Still don’t really understand why deepseek’s launch wasn’t a more long-term market correction. It feels like investors don’t want to miss out on a second Amazon, where Bezos spent years going all in on R&D to build the infrastructure for what it is today.
Because they invested too much to admit they screwed up. Also, Amazon just sold at a loss until their competitors couldn't afford to keep up making no money. That's not infrastructure or innovation. It's literally the opposite.
Not sure why your comment got flagged as rude, it’s not, but Amazon’s early years were marked with exponential rev. growth, but no profitability because it was reinvesting its cash flow into capex and R&D. I’m not a fan of Amazon/Bezos, but early investors that stayed long were rewarded handsomely.
Their current economic position isn't a result of developing new technology that paid off the investment. It's just monopolistic behavior that was enabled by operating at a loss, funded by that investment. When you don't have to make a profit and your competitors do, you can just wait them out.
I think the hangup here may be the term "R&D". Amazon did a lot of research & development, but it wasn't in new technologies. It was in integrating existing technologies. It wasn't stuff they could sell, except in the sense of, e.g., running Barnes & Noble's online store.
He innovated owning so much that you just charged everyone feudal rents instead of producing anything.
Yep, appreciate the clarification. The investment was in acquiring massive amounts of real estate and building a distribution and server facility network.
I don’t follow your 1st paragraph. In the early stages of Amazon, they grew their revenue year over year, but still reported losses because they were building out the distribution and facility network that we see today. The majority of investors long on Amazon saw it for what others thought it was:
An online bookstore. They didn’t envision an online mktplace for everything (& a massive cloud hosting service). That said, more recently they clearly use their market position to engage in anti-competitive behavior, which does stifle competition. I’m just saying they didn’t start that way.
It’s because tech has no backup plan for when AI goes bust, and the US economy has no backup plan for when tech goes bust. Everyone’s just inflating the bubble and hoping to cash out before the mess gets on them.
Yup. There is no next horizon that wouldn't be reliant on scientific advancements that the techbros have no idea how to make. They've run out of existing technology to commercialize.
Same as it ever was, same as it ever was Same as it ever was, same as it ever was Same as it ever was, same as it ever was
Don’t disagree that it’s a bubble, but do disagree that AI will go bust (despite my concerns about its impact on humanity). One of these AI companies will achieve market dominance, and I’d have to imagine that investors of companies that don’t gain that perch will be left with pretty huge losses.
0 AI companies have reported a profit, ever. We are 3 years in. Reports that R&D will innovate are proven false. The deck was already stacked in their favor in terms of politics. What path is there to profitability?
That’s why I made a comp to Amazon, which didn’t make a profit for 6 years after it went public. AFAIK, OpenAI’s financial statements aren’t public, but I’d love to see their SCFs. I assume the most sig. costs are related to server usage, which is why I referenced deepseek as a course correct.
Yeah neither of those comparisons apply to American AI. Its already bought too much hardware to replicate the deepseek model (that’s setting aside the massive paradigm shift it requires). And Amazon had the opposite issue — their product was fundamentally profitable but hadn’t grown to scale.
“internet supermarket” was never in doubt as a valuable thing to win the race to, Amazon just needed time to leg it over there. AI on the other hand is not profitable on the basis of anything concrete but just the assumption tech wizards will one day make something profitable.
I think your point about being too far gone into the business model is right. But directors and investors should be asking the question, regardless. Point taken on Amazon, I was only pointing it out as a comp. whose business model called for massive capital investment for their larger aims.
Fair; my point is where the comparison breaks is Amazon had a path to profitability — build it to scale basically. AI has already been shoved in every corner of the internet and still can’t make money. It’s been stringing investors along on “R&D will release magic soon” but signs point to no.
That’s a good point. Even as an online bookseller, you can argue Amazon would have been profitable. I think what they’ve become, outside of Bezos, his management team, and probably a select group of private equity investors, was completely outside the scope of what we could imagine.
The cost structure required (or at least that were told is required) of AI doesn’t seem sustainable at all to me. But it certainly seems to me like something the tech industry at large and its investor base are full systems go on.
Investors are in on it because from the mid 20th century until ~2020 the law held that hardware improvements meant whatever tech was up to would sort itself out eventually; this created a perception that it’s a land of wizards who will always innovate. People don’t realize the party is over.
But how can they monetize? I think that's the big issue, all these projects seem to be burning cash at alarming rates, and no one is making money except the ones selling the shovels.
Even the shovel-sellers aren't making money this time! Saw a report on CoreWeave just a couple weeks ago; they're juuust about out of runway.
I would assume there would be a similar model in the nearish-term to a search engine, and over the longer term the revenue streams would become even more nefarious, unfortunately. You’re absolutely right that the costs are huge for U.S. companies. That’s why Deepseek seemed like a game changer.