Dear investors and well-wishers,
The fund returned 20% in January.
Multiple waves of change have rippled through the tech industry over the past two decades. E-commerce, search, mobile apps, social media, SAAS, and the cloud were each revolutionary in their own way. Even the largest companies had to adapt and flourish or in many cases suffer almost total loss of market value.
The leaders of each wave drove investment returns for years, and were often the major contributors to market indices, barring the last year or so of course.
It’s increasingly clear there is a new pillar of growth: artificial intelligence and specifically large language models like ChatGPT. In cultural terms, this makes the other advances look like stepping stones.
ChatGPT has already tapped immense consumer demand
An obvious obvious example is Amazon versus Sears (as a stand-in for physical retail).
This chart starts at the end of the dotcom bubble, one of the worst times ever to buy a stock like Amazon
The leadership shift from Nokia and Blackberry to Apple was another multi year trend with trillion dollar shifts in value.
Catching shifts like this was the original driver of our investment philosophy
You didn’t need to be early to any of these movements as they lasted years. And it was usually the obvious play that was the right one.
Social media was similar – many hits and misses, but the market leader generally performed the best, notwithstanding the latest hiccup.
Growth, trends, and also price action were clues, this final aspect was something we neglected to immense cost last year.
Now after a once in a decade tech washout that is still in motion, a new set of clear winners is emerging.
A veritable rainforest of new AI startups is being freshly funded, and while it won’t be clear for some time which will win and which will fall (and in each market many more will be funded than can succeed), one thing is certain: they will all going to need immense amounts of computing power, power that absent this revolution may not have been required for many years.
Best-in-class models trained on static data two years ago will soon be training on a continuous basis, taking in the bulk of human production of digital and physical words and images and calculating the weights and relationships between each word.
And once these models are trained, instead of simply looking up a key-value pair when a user makes a request, answers will need to be calculated, more-or-less word-by-word, pixel-by-pixel, or beat-by-beat, driving a step-change in demand as they move from research to commercial scale.
NVIDIA is a clear winner, announcing last week their supercomputers would be available on demand through the internet, allowing anyone access to their AI models via a subscription, cloud-based model. Notwithstanding a collapse in phones, PCs, gaming and cryptocurrency, the company is profitable and buying back shares.
And gaming may not be dead forever – a new game ‘Hogwarts Legacy’ just returned $850 million in its first two weeks.
NVIDIA is partnering with Mercedes to power their next generation of cars from 2024, including level 4 autonomy, over-the-air updates, and best-in-class safety and convenience applications.
NVIDIA is a small position for us, bought at lower prices, but their 81% market share in AI processing is now more relevant than ever.
ASML has been our second largest position for some time now, holding total market share in the extreme ultraviolet (EUV) lithography machines that make the most advanced chips.
ASML has total market share in some of the world’s most strategic technology
These chips require radiation similar to X-rays emanating from the sun. The shorter the wavelength, the higher the resolution, so extraordinary effort has been made to recreate these helical conditions.
In the heart of ASML’s machines, two lasers are fired in quick succession at tiny droplets in a spray of tin. The first laser flattens the droplet into a pancake, the second vaporises it into searing hot plasma that generates extreme ultraviolet light.
Each droplet is travelling at a staggering 70 meters/second and has to be hit by lasers twice, 50,000 times per second.
The EUV then passes through extremely fine Zeiss lenses and mirrors – the flattest surfaces generated in human history – to hit a wafer, which adjusts its position 20,000 times per second with nanometer accuracy. This precision allows the printing of 100,000 transistors per square millimeter, and enables the latest generation of our phones and computers, not to mention advanced defence equipment.
ASML EUV machines
The next generation lithography machines are expected to cost ~A$400m each. Despite the high cost, our theory demand for such critical, strategic assets would remain strong has held true: ASML is on track to post 40% revenue growth for EUVs in 2023, with non-EUV revenue growth coming in at 30% (given a chock-full orderbook, ASML has decent near-term visibility). ASML is continuing with a $12 billion buyback.
Crocs reported a whopping $850M in GAAP operating income for the year. Cashflow was used to pay down debt, and buybacks should switch on soon. The firm has a good record here: Since 2014 they bought back $1.7 billion of stock at an average price of $37.9.
Like many small cap companies, Crocs sold off 75% last year, but there has been a clear break with the outcome of former growth darlings in recent months.
Acquisitions made pre-crash have mostly been disastrous for growth companies. Obvious losers include Shopify’s acquisition of Deliverr, moving them from capital-lite SAAS to capital intensive logistics. Carvana spent $2.2 billion on Odesa – cash it now desperately needs. And ofcourse Square bought Afterpay for $39 billion, only to report over $500 million of BNPL losses last week for the privilege.
So it’s nice to see Crocs $2.5 billion acquisition of HeyDude is performing well, with revenues of over $1 billion and income of $250 million.
There’s clear dispersion amongst small and mid cap companies, with profitable growers like Crocs and semiconductors so far outperforming growth favorites in say, software.
There are only a handful of moments where new technologies break into the mainstream with the force that we’ve now witnessed from large language models.
Each wave – and we’ve all been lucky to see a number over the last two decades – brought extraordinary investment returns. These mostly came from backing the clear market leaders driving each space forward. I don’t know how it will play out, but I do know AI is startlingly effective and outcompeting the best human experts in a rapidly increasing number of domains.
In some ways this has all happened in reverse: it wasn’t so long ago the intellectually fashionable worried about the plight of truck drivers, but the most impressive advances have come in coding, art and writing. The areas thought to be most resistant are falling first. It’s a very exciting, if disorientating field-levelling moment.
As with prior revolutions some things are clearer than others. Many years ago you could see market share being taken by Amazon, and conclude the revenues were coming from somewhere else. Queues outside Apple storess led to separate profitable conclusions. Often the obvious play was the best.
Now we know that there’s going to be immense capital flooding into AI, and we know that the best of these tools can capture a hundred million users in a matter of months.
We also know that regardless of who wins in each space, every use case will require substantial computing power, and it just so happens that there are highly profitable monopolies over critial parts of the tech stack, notably ASML in lithography and NVDA and AMD in GPUs. The same way Microsoft missed social media, mobile, and search (for a while), Intel has missed the major trends in semis, leaving the majority of the field to Taiwan Semiconductor and to a lesser extent Samsung (it remains to be seen if Intel can turn around – but the lesson from prior shifts was to make the obvious bet). There are also a surprisingly small number of firms building the critical software, raw silicon wafers, lenses, mirrors, quality control systems, and the other key components. After a severe tech crash and slump in consumer demand, these companies are trading cheap.
The information in this note has been prepared and issued by Frazis Capital Partners Pty Ltd ABN 16 625 521 986 as a corporate authorised representative (CAR No. 1263393) of Frazis Capital Management Pty Ltd ABN 91 638 965 910 AFSL 521445. The Frazis Fund is open to wholesale investors only, as defined in the Corporations Act 2001 (Cth). The Company is not authorised to provide financial product advice to retail clients and information provided does not constitute financial product advice to retail clients.
The information provided is for general information purposes only, and does not take into account the personal circumstances or needs of investors. The Company and its directors or employees or associates will use their endeavours to ensure that the information is accurate as at the time of its publication. Notwithstanding this, the Company excludes any representation or warranty as to the accuracy, reliability, or completeness of the information contained on the company website and published documents.
The past results of the Company’s investment strategy do not necessarily guarantee the future performance or profitability of any investment strategies devised or suggested by the Company.
The Company, and its directors or employees or associates, do not guarantee the performance of any financial product or investment decision made in reliance of any material in this document. The Company does not accept any loss or liability which may be suffered by a reader of this document.