, but what follows is an opportunity for those early investors to finally sell and realise substantial profits. At the IPO, demand for the shares exceeded supply, causing the share price to rise in early trading and pushing the valuation well over $2 trillion.³

Part of the reason for this is the story – reusable rockets, a constellation of satellites and orbital data centres. But a significant factor is that only a very small share of the company was sold in those first days (about 4%).1 By September, we expect roughly four times as many SpaceX shares to be available, rising to ten times by the end of the year. This supply of new shares will greatly reduce the shortage, but could it create a glut? If there are more shares available than willing buyers at a given price, then that price needs to ‘correct’ – an investor’s euphemism for ‘fall.’

That question – of price correcting to meet reality – runs deeper than any single IPO. A great technology doesn’t automatically equate to a great investment. Today, thoughtful investors must grapple with a profound central tension: AI may well be the most significant technological shift of our lifetime, but it may also be producing one of the most consequential valuation bubbles in decades. This isn’t scepticism for its own sake – it’s about making sure the story doesn’t run away with the numbers.

Why is everyone talking about these listings?

OpenAI’s ChatGPT is as synonymous with AI as Google is with search, yet unlike Google, has been off-limits to ordinary investors. Listing changes that.

Anthropic, meanwhile, has been pushing technological frontiers further still. Its latest model prompted the U.S. government to issue an unprecedented directive blocking foreign nationals from accessing it – a ban since lifted, but a warning shot nonetheless: the most powerful AI models may periodically be restricted from commercial sale on national security grounds, limiting the addressable market on which colossal investment decisions have been made.

To look at these companies and want ‘in’ isn’t foolish; OpenAI and Anthropic are building the ‘foundational models’ and infrastructure that are changing how we work, communicate and think – it can feel to some investors like a direct way to secure a slice of the future. However, stepping from the private arena into the public market introduces a completely different set of rules.

Who are you really buying from?

To evaluate IPOs properly, investors should consider the mechanics of the transaction. There’s a fundamental difference between buying shares at an IPO and purchasing them later in the secondary market (the open stock exchange).

Publicly traded shares are subject to strict rules that aim to ensure everyone has the same information about the company. But at the point of flotation, investors must buy from the private shareholders, owners and operators of the business – the very people who know the business best and have chosen this moment as the ideal time to sell.

Insiders choose the exact moment to go public – usually when market sentiment is at an absolute peak and the company’s growth story looks at its most flawless. Fund managers frequently warn that a sudden flurry of historic fundraisings often accompanies the very top of a market cycle, as insiders rush to lock in elevated valuations before reality catches up.

What the data actually show

IPOs have a well-documented pattern: a sharp initial rise, followed by sustained underperformance. The excitement of a first day ‘pop’ rarely survives contact with the medium term.

Tracking thousands of U.S. listings over several decades reveals:

• Average three-year buy-and-hold return: roughly 19%
• Broader market return over the same period: more than 20 percentage points higher
• IPOs generating a net loss over three years: nearly 60%

There will be exceptions, but the structural odds are unfavourable.

The pattern becomes more instructive when we separate companies that were profitable at listing from those that were not (defined here as those with annual revenues above $100 million):

	First-day return	Three-year performance
Large profitable issuers	11.6%	Outperforms unprofitable peers; broadly tracks the market
Large unprofitable issuers	19.4%	Underperforms the broad market and comparable listed companies

The initial excitement around loss-making companies is typically higher, but it can reverse sharply. The companies most likely to generate a first-day headline are precisely those most likely to disappoint over the medium term.

What happens when the company isn’t profitable (yet)?

Historically, companies went public to fund expansion after proving their business model worked and generated profits. Today, firms are scaling to near-trillion-dollar valuations while losing billions.

For example, despite Starlink’s roaring success, SpaceX reportedly lost nearly $5 billion in 2025 owing to a sharp increase in capital investment.⁴ Anthropic, despite its substantial revenue, was heavily unprofitable and was bracing to not turn a profit until 2028 or beyond. For OpenAI, the path to profitability is expected to be even longer. As investor concern has grown, these companies have stepped up efforts to demonstrate that profitability is possible.

How successful are investors when backing these giant, unprofitable debutantes? History offers a cautionary note.

When a company lists at a valuation that already assumes absolute global dominance, there’s almost no margin for error. If a company is valued at $900 billion before it makes its first pound of profit, it has already ‘priced in’ a decade of flawless execution. If growth slows even slightly, or if a competitor launches a superior AI model, that valuation can evaporate rapidly.

How could that come unstuck? For those companies relying heavily on AI models, there is an arms race underway to build or preserve technological dominance. Specifically, we’re seeing seemingly speculative, eye-watering sums being poured into data centres, microchips and energy infrastructure.

The AI arms race and its contradictions

Why are tech giants investing hundreds of billions in AI infrastructure if doing so actually threatens to reduce their future profitability?

The question: Why keep spending if it reduces future profitability?

↓

The trap: Every major player spends $50bn+ on AI infrastructure

↓

The consequence: Oversupply of computing power drives down the price of AI services

↓

The bind: No single company can afford to stop – falling behind means obsolescence

↓

The result for investors: Capital is consumed in a war of attrition, not deployed toward returns

The parallel with solar energy is instructive. The World Solar Energy equity index, which has tracked returns of companies in this transformative sector since its inception in 2004, began with enormous hype and formed a substantial bubble. Investors eventually came to understand how intensely participants would compete on price, making it structurally difficult for any single company to sustain high margins.

How to invest in AI

As investors, our job is to balance the excitement against structural realities: the acute information asymmetry of an IPO, the risks of investing in unproven financial models at premium prices and the broader macroeconomic tensions of the AI arms race.

The most confident investors are those who can admire a technological miracle while calmly asking: ‘Yes, but at what price?’

Sometimes the upside of structural innovation can be captured more easily or more safely in an adjacent area. Historical estimates suggest that fewer than 5% of Californian gold-rush prospectors made a meaningful profit. And yet, whether or not gold was found, a healthy profit was earned by those equipping and sheltering the prospectors.

The most obvious modern-day analogy is semiconductors – sold to AI model developers as they race to outbuild each other. But the opportunity extends further. Companies supplying electrical components to the solar industry are now benefiting from the construction of vast AI data centres, while a range of businesses work to meet those centres’ enormous power, backup and cooling requirements.

The AI revolution is real, and its transformative power is not in question. What is in question is whether the current wave of landmark public listings represents the opportunity of a generation or the peak of a cycle. The answer, most likely, is somewhere in between and depends on how, where and at what price we choose to participate.

The investors best placed to thrive will be those who apply the same rigorous, dispassionate analysis to a trillion-dollar AI company that they would to any other asset: scrutinising the valuation, interrogating the path to profitability, and asking who, precisely, is on the other side of the trade.

The gold was real. The picks and shovels were real. But most of the prospectors went home empty-handed. Transformative change and rewarding investment are not the same thing – and in this extraordinary moment, perspective is the most powerful tool an investor possesses.

---

About the author

Guy Foster

Chief Strategist

Guy joined RBC Brewin Dolphin in 2006 and has previously served as Head of Research before becoming the Chief Strategist. His responsibility covers the investment strategy, providing recommendations on tactical investment strategies to our investment managers and leads the Investment Solutions business.

---

¹Initial Public Offerings: Updated Long-run Statistics, Jay Ritter, 7th April 2026

²How Claude Is Helping Anthropic Overtake OpenAI in Enterprise AI and the $1 Trillion IPO Race, Marketwise, 2 June 2026

³Bloomberg

⁴xAI Spending Pushed SpaceX to a Nearly $5 Billion Loss, The Information, 9 April 2026

---

---