Amazon plunge continues $1T wipeout as AI bubble fears ignite sell-off | Hacker News Reader

Amazon plunge continues $1T wipeout as AI bubble fears ignite sell-off

cnbc.com

92 pointstruegoric90 comments

Question: does Amazon's retail business matter for analysts at all? It drives the majority of revenue for the company but a much, much smaller amount of profit. Does Wall Street care about Amazon's core business one iota?

It definitely boosts their overall value as a company. If one share equals a slice of "what the company is worth now" + future growth, steady long-run revenue sets a solid baseline for the stock price.

Big revenue + small margins in a stable business, IMO, is a massive liability for the bottom line; any downturn in business and that becomes big revenue + big losses. Even if cloud is making money, it can wipe a lot of that out.

From the point of view of running an enterprise that lasts, though, diversification is important. Financially diversification is probably, in general, bad for EPS. But if you want to run a lasting empire, it's best to not tie it to just a narrow thing.

That depends on the business. People are not going to stop eating so small margins in the grocery business isn't a negative - the revenue is mostly recurring and recession proof (some people might switch from buying meat to rice+beans, but other people are going to stop eating out and so it balances).

Just because people need a grocery store doesn't mean that you're guaranteed to make money running one.

multiplying huge revenue by a small percentage to get a big positive number

to multiplying huge revenue by a small negative percentage to get a big negative number

So that's how Kroger managed to lose billions over the last couple of quarters, or how small changes in shoplifting/shrinkage based on store makeup can cause losses to some chains, etc.

https://massmarketretailers.com/kroger-delivers-solid-quarte...

They didn't lose money because of their grocery operations. Those margins have increased slightly.

They lost money because of their grocery delivery operations.

Their margins have increased slightly if you ignore the part where their efforts at grocery fulfillment blew up.

Just because an industry is essential doesn't mean that every firm in it makes money.

I agree, all kinds of grocery stores have failed over the years. Kroger's just isn't a good example of a failing store for a lot of reasons. They aren't a stellar investment, but they are also up from 1 year ago and 5 years, so investor's don't look like they agree with your summary of the business.

With tech companies, investors buy future growth, not stable businesses.

The margins on tech/cloud are just so astronomically higher than retail. Places like Walmart or Costco are fighting for <5% profit margins, IIRC its closer to 2%. Quick search says AWS has about a 35% profit margin and about 50-60% of the operating profit of amazon but 17% of the Revenue. With those numbers it makes sense to focus on AWS.

IMO AWS should be spun out as a separate company.

It doesn’t help that now other potential growth markets for AWS, like the EU, now are getting pushed to have more and more data sovereignty due to the administrations antipathy towards allies. Musk can whine all he wants about the EU, but that’s like complaining about customers that don’t want to buy your products instead of building a good product they trust and want to buy.

This makes me wonder if Amazon the retailer requires having access to AWS services “at cost” in order to be profitable. If AWS was spun out completely, would Amazon proper be able to afford their AWS bill of AWS had a profit margin on it.

I’m sure Amazon.com would be fine, but it would take a chunk out of their margins. I’m also sure that their X% ownership of the spun out AWS would cover the difference.

Consider the reverse. If AWS didn't have a large guaranteed customer in the form of Amazon, would they still be able to develop their products with perfect knowledge of the needs?

If AWS becomes a separate business now, they may be able to build better products, given a bit less focus on one large customer (Idk if Amazon.com has any priority in their product roadmaps now)

I don't have the answer but this question has an answer, as Amazon's earning reports are public.

If they were separate businesses it would make no sense to merge them. It would be like Microsoft buying Walmart.

Don't give them ideas.

Growth is all that matters. There is perceived to be much less potential growth in retail than there is in tech. You have to remember, most people literally think of computers in magical terms, and what's possible is usually more anchored by what they see in movies than what they experience in real life. So believing that Sam Altman is going to manage to capture all economic output of labor is seen as a realistic belief. Believing that Amazon will replace all retail in the world is obviously never going to happen.

I'd run the numbers and AWS and Amazon's ad business together accounted for 100% of Amazon's market cap. The e-commerce division is essentially worthless as far as Wall Street is concerned, other than a loss-leader inventory source for the ad business.

This insanity was going to break at some point. Now hopefully these trillions of dollars of losses might finally allow the price of old DDR4 memory I've been trying to acquire to finally recover.

They always say it's about "AI," but it never turns out to be about AI. I wonder what's it about this time?

The collapse of the yen carry trade.

I have read that numerous places and it seems plausible but it is beyond my investing experience.

I think the new nominated Fed Chair is also a hard money advocate and is spooking USD alternatives (gold, silver, BTC, etc.) But hard money can be quite hard on the economy, so that could limit growth.

We should just blame "the willies".

Investment is all about belief. When the root cause is the willies, then we hallucinate reasons together.

Crypto and memes have demonstrated us a lot about the drives of individual investors.

Unfortunately it seems that professional investors are not that much more rational (from my limited personal experience with a small hedge fund, and from my years of looking at markets).

My favourite term has always been "taking profits" which is generated by the technical analysis (I loath that term) of looking at the prices: taking an effect and publishing a cause (trying to sound smart).

We are deeply irrational beings; often the more you go up a professional ladder the more rationalisation you see.

During unstable periods, we see lots of weird side-effects and there is a lot that doesn't seem to make sense.

Edit: a better meme could be "The use of AI by funds is destabilising markets"

Disclaimer: I am not a professional investor. I am a cynic.

Speculation involves belief, sure. However, there is hard math involved in markets as well, and the yen carry trade equation exists whether you believe in it or not.

I love your juxtapositioning!

Can the yen carry trade equation exist outside of belief?

Edit: I just noticed your comment is the one result for your "equation"!! https://duckduckgo.com/?q=%22yen+carry+trade+equation%22

Not saying you're wrong, necessarily, but as I type this the Dow is at 49,700. Would you expect the collapse of the yen carry trade to cause the Dow to collapse as well? Or is the Dow not high-growth enough for people to put yen-carry money into it?

> Would you expect the collapse of the yen carry trade to cause the Dow to collapse as well?

USD devalued ~10% last year, so some of the losses are already priced into the DJI. When you account for that and sprinkle in ~3% inflation, it has lost value despite being up ~11% in the last year.

Anecdotally, read somewhere that delivery people are going idle. The Orange One wrecked everything with tariffs, and the ripple of destruction is slowly taking its course. That's before we take into account massive outflows of cheap labor and fired government workers.

wall street analysts are starting to realise that software companies shouldnt trade on a P/E of 300.

DocuSign is currently valued at 30 times its annual earnings. Adobe is currently 16. Amazon is 28 -- has been as high as 50 recently. NVDA is 44.

Investors are basically starting to realise that enterprise are not going to subscribe to software like DocuSign for 50 years. They'll probably just move to odoo or zohosign or something and save a lot of money. So its probably a better bet to put that capital into something like Nvidia or Tesla or whatever. it also looks like the US Fed isn't going to cut rates, so capital is getting more expensive.

Of course, if you are a CEO its great to blame all this on AI and then tell your investors you are increasing AI in your business (see: salesforce whose stock price is down 42% in a year and is now trading at 25x earnings)

> software companies shouldnt trade on a P/E of 300

You are playing pretty fast and loose with your definition of a "software company" when you include Amazon and NVIDIA in your list. Amazon is many things but it is not a "software company" and neither is "NVIDIA".

50% of amazon operating profit is from AWS. NVIDIA's GPUs aren't really that much better than AMD if it weren't for CUDA.

Software company is a pretty good description for both.

I’ve heard the same about Nvidia, quite a few times, but have never really understood it.

I don’t suppose you know a good “for dummies” explanation of why CUDA is such an insurmountable moat for them?

Like, what is it about that software that AMD can’t produce for their own hardware, or for a most important subset, with these $1T market stakes?

My understanding on this may be spotty (and I appreciate it if someone corrects me), but CUDA is not the software layer that allows you to use NVIDIA GPUs for AI processing?

AMD may develop their own software layer, but a lot of things already work on CUDA, and the job to port this to a different platform may be non-trivial (or even possible depending on the level of feature parity).

> I don’t suppose you know a good “for dummies” explanation of why CUDA is such an insurmountable moat for them?

Theoretically the moat isn’t insurmountable and AMD has made some inroads thanks to the open source community but in practice a generic CUDA layer requires a ton of R&D that AMD hasn’t been able to afford since the ATI acquisition. It’s been fighting for its existence for most of that time and just never had the money to invest in catching up to NVIDIA beyond the hardware. Even something as seemingly simple as porting the BLAS library to CUDA is a significant undertaking that has to validate numerical codes while dealing with floating point subtleties. The CPU versions of these libraries are so foundational and hard to get right that they’re still written in FORTRAN and haven’t changed much in decades. Everything built on top of those libraries then requires having customers who can help you test and profile real code in use. When people say that software isn’t a moat they’re talking about basic CRUD over a business domain where all it takes is a competent developer and someone with experience in the industry to replicate. CUDA is about as far from that as you can get in software without stepping on Mentor Graphics’ or Dassault’s toes.

There’s a second factor which is that hardware companies tend to have horrible software cultures, especially when silicon is the center of gravity. The hardware guys in leadership discount the value of software and that philosophy works itself down the hierarchy. In this respect NVIDIA is very much an outlier and it shows in CUDA. Their moat isn’t just the software but the organization that allowed it to flourish in a hardware company, which predates their success in AI (NVIDIA has worked with game developers for decades to optimize individual games).

Maybe nobody has reputably released non-fortran versions but they probably exist.

Lots of other versions exist including reputable ones like Intel’s MKL. The hard part isn’t reimplementing it, it’s validating the output across a massive corpus of scientific work.

BLAS is an example though, it’s the tip of an iceberg.

the first problem is a whole generation of people learned to code ai applications by fiddling around with the gpu in their gaming pc 10 years ago. so an entire generation of talent grew up with cuda

the second problem is that so many libraries and existing software is cuda only. even some obscure hardware stuff. i discovered the hard way that some AMD thinkpads dont support thunderbolt transfer speeds on their usb-c ports, whereas nvidia ones do

the third problem is that the cost to develop a cuda equivalent is so great that its cheaper for companies like google to make TPU and amazon to make Trainium. its literally cheaper to make an entire new chipset than it is to fix AMd. i dont see companies like apple/amzn/goog etc fixing AMDs chips

>its literally cheaper to make an entire new chipset than it is to fix AMd

Is it? Or does AMD expect to make a profit and it's cheaper to make your own chips at cost?

i mean its cheaper from an enterprise customer perspective. if a company is training an LLM, writing their training programs to use AMDs hardware instead of just using CUDA is so expensive and time consuming that it is cheaper to pay four times the price and use nvidia hardware. in this space its important to move fast, although that economic will shift over time

which is why nvidia hardware trades at a 4x premium to AMD

its not necessarily cheaper to make chips at cost either. nobody is making them, only designing them. so first you have to design your new chip, then you have to get a minimum order in with the chip fab so big it competes on unit economics, and then finally you have to get your dev team to write a CUDA equivalent software that is a problem so hard its only really been solved by apple, google, intel, and nvidia

only companies with big fab orders can get priority too.. if a company did all of the above and was ready to go, they probably wouldn't get fab capacity until 2030

CUDA is a GPGPU software layer that is very mature, and integrates with C,C++, Python, Fortran very well. AMD just never really got the same quality of GPGPU software in the last 20 years. 99% of scientific computing that uses GPUs (which is a lot since they are so much faster than CPUs for linear algebra) have gone to Nvidia because of this. All of the big AI libraries (Tensor Flow, PyTorch) basically ended up writing around CUDA, so they just didn't write things for AMD. Plus if you go and look at a job for signal processing or whatever at say Lockheed Martin or Raytheon, they specific CUDA.

AWS is not a software company either, it is a capital intensive computing infrastructure company. NVIDIA as well is a capital intensive computer hardware company.

Just because software plays a role, doesn't make them software companies. It is like saying all companies are "electrical companies" because they require electricity to operate.

Is a seal a sea creature or a land dweller? Well it eats at sea.

Argh, we first ignored the fact that most of its income comes from being the lead e-commerce retailer to just focus on AWS and then we need to discount the fact that the majority of AWS is CapEx towards hardware / datacenter (expected CapEx this coming year is $200B), to just leave the software. Whatever.

One thing is for sure: if the seal builds and manages hundreds of billions of dollars worth of computer infrastructure across seven continents, it isn’t just a software company.

youre right, i should say tech company. but at least my flawed epistemology reveals my humanity

although one could argue disingenuously that nvda is a software company because the product they ultimately manufacture is a bunch of blueprints they email to tsmc or samsung who then actually make the chips

> but at least my flawed epistemology reveals my humanity

Plenty of AIs have flawed epistemology. But nice try.

I actually don’t even think it’s a IaaS company. I mean it is, but Amazon seems to always excel at operations: whether it’s organizing a retail business, operating a logistics company, or managing a hyperscaler, it just seems like its real secret sauce is running ops.

(Which makes sense because all of their end user products suck)

I've always found it confusing how run of the mill SaaS trades at multiples assuming decades of doing business. The amount of change in software businesses has been massive and being able to run a successful software business even for 15 years from 2010-2025 requires a great deal of strategy and foresight and more likely than not that's not enough. Considering how these dynamics have been accelerating as technology accelerates it just seemed so off that the market was landing on a 20-30x multiples for software businesses that don't have much moat (e.g. swathes of B2B CRUD apps).

Investor analyst looks at earnings growth and determines Customer Acquisition Cost (CAC) and Customer Acquisition Cost Payback Period (CACPP). They determine that ABC Software Corporation has no marginal manufacturing cost because it makes software that it sells online, so if it invested 90% of its profit margin into marketing it could grow its ARR by 140% a year. Then they extrapolate that for 30 years and say ok the NPV of 30 years of 140% ARR on current CAC, etc etc...

If everyone in the industry benchmarks on more or less the same multiples, it becomes a good idea to buy any b2b crud saas trading at 10x earnings because if the big boys see it they'll probably bid it up to 30x

the other classic move is to take a business which really isn't even a new technology, like revolut, and call it a tech business. now suddenly a bank can trade on a 50x earnings multiple instead of 15x like say a bank. many such cases~

Hes not admitting his crimes, hes bragging!

Why would you put more money into Nvidia or Tesla right now? Don't you think they are priced in already?

we are only examining the valuation metrics here, not comparing the companies themselves as investments.

you could decide that if you are a very large company, building software internally to replace a SaaS product is a path forward. Or replacing a premium software like DocuSign with a cheap one like Zoho sign. or just building your own proprietary electronic signature app

It is however impractical for big company to start manufacturing cars or designing competitive GPUs

so the earnings of tesla and nvidia is theoretically more 'stable' than a software application company like salesforce, adobe, etc.

this analysis ignores both the size of the company, and what it does, or whether or not any one of them is a good investment

Every tech company assumed they would be the benefactors, not victims, of AI. And investors now see that without the alleged AI growth, these companies at best look like stable utilities, not high growth stocks. At worse companies look like they make highly replaceable software as software stops being a moat.

Moreover they look like large, inefficient organizations with a lot of human veto points that prevent innovation (requiring more human coordination is an anti moat now)

Was software ever a moat? Software typically only gave companies a small window of opportunity to turn a fleeting software advantage into a more resilient moat (network effects, switching costs etc.)

Yes, I would argue good (stable, fast, easy to use) software was somewhat of a moat and much harder before coding agents.

Stripe, Square, Shopify, Google, all thrived in some part because their services take a hard problem and make it easier to use. Now more people can take a hard problem and make it easier to use.

All you have to do is look around (esp 5+ years ago) and see the many many BAD, unstable, hard to use, slow, etc versions of these companies

Windows was a moat but it looks more like an anchor now.

Windows' moat was not the operating system code, but that they were able to get distribution via IBM, and then grow an ecosystem of applications that were targeted at Windows, which created a snowball effect for further applications.

Yes though still it was a big barrier to build an OS.

Hmmm. And I wonder who bought all that NVDA today and what's suddenly so compelling about the price.

Certainly couldn't be someone in government trying to pin the NDX100 index.

https://youtube.com/watch?v=LZ-bkTG30Ns

It’s not an AI bubble - it’s an inflated P/E bubble.

Yeah, the stock price is still higher than any price it obtained prior to Oct 2024. I think people are just shocked that stock prices may go up and down rather than up only.

Amazon missed earnings and promptly doubled down on AI spending:

https://finance.yahoo.com/news/amazon-plans-200b-ai-spending...

It is encouraging to see that investors are punishing what is the greatest misallocation of capital since the dotcom bubble. Investors have figured out that AI is limited to probabilistic and annoying chatbots that are for entertainment and for looking up trivia questions.

I know it feels comforting to say this, but deep down you have to realize that saying things confidently does not cause them to become true.

Was that comforting? At least the commentator came with a source

Theres a real problem with people who are too tech-induced: youre disconnected from how the average person interacts with stuff.

If only that is what investors have figured out.

Unfortunately, it seems investors now think that all paid software will be replaced by AI generated software, somehow open source projects laundered through generative AI models should finally convince enterprise customers to go with free.

Why should anyone take your sensible first statement seriously if your second statement is so easily verifiably false?

Some of these AI critical posts really are an exercise in Gell Mann Amnesia, man.

Im still waiting to read about macro-level mass-lay offs or insane productivity leaps.

Where are the results, tell me? What insanely great products have been shipped by people leveraging/building on top of LLMs...?

Yeah, silence. As usual.

There's literally a billion ChatGPT users already, the worlds fastest growing product. Do you think they're all just playing around in the sand? Ask anyone in education, it has completely upended every student's workflow.

Just because people aren’t talking about how LLMs are boosting their productivity to ship “insanely great” products doesn’t mean it’s not happening. I don’t talk about every tool I use, and neither do many of the best teams who keep their eyes on the prize and relentlessly ship.

Meanwhile this story just came out: https://newsletter.semianalysis.com/p/claude-code-is-the-inf...

You can find posts on this site every day with people say they're seeing big productivity gains.

I'm sure they'll all lying though, right?

Or, because it hasn't solved all of the world's problems yet then I guess it doesn't count?

Maybe you're the one overhyping AI.

I think their point was: at this point we should see clear macro effects. There should be new software that wasn't possible yesterday (or feasible, etc). Something big and new, and high quality.

AI helps me too, and I feel like I ship more now. But what if it's just a feeling and I'm not doing anything groundbreaking? What if I'm just doing what I've always done but at slighly more speed?

> There should be new software that wasn't possible yesterday

This is exactly the overhyping the critics are continuing to strawman their complaints with.

Speaking for myself, I think AI is better than the average dev. And yes I think it is affecting hiring.

But this is very different from saying I think it will do new things that weren't possible yesterday.

Oh boy, are you going to be in for a rude awakening. Might I ask what is your exposure? Because this does not line up with what I am witnessing day to day at all.

This type of commentary reminds me of the people during the dot com boom who were adamant that e-commerce was all film flam and would never take off.

Consider that it is possible that both (1) we are in an investment bubble and (2) we are underestimating the long term impact of LLMs and perhaps mispredicting where they will land.

By your logic the only way to achieve number 2 is to keep pumping number 1.

This type of commentary reminds of people propping up these LLM MLMs.

In what way is the long term impact of LLMs being underestimated? If anything, it seems that it has been overestimated in the past years and that something other than LLMs will be needed to reach the original scaled LLM hope of AGI.

Back when the Internet was America online and some CGI bin perl scripts, there were a lot of very lofty things said about the potential of the Internet in the future. I don’t remember any of them predicting the power of the tech would have over business, politics, media, and hours of every single day for billions of people. Even without AGI, it’s quite possible that were still underestimating. The effects of predictive, probabilistic computing 20 or 50 years from now.

The internet alone didnt change sh!t. Without smartphones, unified app stores, cellular network innovation et al internet traffic would not be so high.

Funny how people leave this stuff out. Yawn. Basic simpleton analysis and takes.

The Internet created the backbone that allowed for rapid experimentation in communications technologies, and created the ability for anyone to create and share technologies and reach a huge audience very quickly.

Without the Internet, most consumer electronics would have been far more expensive to build, and would have been strictly controlled walled gardens, but the Internet in general and the Web in particular allowed so many inventors to flourish. Ever since that Genie was let out of the bottle, corporate and government interests have been trying to put it back in, and most companies are trying to build and reinforce walled gardens under the banner of unified app stores that extract insane rents.

Wow this is like going on a medical forum and saying "medicine didn't change shit".

They were replying to this particular underestimation:

> AI is limited to probabilistic and annoying chatbots that are for entertainment and for looking up trivia questions.

That is not a rational assessment of the utility that the technology provides, even today.

You’ll get downvoted for your second statement. I think investors are struggling to see how AI turns into more money for consumers if it. It’s one thing to exclaim how your productivity is up, but does that translate into more profit and larger customer base if you’re a business? I very much doubt consumers will pay more than dollars a month for an LLM and I also very much doubt the ad market can grow large enough to cover the spend on that (ad market is plenty big and driven by other economic factors)

Many people are spending significantly more time every day engaging with AI chatbots than they spend engaging with Google, and Google is one of the most valuable companies in the world.

By that exact logic Tiktok should be the most valuable company on the world, yet its not even in top 10. Attention doesn't automatically generate profit especially since most of these companies are yet to monetize attention. They are still burning billions and the public opinion on this outside of tech bubble is very negative. It wouldn't surprise me if the money on the internet falls to pre-2020 levels in next few years. Subscription pricing model is also becoming increasingly cumbersome for companies and individuals which is another worrying trend.

I suppose people don’t realize how much of their day is actually engaging with Google through their trackers, their email, their phones, YouTube, etc.

Yeah.... stick to conversations strictly re. technology pal.

I was disagreeing with just your last statement, but then I did a little research: about 80% of revenue is from chatbots and 20% from APIs. So +1 on your comment.

Bear with me here, I actually do have a point to make: I took my stepdaughter out for breakfast this morning. She is a financial wizard specializing in running large cities, and to explain to her the current craziness of overspending on AI infrastructure, I described "exponential spending increases for linear economic value increases." I may be wrong about this, but I am all for targeting the sweet spot of more efficient smaller AI models that are fit to purpose for specific use cases.

You're agreeing with GP but it's a flawed premise: a chatbot doesn't have to be limited to entertainment and trivia, as anyone who has used one for productivity knows.

Reply on news.ycombinator.com