Over the past few months, I’ve been getting more interested in both AI and the stock market. Naturally, I started thinking about how to benefit from that financially. But the more I think it through, the less convinced I am that it actually makes sense.
I keep coming back to a few different scenarios, and strangely they all seem to point to the same outcome: disappointing stock performance for LLM companies like OpenAI, Anthropic, and similar players.
Scenario 0: The pessimistic view
Maybe the skeptics are right, and intelligence doesn’t scale the way people expect. We run into hard limits, progress slows down, and the whole space cools off.
Result: the hype fades, IPOs underperform, and valuations slowly decline.
Scenario 1: Everything works, maybe too well
Let’s say the opposite happens. Companies like Anthropic or OpenAI make major breakthroughs and get close to something like superintelligence. Productivity increases massively across industries.
Result: goods and services become much cheaper as intelligence gets widely accessible.
The problem is that if everything becomes cheap, margins shrink. In an extreme version, even the way we think about pricing and value could change. So despite huge success, stock returns might still be underwhelming.
Scenario 2: Strong progress, but competition eats the upside
AI improves a lot, but in a more gradual way. Companies still compete, protect pricing, and try to behave like normal businesses.
Result: competitors, especially open source and lower cost players, keep catching up and undercutting prices. This is already starting to happen coming from mainly China but it could be anyone geoplitically wise or domestic. Margins get compressed, and strong tech doesn’t translate into strong profits.
No matter how I look at it, I end up in a similar place. Either the tech disappoints, succeeds too much, or succeeds in a way that still doesn’t generate great returns for investors.
I’ve tried to challenge this thinking, including asking LLMs, but most answers felt overly optimistic or didn’t really address the core issue.
So I’m assuming I’m missing something obvious.
What’s the real bull case here? Why would someone still invest in upcoming LLM IPOs given all this?
Does investing in upcoming LLM Stocks even make sense longterm?
byu/fennforrestssearch instocks
Posted by fennforrestssearch
15 Comments
I don’t think they can all survive. Just not enough differentiation, and my very amateur opinion is Anthropic seems to be pulling away with Claude and at least has a CEO who understands the concerns behind what he’s building.
Power, its power and it has always be power
No power, no AI
No power, no internet
No power, no onlyfans
No power, no crypto rubbish
etc…
This is why many people choose to buy the “picks and shovels” instead. Buy the A.I infrastructure stocks instead or invest in QQQ so that you have some exposure to the winners regardless of who they end up being.
Bubble
I wouldn’t touch these IPOs personally, they’re dumping them on the market at insane valuations at the top of the most ridiculous hype cycle the world has ever seen.
Ai is real, and it is at the beginning of what it will be, but that doesn’t mean it will just go straight up. If you own Voo, you already have exposure to ai risk and once these companies go public they’ll be so large they’ll be added to indexes pretty quickly.
In my opinion voo is already concentrated with ai risk, I personally don’t want any additional risk than I already have from a direct investment in a company with this much influence over the market and also these financials.
As others have said, without power we can’t plug the gpus in that will power the next wave. If anything I’m betting on small scale nuclear but that is also pretty risky at the moment.
All the mag7 ran up from a major drop the last 2 weeks. I fully expect that rotation to move back into saas, which is being treated like shit. Theyre not going away. People dont recognize the cost to run a agentic ai to do work, more or less to replace enterprise software
The best risk reward you’ll get right now is stuff like ASML, AMAT, KLA, AVGO, MRVL, and to a lesser extent MU. Couple weeks ago would have been much better valuations for all of those but the way things are going, you might get another shot in a couple more. Then in the lower risk level, like others are saying, power and related stuff. ETN, NEE, VRT, VST, and when it calms down a bit, CAT.
Bro how would you even invest in llm stocks, they’re all basically private and doesn’t exist on the open market unless you’re an accredited investor.
Also, your scenarios are all very all or nothing, it can be a mix of multiple, and closed source becomes premium pricing with strong margins while open source has a life of their own – sometimes, high intelligence is worth 10x the price. And claude is distinct in behaviour from chat which is distinct from gemini. They’re building lock in and ecosystems of their own, much like google ecosystem is distinct from microsoft ecosystem.
Just keep buying SNDK and win in the AI game. AI infrastructure is going to be king for the next few years. The big hyperscalers will not be that successful for a while as they keep investing in AI capex where infrastructure companies like SanDisk benefit.
IPO will be exit liquidity for the pre-IPO investors.
The only top lab I’d invest in is Google. Otherwise semi, datacenters, etc. Infra.
Presently the main strategy in the private markets seems to be the thesis that: “One or some of these companies will win and take over market share while most will fail, therefore it’s worth allocating smaller amounts to many companies instead of large amounts to some.”
If you’re willing to make an active bet on any product and company then that’s no different than what stock investors have done for all of market history. Of course you could pick wrong and then run the risk of a very steep loss. The real problem is just that nobody knows who will take over market share. These are very fun markets for active stock pickers but very easy to get wrong, so that will probably be a personal matter and not something that can be answered absolutely. If you do genuinely have some unique insight then you might make a great return off of it, but there’s real risk and that uncertainty is fundamental to why the market is the way it is right now.
Makes sense if you enjoy being exit liquidity.
I am also pessimistic. Until companies can build a better moat to prevent distilling, i don’t see these big players being worth trillions.
I don’t really trust any of the new IPOs. The direct AI money I have in the market is in Google. I’m not sure I believe in LLMs as the long term future of AI, but the concept of AI I think is coming. So I put my money in the company that has its fingers in every AI pie. They’ll get a piece of whatever hits.
No