The bizarre comings and goings at OpenAI have been dominating business headlines recently. I must admit, I find the ongoing delirium surrounding artificial intelligence somewhat tiring. While I’m certainly not qualified to debate the finer details of the technology, it seems to me that for most people, the extent of it so far amounts to firing their stream of consciousness into a chat bot and making silly pictures. Of course, this negativity has nothing to do with the fact that I have watched shares in NVIDIA triple in value since I sold them, or that the bulk of the technology is beyond my comprehension, but I digress.
Credit where credit is due, the AI extravaganza has almost single handedly given US mega caps, the recently ordained ‘magnificent 7’, a new lease of life. Led by poster child NVIDIA, the market cap gains in the space have been nothing short of sensational. The trajectory of the AI trade can be traced back to the company’s groundbreaking first quarter results where they announced an astonishing 50% increase in revenue guidance, leading to one of the largest single session market cap increases in recorded history. This momentum has continued apace, with its just released Q3 results recording sales growth of 206% year on year and 34% sequentially, a result which one strategist aptly described as “so stupendous as to make comparisons meaningless for the purposes of gleaning anything beyond the self-evident conclusion that things must be going well”.
This makes me think that perhaps I am missing the point, and my characterisation of AI made above may be slightly off the mark as it seems unlikely that making cartoon pictures of my cat would generate $18bn in revenue in 3 months (as NVIDIA has done). Although I had similar reservations around web3, and crypto, (which continue to make an inauspicious fist of things, see here, and here for a taste), these other technological ‘innovations’, continue to be priced on nebulous theories of de-dollarisation and the metaverse, while broadly speaking, the AI trade and NVIDIA can be valued in dollars and cents. This begs the question, are NVIDIA and the other AI behemoths existing in a bubble of epic proportions, or is this really the internet 2.0, or, as exclaimed by one excitable proponent, ‘the new fire’.
On the surface, headline multiples for these companies appear daunting, with NVIDIA, trading on an eye water trailing PE of 83 (and even this is likely generous given it uses non-GAAP earnings, excluding things like stock-based compensation, which is how these types of companies tend to capture the best and brightest while conveniently keeping the cost off book). Valuations for contemporaries, Microsoft, Amazon, and Google (via their computing platforms, Azure, Web Services, and Cloud), are less clear given they are either lumped in with the broader conglomerates, or are still focused predominantly on top line growth. Nominally however, all three have seen significant multiple expansion coinciding with the AI rally that has pushed valuations well ahead of recent averages.
As any tech investor will tell you however, these companies, especially NVIDIA, are valued not on what they are doing now, but what they may do in the future, and if their performance over the last six months is anything to go by, that future is looking increasingly bright. Current market pricing has NVIDIA growing earnings by 260% for 2024 and a further 65% in 2025, which would see the roughly $3 per share of earnings today turn into more than $20 in the next two financial years. That rate of growth is astounding, and you could argue, actually makes NVIDIA cheaper than the market at current prices, relative to its potential growth (on a price to earnings growth multiple). It’s a similar story for the other tech titans (although much less dramatic given their existing scale), which, based on these same calculations, are also cheaper now than they have been in recent years despite all three steadily retracing their previous all-time highs.
The success of Microsoft’s Azure is particularly notable with the company’s computing platform dragging the not so lumbering giant into the 21st century. Of course, its software business stands as one of the world’s most dominant monopolies, providing the cornerstone of a credit rating which is, somewhat ironically, higher than that of the US government. However, its cloud native Azure is now, by a stretch, the dominant growth engine of the company, expanding twice as fast as the wider business and around five times faster than the broader market. Granted, only a fraction of these revenues are directly tied to AI, however its contribution is accelerating much more rapidly than initially expected, thanks mostly to Microsoft’s $13bn investment in OpenAI (the company that brought you ChatGPT, and all those fun cat pictures).
The scope for this to grow is substantial as the technology is integrated into the company’s product suite including Excel, Word, and Teams. The development is described by investment bank UBS as “easily the most anticipated GenAI-based software application launch”, and should provide ‘technologically challenged’ users like myself an opportunity to experience first-hand the much touted productivity benefits that are expected to drive these massive earnings growth expectations.
The trillion-dollar question is if these forecasts are both accurate and sustainable which very much depends on your view of the broader AI ecosystem, and for NVIDIA, how dominant the company’s well documented technological lead really is.
One trust which has put its money where its mouth is, Manchester & London (MNL), has almost 50% of its NAV across NVIDIA and MSFT alone, while ensuring the rest of its portfolio has AI core and central to its business purpose. The trust defines the progress of AI so far as “embryonic compared to its immense, era defining potential”, while predicting that ‘’in a couple of years, many investors will rue their underweight position in NVIDIA today’’. While the managers acknowledge that this allocation naturally brings with it significant concentration risk, they counter that the winners from the AI era may, in time, be counted “on the fingers of two hands”, so diversification would be the antithesis of their bullishness. The Polar Capital Technology Trust (PCT) has favoured a watered-down version of this approach, holding a still chunky 18% of its NAV directly invested in the two companies, in addition to large positions in Alphabet, Amazon, Advanced Micro Devices, and TSMC (which all boast varying degrees of exposure to the AI trade).
Other funds, such as Herald Investment Trust (HRI) would argue not against the overarching premise of AI, which it believes too to be a generational paradigm shift, but that picking the winners may not be so straightforward. The trusts instead favour the ‘picks and shovels’ end of the market, with significant investments in semi-conductor development and manufacturing, and the infrastructure that allows the technology to flourish. Private equity infrastructure trust Pantheon Infrastructure (PINT) is another targeting this end of the spectrum with substantial investments in digital infrastructure, such as data centres and transmission networks. PINT notes that the already impressive growth in demand for data centres will be driven higher by AI. There is also additional revenue to be extracted as data centres are retrofitted to cope with the additional cooling requirements that AI chips need.
There are certainly arguments to be made for both cases. The cautionary tale of Cisco, which for years was at the forefront of the internet revolution before collapsing under its own weight, stands as a reminder of counting your chickens. Today, shares of the company still trade below their peak reached back at the turn of the century. On the other hand, as we discussed above, it is entirely possible that companies like NVIDIA are actually still cheap, particularly if you believe, as MNL does, that we’re still very much in the earlier innings of this revolution.
Considering my trading record on NVIDIA, I’m hardly in a position to say, one way or another, however as someone who likes to have their cake and eat it too, I’ll suggest that diving headfirst into market euphoria is often a recipe for disaster, regardless of the technological potential. There’s every chance however that, as MNL so colourfully put it, I’m just a braying, misinformed Luddite.