AI looks likely to cause major changes to society over the next decade.
Financial markets have mostly not reacted to this forecast yet. I expect
it will be at least a few months, maybe even years, before markets have
a large reaction to AI. I'd much rather buy too early than too late, so
I'm trying to reposition my investments this winter to prepare for AI.
This post will focus on scenarios where AI reaches roughly human levels
sometime around 2030 to 2035, and has effects that are at most 10 times
as dramatic as the industrial revolution. I'm not confident that such
scenarios are realistic. I'm only saying that they're plausible enough
to affect my investment strategies.
Google's DeepMind seems a bit more likely than any other company to
produce valuable AI. That's a good enough reason to hold a modest
position in Google. TPU and
Waymo are additional AI-related
reasons for holding Google stock.
OpenAI is a serious contender. Microsoft has a $1 billion stake in
OpenAI. But profits from that seem to be capped at
100x, so the best case scenario
seems to be that Microsoft shareholders get $13.40 per share from that
stake. Maybe Microsoft will invest more in OpenAI. I only see Microsoft
getting a modest boost from AI.
Conjecture sounds quite
promising. I offered to
invest $100k in them this summer. They don't seem interested in
investments that small.
There are other new private companies that I haven't been able to
evaluate. Here are some that seem worth paying some attention to:
That's a dramatic pace of new startups being founded that look
potentially important. VC's seem to have been throwing enough money at
them in mid-2022 that they likely had little reason to talk to more
ordinary investors such as myself. Investment in such startups has maybe
cooled slightly now that FTX has stopped throwing money at them, but
I'm guessing it hasn't cooled much.
I see a 50% chance that one or more of these new startups will make
DeepMind and OpenAI irrelevant. I have little hope of being able to buy
a diversified portfolio of such startups, so I don't expect to profit
from direct investments in companies working on AI.
The most promising investments involve the hardware needed to power AI.
That mostly means semiconductors.
Semiconductor capital equipment companies seem more promising than
companies that are more directly involved in making chips, as capital
equipment is more cyclical, and less dependent on specific uses.
My current semiconductor-related investments are (in alphabetic order):
AMKR, AOSL, ASML, ASYS, KLAC, MTRN, LSE:SMSN, TRT.
I'm likely to buy more sometime in 2023, but I'm being patient because
we're likely at a poor part of an industry cycle.
Others that I'm considering buying: ACLS, AMAT, INTC, LRCX, MU, and
Some of you will be surprised that I didn't suggest NVDA. Its PE and
price/sales ratio are high compared to my other semiconductor
investments. I can imagine buying it if its stock price drops relative
to other semiconductor stocks. But I suspect its fans underestimate the
risk of competition from hardware that is optimized more specifically
for deep learning.
I've invested in privately held Fathom
Radiant, which seems to have a shot at
NVIDIA-like success in AI-related hardware. I'm unsure whether they'll
want any more investment in 2023.
I've seen some noise about neuromorphic computing. I suspect that such
research is far enough from commercialization that it's hard to figure
out how to invest in it. (Also, I'm nervous about helping to speed up
AI, as it's already coming fast enough to worry me.)
I have a tiny position in Everspin Technologies (MRAM), mainly because
of vague suggestions that their technology will support neuromorphic
computing. I haven't found much substance to that speculation, so I
think of it as a real long-shot. Maybe a 2% chance of a 500x return this
decade, without a high risk of the stock becoming worthless?
INTC also does some neuromorphic work, and seems to be a relatively safe
investment even if it lags behind the cutting edge.
Datacenters seem likely to become a major industry, but I don't see
great investments here.
AMZN, MSFT, and GOOGL are likely to get some benefit from datacenter
growth. I guess I'll expand my GOOGL holdings and buy small positions
in AMZN and MSFT sometime in 2023, when I see signs that their stock's
downward momentum has dissipated.
EQIX is an example of a company whose revenue growth seems likely to
accelerate. But that will be capital-intensive, causing the company to
sell enough shares that revenue per share growth doesn't look very
impressive. Also, its PE is high.
AAOI is a floundering company that gets over 30% of its revenues from
selling products to datacenters. It might recover and profit from
datacenter growth, but I'll wait for its losses to shrink before
Semiconductor stocks seem to be strictly better ways of benefiting from
datacenter growth than datacenter-specific stocks.
AI seems to be on the verge of turning robotics into a major industry.
But it's still hard to see when and where the key advances will happen.
I suspect the main winners will be companies that aren't yet public, as
I'm not too impressed by the opportunities I see so far. I'm playing
this mainly via tiny positions in LIDAR companies (INVZ, OUST, LAZR) and
I have a moderately large position in
whose ruggedized computing products have some robotruck applications. I
should look more carefully for other companies that make
Electricity use is likely to become a larger fraction of the economy as
AI takes off. AI is energy-intensive, and will catalyze much more
Utility-scale solar seems likely to supply an important fraction of that
My two biggest solar bets are CSIQ, and SCIA. I have smaller positions
in JKS, DQ, SEHK:1799.
For power grid infrastructure and solar farm construction, I have
positions in MTZ, MYRG, PLPC, and PRIM.
I consider it somewhat likely that nuclear energy will somehow grow more
important as a result of AI, but I'm pretty fuzzy on the details.
Fusion is likely to be feasible in the 2030s, but I have little idea as
to whether it will be competitive.
For fission, there's SMR, uranium mining companies, and uranium
futures. I haven't invested in these, and I'm not in any hurry to
decide which of these to buy.
Here are some weak guesses about how AI will affect existing companies
that make up significant parts of current portfolios, by industry:
We should also think about the effects of an AI-induced general
Some of the biggest winners in that scenario will be commodified
industries where capacity is slowest to change: oil and gas, mining, and
I see many changes coming that are hard for investors to exploit.
For example, many forms of therapy seem ripe for being replaced by AI.
Most patients can't find top-notch therapists. Therapists have little
ability to document their successes. Even when they are able to get a
reputation as a great therapist, they can't scale up to serve more than
a tiny fraction of the clients that need them.
AI can eliminate the latter problem. The reputation problem is much
harder. If one out of every 10,000 patients publishes informative
reviews of the AI, that will generate reputations that are much more
informative than those of current therapists (assuming that fewer
reviewers are actively creating noise).
If the system gets good enough feedback about the results of its advice,
it will only take a few years to accumulate better knowledge than the
top 1% of therapists. I imagine an AI therapist who has treated a few
million patients will be able to usefully sort patients into thousands
of categories, and, for the average category, the AI will have tried a
dozen different approaches, and will remember which ones worked best.
The AI doesn't need an IQ of 100 to translate that specialized
knowledge into a large advantage.
Some countries will adopt enough regulations that prevent AI therapy.
Will they be able to prevent patients from learning about and finding
ways to use AI therapists that are launched in other countries? I'm
mildly optimistic, but not willing to bet much.
That might cause an important decrease in prescription drugs for mental
health. It seems premature to short the relevant stocks, but I'm going
to start paying attention to that area (see the new etf SANE).
I expect that in something like 5 years, several AIs will be created
that are able, given enough capital, to make stock markets too efficient
for me to be able to measurably outperform the market. Should I prepare
to retire from this career? The "given enough capital" part won't
happen instantly. There's lots of investment funds that will be slow to
adopt AIs, and there are a moderate number of market inefficiencies that
will remain until a lot of capital is mobilized to eliminate them. So
I'm unlikely to retire as a stock market speculator before 2030.
Can I be one of the leaders in applying AI to the stock market? I plan
to think more about this, but my current guess is that the superior
access to large datasets of big institutions gives them too much of an
advantage for me to compete with, and I'm not comfortable with joining
Two thins the post made me thing of (but not yet very much about).
To play devil's advocate, would you have said the same thing about computers and the internet (improving productivity in a lot of things)? If so, would you expect it to impact GDP? Because it's not clear that it did. https://fred.stlouisfed.org/graph/fredgraph.png?width=880&height=440&id=RTFPNAUSA632NRUG
I don't think one can easily unpack the impact from either computers or internet (which I'm honestly not sure really has significantly increased productivity) impacts on aggregate productivity just by looking and a graph. GDP is nominal prices basically so technology changes that might well increase output or increase output quality while also working to lower or hold prices constant will be masked in simple GDP traces.
I think you'd need to look at some older models, perhaps like Solow's Growth Model, that include a technology term and see how that is moving around. Total productivity seems like it would be driven by labor, capital and technology state. If one assumes human productivity is pretty constant and the installed capital base is likewise pretty set then innovation like computers, internet and AI should show up in the technology component of the model.
Independent of effects on GDP, the internet (nasdaq100) has still strongly outperformed the overall US stock market (sp500).
I would call it a tie:
Coincidentally, I made an app to do exactly these types of historical comparisons of returns, with much greater fidelity.
Input ^NDX and ^GSPC (Nasdaq-100 and S&P500 respectively) as the input tickers. These are Yahoo Finance's codes for those respective indices. Alternatively, you can input QQQ and SPY, which are ETFs that track those indices but there will be less historical data since ETFs come after indices.
The wiki page says the Nasdaq-100 started in 1985? Where are you getting your 1972 data from?
Nice post. I've been thinking along the same lines and made some investments last year. Perhaps a Discord chat group would be better suited to discuss this, I'd join.
TSMC is a giant when it comes to semiconductors that cannot be ignored, but there are risks related to China-Taiwan tensions. I'm invested, but I also have out-of-money puts for protection.
When it comes to robotics, have you looked at ISRG?
There's a Discord channel called EA/LW Investing with about 150 people. I don't have permissions to generate invite links, but if you DM me your Discord username, I can ask someone to you add if interested.
I have not looked at ISRG for quite a while. I occasionally look at smaller startups doing medical robotics. But the FDA slows down innovation in this area enough that I'm unlikely to invest in this area anytime soon.
I’m surprised you don’t expect any effect on real estate. I would expect it to be greatly affected by such a transformation, although I’m not sure in which direction.
If unemployment skyrockets because AIs can do everything that most humans can and cheaper, then fewer people would have to live in cities, potentially leading to crashing house prices in cities. The same could also happen if working from home becomes the norm due to VR.
On the other hand, if human employment stay high, and productivity increases and we still have to be physically at work, then city real estate prices would likely increase with productivity.
Disclaimer: I have almost no track record in active investing, so don’t take anything I say too seriously.
VR will likely shift the locations where housing is demanded. I consider that somewhat separate from the effects of AI. A shift away from the most expensive cities will mildly increase total demand for housing, by making it easier for people to get the kind of housing that they want. I don't know of any publicly traded companies that are focused on housing in expensive cities. Most housing investments are focused on regions where new homes can be built: suburbs and cheap cities that allow new homes.
I like this post a lot, partially because I think it is an underdiscussed area, partially because it expands beyond the obvious semiconductor type companies. One thing I would add, is that with almost no technology advancement, existing and soon to exist LLMs might make investing in social media (and internet adjacent) companies much more volatile. This is because as far as I can see, the bot problem for these companies should only become worse and worse as AI can more perfectly mimic a real user. This could lead to a kind of dead internet scenario where real humans are somewhat drowned out by noise and have grave implications for whatever stocks are involved-who of course depend on ad revenue. The reason I say more volatile is because there is obviously also upside for these same companies to utilise AI themselves further. My views on it are fuzzy though and it could go either way.
AI Therapy isn't the first domino to fall, AI Customer Service is (it's already falling).
95% of customer service humans can be replaced by a combination of Whisper+GPT; they (the humans) are already barely agentic, just following complex scripts. It's likely that the AI customer service will provide a superior experience most of the time (less wait times, better audio quality at a minimum, often more competent and knowledgeable too, plausibly capable of supporting many languages).
Obviously huge cost savings so massive incentive for companies to replace humans (and why it's already started with even weak chatbots).
Investing in it is tricky, same problem you mentioned at the start - picking which horse is going to win this race, most probably either don't exist or aren't publicly tradeable.
Zoom is a potential frontrunner, they acquired Solvvy last year which suggests some strategic awareness of this trend/potential market.
A related thought I’m having: assuming we have a scenario where personal finances still matters, am I better off keeping my real estate loan and my stocks or is it better to sell 3/4 of my stocks to pay off my loan? I’m particularly worried about a scenario where AIs mean that I (and presumably most other people) will be unemployable. My loan is fixed rate.
Stock market feels like one of the last places that modern AI will have a huge impact on (beyond the basics that are already in use). All the breakthroughs seems to be coming in the form of models which are really bad at non-stationarity and extremely high noise. Could be interesting to build a stack of LLMs that ingests 10K's and macro econ data and tries to infer causal models, but I'm skeptical it'll be better than humans at this for a long while. I think you're safe for now :)
I assume you're not a fan of the LRNZ deep learning-focused ETF, since it includes both NVDA and a lot of datacenters (not to mention the terrible 2022 performance). Are there any other ETFs focused on this sort of thing that look better?
I don't look very much at industry-specific ETFs. They're often weighted by market cap, which is generally bad (it tends to weight highly stocks that are in bubbles). Often the larger companies that get included are not at all pure plays on the ETF's theme. For LRNZ's top holdings: SNOW looks like it might be helped by AI, but I don't understand it well enough to bet on that; CRWD looks more likely to be hurt by competition from AI than helped.
And with LRNZ, price/sales ratios are quite high.
There's SOXX, which is a cap-weighted semiconductor ETF and SOXL which is a 2x leveraged version of SOXL.