2023 Theme: The Big Getting Bigger, Tech Edition
We were thinking about one of our favorite themes - The Big Getting Bigger - recently and realized that there was an important aspect of this trend that we hadn't appreciated.
In brief, we had completely ignored an important reason driving at least a portion of the rise in big technology share prices this year.
By now, everyone knows artificial intelligence is expected to be the "next big thing." But providing it happens (it might not!), what is a little less understood is why the biggest tech companies - other than elite chip maker Nvidia - are supposed to be the biggest beneficiaries of this trend.
That is unusual and also reveals something important to remember about this latest development in the forward march of technology and progress. As Ted Kaczynski noted some 30 years ago, human societies are far more concerned about who OWNS the ability of machines to think for themselves than the scary reality of machines being able to think for themselves.
But first, a recap:
As regular readers will recall we first introduced this Big Getting Bigger theme (trademark pending) during the late winter and early spring. We principally framed it originally as something that was being pushed - inadvertently perhaps - by the US government and its desire for a new industrial policy.
THEN we added to this theme in the wake of the banking failures in March by pointing out that - once again unintentionally perhaps - both the crisis AND our rescue approach AND our regulatory push were only going to make American banks still bigger. This was ironic, of course, since the very same regulators and politicians talk frequently about the importance of "breaking up" these very same banks.
Lastly, we pointed out that both of the above points were being reinforced by investors fleeing to the largest high growth companies for a variety of reasons that were, over time, becoming self-reinforcing.
On this last point, here was the Big Tech Super 7 trajectory faithfully recorded in these pages:
Investors bought Microsoft in February because it would benefit in an absolute sense from the boom in expected profits from advances in artificial intelligence.
Then in March and April people bought the computer company's shares because it was expected to do relatively well in an economic downturn.
And then when that didn't transpire by May and June they were buying because, well, it had done really well and people had serious F.O.M.O.
FOMO = Fear. Of. Missing. Out.
It is worth remembering all this for the future:
A company's share price starts going up for one reason or a few and can then continue to go up because it achieves a momentum of its own and investors and algorithms can continue to buy sending its price ever more into "irrational" territory.
So, the next time you look at a company's share price and think "this is nuts, when's the crash" ask yourself whether the firm has the above type of momentum behind it and if so, it might be awhile.
This was our previous work on "The Big Getting Bigger" but now we turn to the technology sector.
Previously, what we did NOT cover directly was the fact that today's boom in AI companies is pretty different from earlier waves of tech-led disruption. At the center of this is the fact that small and disruptive "Generative AI" companies have a structural Achilles Heel that is being eagerly exploited by largest technology companies and is worth fleshing out.
In theory, disruptive technological innovation usually doesn't happen in large companies. In fact, it almost never does.
Rather, the cliche is that only small, agile and fast moving companies can innovate quick enough and take the necessary risks to really come up with a significant challenge to existing incumbents.
From one perspective, this is happening this time as well. Small and nimble start-ups like OpenAI, Cohere, Anthropic and Inflection AI have built incredible products based on large language models that are as good if not better than anything the large tech behemoths like Google or Meta (Facebook) have managed to produce.
But there are two wrinkles to this narrative:
The first is that the largest tech firms are benefiting as much if not more than the small AI start ups.
Second and perhaps most interesting of all, is WHY they are doing so.
As this newsletter has discussed at length, the biggest tech companies are doing extraordinarily well this year. This is sort of puzzling, in a way.
Baked into this performance has been an expectation that they could make even more profits by cornering the next generation of artificial intelligence-led products. That is weird since they are, of course, not really building those products.
Now from one perspective you could look at this and simply think, well, this assumption could be wrong! And it could be. There is no real reason that the likes of Alphabet (parent company of Google) or Microsoft necessarily have to win the race to corner AI profits.
In some ways, many investors seem to agree. This is why we have written about how Nvidia - the shovel seller - has done far better than the gold miners like Meta, Microsoft etc. In any gold rush, it is simply a better idea to bet on the profits of the shovel company than whether even the largest and most skilled gold miner will hit pay dirt.
But regardless of those risks, lots of investors are still betting that the profits at Microsoft or Alphabet (Google) or Meta will be meaningfully higher.
This brings us to the WHY aspect of all this, which is by far the most intriguing and also what we missed discussing in our earlier efforts.
One of the interesting reasons that big tech companies are assumed to profit from the "AI boom" is because they have signed partnerships with many of the best of the new generation of "generative" AI start ups.
These partnerships usually involve a straight up trade of an investment of big company cash and computing power for equity in the company and access to the new AI products and capabilities that these start ups are busy building.
There are two reasons that these partnerships have occurred.
The first is that generative AI technology is actually quite different from other recent trends in software technology in one key respect: they are capital intensive.
The second is that they require access not just to large amounts of money bit also tremendously sophisticated computer chips and computing power.
On the first point, generative artificial intelligence companies have a desperate need for lots and lots of capital. This is very different from other software companies over the last few decades that could employ a small number of engineers and have a hope of developing a product that can scale with their customers.
This isn't how generative AI start ups can operate. They first need to train their models on enormous amounts of data whether it be text or sound or images or what have you.
To simplify somewhat crudely, these new ventures cannot make advances with just a little bit of money and then receive another round of investment to get to the next stage. Instead, right from the get go these small firms need sizable amounts of capital - measured in the billions - to train their large language models to produce even the first generation of a working product.
In this respect these new ventures might have more to do with Uber or Tesla or other companies that were forced to spend big to get big and do so racing against the clock pincer move of regulation one side and competitors on the other. For years the question was whether Tesla could become profitable or whether it would run out cash and everything would come crashing down with legacy car companies smugly dancing on their grave.
Understandably, these tiny artificial intelligence companies would rather avoid that stress and uncertainty and so are signing partnerships now to try and derisk themselves with regards to the critical vulnerabilities they have as businesses.
A significant partnership with a Big Tech company does that. It also aligns a Google or a Microsoft to being interested in their continued existence rather than viewing them as potential rivals to be squashed.
Second, and perhaps more importantly still, these companies require incredibly expensive sophisticated computer chips. This is, after all, why Nvidia's (and other chip companies) share price has done so well.
The problem is having money isn't necessarily sufficient to get access to a large number of these processors. It helps to be a long standing and fabulously wealthy customer if you want to be able to buy thousands and thousands of Nvidia H100 or A100 processors.
So, having the money isn't sufficient for building the product of their Generative AI dreams. They need the access that only a larger and powerful tech firm with pricing power, significant market influence and especially long standing relationships with the few chip companies capable of satisfying their requirements.
It isn't just that investors are assuming: Google/Microsoft/Meta are smart and rich and will eventually either figure this out or buy companies and thereby come to dominate this space in the same way they have with email (Gmail) or encrypted messaging (WhatsApp) or cloud computing (Microsoft Azure).
Rather it is the very nature of this new class of Generative AI start ups that structurally require some sort of large, deep pocketed company that has access to both computing power and ideally also huge volumes of data.
The big US tech companies obviously fit that bill and are quite happy to gain privileged access to what could be the next generation of an incredibly advanced technology and one with disturbingly powerful implications.
Even better the Big Tech firms are also giving money to companies that will, almost by default, turn around and spend that money on the Big Tech companies' products and services. After all these companies will need cloud computing and other services that both Alphabet (Google) and Microsoft supply.
This is a "win-win" no doubt makes the (sizable) investments that Microsoft made into OpenAI or Google made into Anthropic to be among the easiest financial decisions these companies will make this year.
It is really just a vote to give themselves more business and one with a potentially huge payoff down the road:
Heads you get a reliable stream of new business and tails you get that and potentially game changing new technological advance that will further entrench your dominance and incredible profits.
Is it any wonder that their stock prices are up so much?
*******
Have questions? Care to find out more? Feel free to reach out at contact@pebble.finance or join our Slack community to meet more like-minded individuals and see what we are talking about today. All are welcome.
