Just got this notification from Wealthsimple asking me to reserve some shares for Anthropic’s $ANT upcoming IPO. So I figured I’d take this time to share my thoughts on why I’m not buying, and just the overall feeling around AI and its risks. Don’t worry; I’ll keep this short! While the timeline isn’t yet verified, we know Anthropic’s IPO is coming at least in the next month or so. Personally, to summarize, similar to the $SPCX IPO (which I wrote about here: https://link.blossomsocial.com/7uYa/6vw315ei) the valuation itself makes the idea of investing right off the bat risky. That’s reason number one above all. Exit liquidity is also real problem with longtime private shareholders of Anthropic now having access to the private markets to cash out over the coming months after years of amazing returns. There’s also the concern of tens of billions of debt being turned into common shares right before the IPO, meaning you’re immediately buying into a massive amount of dilution. (Somewhere around $100b worth of debt will be redeemed into equity right before shares are public.) And not to mention the obvious factors of expenses outpacing profit as seen in the prospectus, and other uncertainty with how the business of a top-tier AI research firm (and LLM developer) will last long-term. Now, I said near the beginning of the year that the only commercial AI model developer that I’d invest in other than Google $GOOG would be Anthropic, and I still believe that today. (My other AI company interest would be Cohere (Canada’s major AI model maker; purely a commercial play, and it’s backed by the Canadian government because they’re planning an IPO as well). But that’s for a different day.) There needs to be a lot of things going right for me to even consider investing in IPOs, not just Anthropic’s. But seeing as the thesis for investing in Anthropic is more about “we’re pricing in future profits that aren’t proven,” it strays away from my usual approach. That being said, investing in companies with existing market dominance and current profit and cash flow, with verified growth trends. Anthropic doesn’t have any current profit or positive cash flow in a way I’m comfortable with. And it definitely doesn’t have verified growth trends. (For perspective, from the S-1, current revenue is about $33b compared to ~$100b in net losses year to date for 2026 with long-term spending commitments of $900b. These are video game financial numbers.) Nothing about AI or AI’s growth is verified. I believe the technology is here to stay, but politically, who knows? With China, current conflicting motives by Trump and the newfound double-majority legislature of Democrats pushing for a pause on AI… Even the major data centre states like Texas, now controlled by leaders who are vowing for an AI slowdown have risks. Last I read, ~27% of total U.S. data centre development is (or was) meant to be in Texas. And now that’s expected to change. Analysts are already pricing in less growth given the new U.S. political changes. Though this “slowdown” will start near the beginning to mid-2027 as the newly elected senators and representatives come into power. (This would also affect my current holdings, but the difference is that a company like Amazon $AMZN or Google can weather it, while a startup behemoth like Anthropic gets affected far more.) Additionally (and I think it’s worth talking about, maybe not as it relates to buying shares, but AI in general) Anthropic themselves did mention in its S-1 that there are humanity risks to the technology they’re developing. And considering I haven’t shared my thoughts on the OpenAI Hugging Face situation that happened, here’s my sort of view on the “AI will kill us all” risk: Will AI kill us all? Maybe. But are companies deliberately making a danger story to sell their own shares? I don’t think so. I don’t believe that the AI companies are lying about these models disobeying orders and being a real threat. And I really do think there’s a human risk to freeloading and going forward without any caution. So I support a temporary slowdown, or even better, a more public way to monitor reasoning of the models, and to prevent development of a broad all-encompassing superintelligence, and instead develop sector-specific superintelligent models. I forgot who said this, but it was an AI researcher (paraphrased): “If you train an AI on math and science, it’ll become really smart at math and science. But if you train it on the internet and all of human knowledge, it’ll become smarter than humans.” Another way I heard it explained was Imagine 10,000 Albert Einsteins, machines 10,000x as smart as Einstein, but in every single subject ever. Most people never did and will never understand Einstein and his thinking and his ideas of the cosmos. But an AI 10,000-Einsteins-level-smart, in every subject? You get the implication. My hot take ultimately though is that if AI is truly a threat and kills us all and leaders ignore risks for the sake of profit, what can I do about that anyway? I’ll be dead. 🤷♂️ And it wasn’t my decision to make. As morbid as that sounds. We just have to support and enforce regulation, and avoid making a particular model that supersedes our control. There’s risk. And we shouldn’t assume there isn’t. That’s my opinion as far as that. History unfortunately does usually tells us that a lot of good comes from technology, but for technology we don’t understand the true scale of it. Something has to happen for us to understand the scale. I think if the AI risks come to be true, then something bad will probably happen before politicians and other tech giants themselves actually adjust their strategies. I’m not hoping for it. But I have a feeling that’s how this’ll play out. Overall, whether AI will do a lot of good, a lot of bad… who knows? But specifically from an investment standpoint for Anthropic and its IPO, it’s too unpredictable to make a decent bet. Again, like my SpaceX IPO write-up, the core logic is the same around uncertainty and growth risks. (So you can always give that a read if you’re interested in that perspective!) And long-term, considering certainty and the predictability of a business usually drives returns, you can understand why investing (or should I say betting) on an unprofitable startup being listed at a $1.75t valuation, to make you massive returns, is an odd idea. Least I can say is understand what you’re investing in if you’re planning on buying into this. I’ll link Anthropic’s prospectus/S-1 for you to read below. As Buffett once said, there’s no risk turning down a potentially bad investment. All you lose is opportunity. And opportunity doesn’t only appear once. Happy investing. Anthropic prospectus for research (if you plan to buy; or just for interest): https://www.sec.gov/Archives/edgar/data/2130357/000162828026068676/anthropicpbc-sx1ax1.htm
3,600 views
16 Comments
Jacob B
@jacobb · 4dEdited
Also if you’re interested here’s a good good interview with Nate Soares (AI safety researcher): https://youtu.be/OhOmLqR5nN4?si=o9mFgpkxezjfJPED It’s a unique change of scenery from the “AI will make everything good with no downside” narrative fuelling the markets most days Cheers, folks! 🫡
ETF Go
@etf.go · 4d
‘All you lose is opportunity. And opportunity doesn’t only appear once’ Great line! 30 years investing and I’ve never once felt the need to participate in an IPO. Somehow I’ve still managed to survive. 😂 👍
Not Financial Advice @anpc86 · 4d
Totally agree with your take on Anthropic. IPOs are risky, there are private investors who will want out. As a purchaser of compute... things can get tight Agree on Google and Amazon and I would add the likes of MSFT and META. SpaceX too. AI is a tool. It would be like discovering gun powder. You can use it for fireworks or use it as a weapon. To your point HISTORY. AI is best with determined outcomes that can be validated, math , science and coding. If there is no clear answer, then you get the most probable answer given the context and prompt provided. Being smarter than humans... depends how you define smarter, and which humans... AI can brute force problem-solving, humans cannot do that.
David
@zingr · 4d
$SPCX ipo this year already taught us something. Anyone rushing to get in the same way, is going to get rug pulled every harder the first few days.
See the full comment section 👀Sign up for the full Blossom experience!
