Big Tech Won't Be Monopolizing AI Evolution
While the current market rally suggests "consolidation" to some, a large number of tech players (both known and unknown) would vehemently disagree.
An expansion on macro/market commentary tends to deliver a more coherent context than the eventual soundbites can convey. Not every soundbite makes it; c’est la vie. Here’s one about AI players that didn’t make it into Bloomberg.
Which sectors right now are ripe for AI potential, and that investors should watch?
AI isn’t science fiction; an AI algorithm fundamentally analyses a large corpus of actions with “tagged” outcomes to arrive at a probable outcome for the actions it encounters that are “untagged”. Hence, sectors involving a large volume of human activity in analysing largely-predictable set of actions are very, very ripe for AI-driven optimization. There are two sectors to watch: technology services (which range from Business Process Services to Sales Management) and legal services.
Technology services firms have long-operated on razor-thin margins while legal services have long been over-priced for the amount of legal research capabilities that require to be maintained. There is significant scope for margin improvement and cost rationalization in these sectors, respectively. In fact, LexisNexis and Thomas Reuters have both been offering AI-driven tools for legal research.
Once this current hype with generative AI dies down, which stocks/sectors do you think are well positioned to emerge as long-term AI winners?
Now, Gartner did famously estimate in September 2022 that Generative AI will reach a “Plateau of Productivity” in 2 to 5 years:
While artists do quibble about the fact while the time from “imagination to creation” has been reduced to seconds, the results aren’t always quite what was envisioned. This doesn’t mean, however, that there isn’t scope for Generative AI outside the art world. As improvements continue to roll in, the hype behind its applicability will continue.
The technology services sector is a goldmine of opportunities since it comes primed with technological adaptability. Outside of the usual suspects such as Meta, Nvidia, Amazon, et al, a number of other companies have been busy hammering out AI solutions have laid substantial solid groundwork in the field, which have strong potential to boost margins in the labour-intensive tech services sector. In this year alone:
Salesforce announced a separate CEO for its AI division to continue building out the company’s journey in Generative AI in CRM via its Einstein GPT product,
Adobe announced the integration of Generative AI across its Creative Cloud suite,
Atlassian has introduced an AI-driven assistant in its collaboration software
Arista Networks introduced an AI-driven network identity service for enterprise security and IT operations
It bears noting, though, the technology services sector is a gold miner itself. Companies such as IBM, Epam Systems and Accenture have been aggressively buying up startups with targeted capabilities in AI solutions-building that will take some of the edge off the narrative that AI will be cornered by the top 7-10 American tech companies.
Even chipmakers such as Nvidia and AMD face uphill battles in the future from the likes of:
Cerebras, which released the world’s largest processor last year and seven open-source ChatGPT-like models for its Andromeda supercomputer which delivers more than 1 Exaflop of AI compute and 120 Petaflops of dense compute
Sambanova, which recently delivered open-source generative AI models after proclaiming the development of Nvidia-beating chips last year.
GraphCore, which recently invited offers for its South Wales-based datacenter 3 years after the rollout of a chip that was 16 times faster than those from Nvidia.
This doesn’t include any number of startups or conglomerates potentially running in stealth mode in India, where the government has identified both semiconductors and AI as fields where the nation must achieve self-reliance. After self-reliance comes international expansion or perhaps they will occur concurrently. Unlike Chinese companies wherein the potential for State involvement is high, Indian companies have no such potential issues and have consistently set high performance benchmarks in international corporate circles.
Do you think Tesla shares are largely sitting out the AI rally because the kind of AI techs it is trying to develop -- self-driving cars and humanoid robots -- are very different from Gen AI?
No. Tesla’s core competency in AI is yet to be proven. FSD has shown numerous operating issues while humanoid robots are hardly unusual; for instance, Boston Dynamics has marked a number of somewhat-eerie successes. The software within is the differentiator. In Tesla’s case, this is an as-yet unproven field.
Does Tesla's valuation leave any room for a further AI boost? Or does it already reflect any potential for success in its AI ventures?
Ha! Tesla already has a PE Ratio north of 50 and a PS Ratio north of 8 in an industry where the averages are slightly north of 10 and 2 respectively. Any potential for as-yet-unseen potential in its AI ventures is arguably well-ensconced within this massive gap. There are plenty of players in the AI space; there’s no reason why Tesla should be expected to be the “be all, end all” in everything.
Tesla’s recent rally is part of a phenomenon that could broadly described as the “Treasury Stock Effect”, wherein certain companies are being assumed to be irreplaceable over the mid- to long-term (i.e. “at least for now”) owing to their ubiquity and scale in modern life and commerce. Thereby, it is assumed that they have greater “survivability” in a dysfunctional economy. As I reported in my article on SeekingAlpha, a basket of ten Big Tech stocks in equal weights - that could be described as the “FAANG+ basket” - began to outpace the S&P 500 around Q3 2021:
This conviction can further be seen in the comparison of volatilities of the “FAANG+ Basket” vis-à-vis that of the S&P 500. As of 19th of May this year, trends in volatility indicate that the market tends to be around 7-9 times more volatile than the FAANG+. The lowest is around the 6X in this past quarter while the highest is at the 10-20X level seen from Q2 through Q4 of this past year.
Generally, the FAANG+ theme has tended to be less extreme and more bullish than the market in terms of price performance.
In fact, Bank of America narrowed this basket down even further in its Flow Report released in the last week of May. All of seven Big Tech stocks - Apple, Microsoft, Google, Amazon, NVIDIA, Meta Platforms and Tesla - were up 70% in equal-weighted terms in the Year Till Date as compared to the other 493 stocks in S&P 500, which was up only 0.1%.
In another article published on SeekingAlpha this week, I noted that the yield curve evolution of U.S. bond issuances across maturities in the Year Till Date (YTD) paints a fascinating (and yet familiar) picture:
Mid-term bonds, i.e. those with 5- and 7-year maturities, should have a lower yield imputed than those of longer-term bonds. A peculiar form of curve inversion is apparent: while the 10-year bonds have the lowest yields of the whole pack, the 20- and 30-year bonds have a higher yield, with the 20-year leading over the 30-year and with both higher than the 5- and 7-year bonds.
So, market consensus could be summarized thus: the U.S. economy in a 20-year horizon is more uncertain than the 30-year horizon but is less uncertain in the 5- and 7-year horizons. Of the two, the 5-year horizon is more uncertain than the 7-year horizon. However, the U.S. economy is the least uncertain in the 10-year horizon.
Intuitively, this is a highly dysfunctional picture: economies cannot be expected to turn and reverse course on a dime in such a fashion. What’s apparent, however, in the current month is that both mid- and long-term bond yields are beginning to converge, most likely toward the 4.25-4.5 range. Combining this likely convergence with the fact that short-term issuances will likely continue to attract a high yield in the near future delivers a more meaningful outlook: bond investors seem to be of the consensus view that the likelihood of the U.S. remaining the world-leading economic engine is remote. The low yield imputed on the 10-year bond is likely due to it becoming an increasingly preferred choice for long-dated exposure over the 20- and 30-year bonds, in a bid to keep capital relatively free while being compliant with Basel III Highly Liquid Quality Assets (HQLA) provisions. Going beyond a 10-year window runs the risk of significant opportunity cost.
Bonds and equities have an inverse relationship in classical finance: when equities become bearish (and lose flows), bonds get bullish (and receive more flows). Bank of America makes a rather estimate: if “real rates” were to rise another 100-150 basis points, AI's ongoing "baby bubble" would pop.
In a Financial Times article published yesterday, Nvidia’s consistent march upwards - as well as that other Big Tech stocks - was covered along with the estimate that most of the rally was driven by hedge funds taking opposite views on the company’s recent guidance wherein the company posited a massive bump in revenue due to AI adoption. Put options on the stock prior to its results are also rising.
It’s essentially a matter of time before Big Tech rationalizes with the rest of the US market as the hype dies. It cannot be reasonably expected that U.S. companies will own the “AI Wave” nor is this wave nearly as rapid as being reported. Assuming Big Tech will be the only game in town in every town around the world is a bit too hyperbolic; it has never been the case that one company (or country) maintained a complete monopoly on a “technological disruption” in perpetuity.
Note: Other such articles, i.e. expansions of my commentary, include a discussion about Korea’s market evolution and lessons on short-selling (courtesy Adani) made to “Chosun Ilbo”, the American banking crisis made to MarketWatch, and a transcript of my interview with Korea’s HankyungTV about the outlook for the Korean economy.
For a “Big Read” on matters not related to market events, the “Dharma” series - which traces the evolution of Eastern faiths across Asia over millennia - has been piling on views lately. Click here for the first part. Finally, click here for a list of all articles published till date.