Merlin Weekly Macro: AI, mega-IPOs and the indexation trap

Star Date 12-06-26; Space, the final frontier

Elon Musk wants us all to be space cadets. The Initial Public Offering (IPO) of his Space X satellite communications and space exploration company blasts off on June 12^th. Raising $75 billion of new capital, estimates of its potential market capitalisation range from $750 billion to $1.8 trillion. That is a very big gap, one through which you can launch a space rocket sideways.

The company has real assets; particularly through its successful Starlink subsidiary with its “internet constellation” of more than 10,000 satellites in orbit, it already provides critical communications infrastructure. The argument is about the future: how much to pay for a business that recorded $18.7 billion of revenues in 2025, has an income stream which is growing rapidly, but which as a group generated a loss of nearly $5 billion and has big cash outflows while it invests at a current rate of $40 billion a year in AI technology whose financial returns are speculative? The $75 billion it is raising covers less than two years of capital investment at the current burn rate.

We are certainly not offering any recommendation for or against SpaceX or its IPO. There are arguments on both sides. However, the estimated 100% bottom-to-top valuation gap reflects the widely differing opinions of everything in the investment melting pot: corporate governance, key man risk, the bet on the mercurial Musk’s genius but also how a non-executive board manages both the behaviour of such a maverick and the inherent complications arising from his business conflicts of interest; it is a judgement call on the quality of the management team implementing the strategic vision; an understanding of the marketplace and the rapidly changing technology and, critically, not only whether the management can turn aspiration into reality, but also make acceptable financial returns from doing so.

Mega IPOs: exploiting the indexation rules to maximum advantage

The greater the value ascribed by the market when listed, the greater the influence of the company through being included in the principal US and global equity indices including MSCI, S&P, FTSE and NASDAQ. Those who invest through trackers will become space cadets whether they like it or not. Index tracker funds will be forced to own the company: through pushing the rule boundaries on early inclusion, Space X’s advisers have ensured it will go straight into the indices.

Algorithmic purchases by their nature create indiscriminate almost insatiable demand for the shares which feeds an upward vortex in the absence of sellers who tend to hold on to their shares into a rising price trend. “Indiscriminate” here is used in the sense that the appetite of trackers to own the shares is purely driven by the need to maintain an index weighting, there is no discriminating quantitative or qualitative analysis of whether the company is a good business or not. In this case, it is highly likely that size-driven share price momentum created by passive trackers will far outweigh the nuanced opinions of active fund managers who have performed a critical analysis of the company and its faculties.

Space X is likely to be joined in the “mega-IPO” bonanza by other companies racing to develop and exploit the seemingly limitless potential for artificial intelligence in all its iterations; the highest profile of those whose listing intentions are already known include Anthropic (the company whose Claude AI business facilitator service has exposed the extreme vulnerability of all web-based technologies to hacking and data appropriation) and Open AI, the creator of ChatGPT, the indispensable text writing tool employed by almost every modern student to write their dissertation for them. The same indexation factors will be at work as those apparent in the Space X public offering.

The history of trackers and passive investment vehicles says that many (but certainly not all) have made investors a lot of money, so what’s the problem? It is a powerful argument. However, whether this is either a rational, efficient or healthy means of allocating capital is a moot point and a very different argument.

When the tail wags the dog

Indexation is also having a direct effect on corporate governance. In the main, as stewards and owners of companies rather than purely investors in them, most institutional shareholders tend to prefer executive performance related bonuses to be aligned with how efficiently and effectively the management team runs the company: rewards are empirically tied to areas that the management controls (e.g. Return on Invested Capital in which improving profits and margins as the numerator in the calculation are directly linked to balance sheet efficiency and capital allocation as the denominator).

This week it was reported that NASDAQ-listed ARM Holdings PLC (semiconductors, roughly 90% owned by Softbank but with a modest publicly traded free float remaining) has offered its CEO an incentive package linked directly to the company’s market capitalisation: in annually graduated increments he will be paid in shares worth potentially up to $800m if the company’s market cap hits $2 trillion by March 2031. It is currently capitalised at $420 billion. Supporters will argue that making shareholders nearly five times better off will be worth every penny. Detractors will point to the temptation towards the likelihood of short-termism in management decisions and in any case it is markets which determine share prices, not management teams; multiple expansion in the valuation driven simply by passive investment indexation momentum should not be rewarded so egregiously for measures that cannot be attributed directly to the CEO’s personal performance.

Booms and bubbles: history revisited

The current investment obsession with artificial intelligence is significantly skewing share prices. It is not new: in recent years we have had the nominal and myopic focus on the US “Magnificent 7” companies (Alphabet/Google, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla) and their immediate predecessors, the FAANGs (Facebook—which became Meta, Amazon, Apple, Netflix and Google); in 1997-2000 it was the Dotcom and TMT (tech, media, telecoms) bubbles; stretching even further into history, phenomena such as the Nifty Fifty quality growth companies dominated US equity investing in the 1960s and 1970s. In the Days of Yore, there was an investment bubble in the South Sea, a mania with Dutch tulips, a precipitous rush into gold mining in the Klondike and others. Gold and crypto currencies still provide their own periods of intense excitement. Results range from lifechanging fortunes being made to some investors losing their shirts.

But for active portfolio managers, the modern effect of indexation momentum is a pernicious, accelerating and dominating challenge. It is particularly so when companies’ market valuations keep rising when the manager’s own analysis and judgement all point towards avoiding the shares. Managers in such a cleft stick lose absolute performance without those in-vogue momentum stocks in the portfolio; relative performance suffers against other managers taking advantage of the trade. The bigger the company concerned, the more painful the experience. FOMO can become a nerve-shredding factor.

Vodafone: TMT’s Grand Old Duke of York

UK mobile phone operator Vodafone illustrates the point perfectly. An undoubted world-class leader in the field of mobile telephony as it expanded its network and operations, and indisputably a “proper” business, at the zenith of the Dotcom and TMT mania at the dawn of the Millennium the principal question was whether in an increasingly competitive and saturated field, the company was over-paying for new 3G bandwidth licences being auctioned by national governments. Further, whether Vodafone’s $190 billion acquisition of German conglomerate Mannesmann in February 2000 (at that point the largest merger in history) was a) strategically sound and b) worth the price.

In March 2000 Vodafone on its own accounted for 14% of the FTSE100 total market capitalisation when its price reached the equivalent of 526p (the company had a share consolidation in July 2006, the quoted price in March 2000 was 400p). On a like-for-like basis, the shares hit a low of 63p in February 2024. Shareholders experienced value destruction of almost 90% over quarter of a century, more when aggregate capital expenditures are considered and even allowing for the benefit of dividends. Trackers owned it all the way up…….and all the way down.

At the time, Vodafone sceptics were castigated for failing to take advantage of a FTSE100 share price which had rocketed five-fold in little more than two years from late 1997; for some active managers, it was a career-limiting event for whom being proven correct on fundamental analysis grounds was of little consolation (in a case of double jeopardy, one notable unfortunate, a highly experienced manager who today would be described as a “conviction value investor”, got the TMT boom wrong both ways: deeply sceptical of the hype and rocketing valuations he eschewed such companies all the way up to the point of maximum pain, then capitulated buying into the sector at its peak and rode it down the reverse slope; his career never recovered but he was for ever more known as the one who for all the wrong reasons “rang the bell” that called the top).

History might not repeat itself but it does rhyme

In a case of back to the future, many of the same debates are taking place today as were happening in the TMT bubble. Similar to quarter of a century ago, there is talk of a “new paradigm”. Now, it is the revolutionary (or is it evolutionary?) effect of AI on all our lives. As with the advent of the internet, from an investment perspective intense focus is on the technology sector. To use the analogy of the Gold Rush, is there greater financial advantage at the cutting edge of technology among the prospectors (in the lingo, those developing proprietary “ecosystems”) with its much greater volatility of risk and reward? Or is it in the supply chain among the less glamorous but more stable “picks and shovels” suppliers (e.g. chip manufacturers and infrastructure providers).

Making such decisions requires a deep technological understanding in a highly complex environment that is pushing unimaginable boundaries, where commercial and technological innovations are jealously guarded and in an industry in which technological change is happening at lightning speed. And applying displacement theory, which companies and sectors lose from AI as their businesses become redundant?

In the 1990s in a former life as an investment bank equity sector analyst, your correspondent specialised in the Paper, Packaging & Printing sector and its universe of companies. As an illustration of the rapid development of digital technology and the internet, at the time one of the biggest debates was who were the winners and the losers of that industrial revolution. History says that print services changed out of all recognition and companies either adapted or went out of business; reading and communicating habits changed significantly but a look at my desk will reveal the paperless society never happened (and once seen as doomed, bookshops selling physical books are in strong revival); packaging was much more affected by environmental policy than changes wrought by digitisation. In a very old-world, capital intensive, analogue sector, the biggest and least likely winner of the digital age and the new paradigm of internet shopping was a very old-fashioned, very unsexy but ultimately utterly indispensable product: as any former shareholder in David S. Smith PLC will tell you, it was the humble brown corrugated cardboard box.

“We can’t rewind we’ve gone too far”

In the investment industry and our world of Mammon, artificial intelligence is almost the only story in town: how to apply it to competitive advantage and how to make money from it. But such is the technology’s potential that questions are being raised that if left unconstrained, does it pose an existential threat to mankind. Data is power, for good or evil. To own it and manipulate it has the potential to be as corrupting as it is constructive. The Pope has raised concerns about the effects of disintermediating and displacing humanity; visiting China, Yvette Cooper, the Foreign Secretary, argued the urgency for global regulation about both AI’s development and application (how do you regulate concepts yet to be discovered?).

But it was The Buggles who understood. Nearly half a century ago they saw the inevitability of all this coming, predicting Bots and the potential for the family car to be a data-mining machine. It was all in their pop hit, Video Killed the Radio Star: “They took the credit for your second symphony/Rewritten by machine on new technology/And now I understand the problems you can see………….In my mind and in my car/We can’t rewind we’ve gone too far…..Video killed the radio star”. QED.

Last word: NOT a Bot

The author of these weekly macro musings is Merlin Investment Director Alastair Irvine. It will not be lost on readers that his initials are……AI. Rest assured, he’s no Bot: he’s the real deal!

The Jupiter Merlin Portfolios are long-term investments; they are certainly not immune from market volatility, but they are expected to be less volatile over time, commensurate with the risk tolerance of each.  With liquidity uppermost in our mind, we seek to invest in funds run by experienced managers with a blend of styles but who share our core philosophy of trying to capture good performance in buoyant markets while minimising as far as possible the risk of losses in more challenging conditions.