Amazon strategy is to invest in the infrastructure, money is where the machines live. I think they just realized none of those companies have a moat, so why would they? But all of them will buy compute

And electricity.

I would be surprised if a cloud market leader thinks winning on commodity vm rental is a strategy

Except they’re struggling here. The performance of their offerings is consistently behind competitors, particularly given their ongoing networking challenges, and they’re consistently undercut on pricing.

For Amazon “renting servers” at very high margin is their cash cow. For many competitors it’s more of a side business or something they’re willing to just take far lower margin on. Amazon needs to keep the markup high. Take away the AWS cash stream and the whole of Amazon’s financials start to look ugly. That’s likely driving the current panic with its leadership.

Culturally Amazon does really well when it’s an early mover leader in a space. It really struggles, and its leadership can’t navigate, when it’s behind in a sector as is playing out here.

Under what scenario does Amazon lose the beast that is its high margin cloud service renting? It appears to be under approximately zero threat.

Companies are not going to stop needing databases and the 307 other things AWS provides, no matter how good LLMs get.

Cheaper competitors have been trying to undercut AWS since the early days of its public availability, it has not worked to stop them at all. It's their very comprehensive offering, proven track record and the momentum that has shielded AWS and will continue to indefinitely.

AWS is losing marketshare to Azure and GCP. This is big deal, it was unexpected after years of Google/Microsoft trying and failing.

Further AWS is losing share at a time when GCP and Azure are becoming profitable businesses, so no longer losing money to gain market share.

It’s already playing out. Just look at recent results. While once light years ahead competitors are now closing ranks and margins are under pressure. AWS clearly isn’t going away, but on the current trajectory its future as the leading cloud is very much not a certainty.

Because if LLM inference is going to be a bigger priority for the majority of companies, they're going to go where they can get the best performance to cost ratio. AWS is falling behind on this. So companies (especially new ones) are going to start using GCP or Azure, and if they're already there for their LLM workloads, why not run the rest of the infrastructure there?

It's similar to how AWS became the de-facto cloud provider for newer companies. They struggled to convince existing Microsoft shops to migrate to AWS, instead most of the companies just migrated to Azure. If LLMs/AI become a major factor in new companies deciding which will be their default cloud provider, they're going to pick GCP or Azure.

Except for spending cloud budgets on LLMs elsewhere like other mentioned, LLM coding will make it easier to convert codebases from being AWS dependent, easing their lock-in

So does Pichai... Every time there is something new, he forces Google to pivot, upending everything without much to show for it.

When did Google ever pivot?

Social. User-facing AI.

Social: YouTube User facing AI: Gemini, Google photos, NotebookLM and plenty of products.

I haven't followed that closely, but Gemini seems like a pivot based on ChatGPT's market success

Google is leading in terms of fundamental technology, but not in terms of products

They had the LLambda chatbot before that, but I guess it was being de-emphasized, until ChatGPT came along

Social was a big pivot, though that wasn't really due to Pichai. That was while Larry Page was CEO and he argued for it hard. I can't say anyone could have known beforehand, but in retrospect, Google+ was poorly conceived and executed

---

I also believe the Nth Google chat app was based on WhatsApp success, but I can't remember the name now

Google Compute Engine was also following AWS success, after initially developling Google App Engine

Is it really such a pivot when they invested a lot in AI already?

They bought DeepMind in 2014 and always showed of a ton of AI research.

>I haven't followed that closely, but Gemini seems like a pivot based on ChatGPT's market success

"AI" in it's current form is already a massive threat to Google's main business (I personally use Google only a fraction of what I used to), so this pivot is justified.

None of these are pivots. The core business has always been the core business.

If you are defining "pivot" as "abandon all other lines of business", then no, none of the BigTechs have ever pivoted.

By more reasonable standards of "pivot", the big investment into Google Plus/Wave in the social media era seems to qualify. As does the billions spent building out Stadia's cloud gaming. Not to mention the billions invested in their abandoned VR efforts, and the ongoing investment into XR...

Well, "pivot" implies the core business has failed and you're like "oh shit, let's do X instead".

Stadia was just Google's New Coke, Apple's Mac Cube, or Microsoft's MSNBC (or maybe Zune.

When they can't sell ads anymore, they'll have to pivot.

> Well, "pivot" implies the core business has failed and you're like "oh shit, let's do X instead".

I mean, Facebook's core business hasn't actually failed yet either, but their massive investments in short-form video, VR/XR/Metaverse, blockchain, and AI are all because they see their moat crumbling and are desperately casting around for a new field to dominate.

Google feels pretty similar. They made a very successful gambit into streaming video, another into mobile, and a moderately successful one into cloud compute. Now the last half a dozen gambits have failed, and the end of the road is in sight for search revenue... so one of the next few investments better pay off (or else)

The link you posted has a great many very insignificant investments included in it, and nothing I've seen Google doing has felt quite like the desperation of Facebook in recent years.

I didn't really see it at first, but I think you are correct to point out that they kind of rhyme. However to me, I think the clear desperation of Facebook makes it feel rather different from what I've seen Google doing over the years. I'm not sure I agree that Google's core business is in jeopardy in the way that Facebook's aging social media platform is.

G+ was a significant effort that was abandoned.

I'd personally define that as Google hedging their bet's and being prepared in case they needed to truly pivot, and then giving up when it became clear that they wouldn't need to. Sort of like "Apple Intelligence" but committing to the bit, and actually building something that was novel, and useful to some people, who were disappointed when it went away.

Stadia was always clearly unimportant to Google, and I say that as a Stadia owner (who got to play some games, and then got refunds.) As was well reported at the time, closing it was immaterial to their financials. Just because spending hundreds of millions of dollars or even a few billion dollars is significant to you or I doesn't mean that this was ever part of their core business.

Regardless, the overall sentimentality on HN about Google Reader and endless other indisputably small projects says more about the lack of strategic focus from people here, than it says anything about Alphabet.

Google basically invented modern AI (the 'T' in ChatGPT stands for Transformer), then took a very broad view of how to apply broadly neural AI - AlphaGo, AlphaGenome being the kind of non-LLM stuff they've done).

A better way to look at it is that the absolute number 1 priority for google since they first created a money spiggot throguh monetising high-intent search and got the monopoly on it (outside of Amazon) has been to hold on to that. Even YT (the second biggest search engine on the internet other than google itself) is high intent search leading to advertising sales conversion.

So yes, google has adopted and killed lots of products, but for its big bets (web 2.0 / android / chrome) it's basically done everything it can to ensure it keeps it's insanely high revenue and margin search business going.

What it has to show for it is basically being the only company to have transitioned as dominent across technological eras (desktop -> web2.0 -> mobile -> (maybe llm).

As good as OpenAI is as a standalone, and as good as Claude / Claude Code is for developers, google has over 70% mobile market share with android, nearly 70% browser market share with chrome - this is a huge moat when it comes to integration.

You can also be very bullish about other possible trends. For AI - they are the only big provider which has a persistent hold on user data for training. Yes, OpenAI and Grok have a lot of their own data, but google has ALL gmail, high intent search queries, youtube videos and captions, etc.

And for AR/VR, android is a massive sleeping giant - no one will want to move wholesale into a Meta OS experience, and Apple are increasingly looking like they'll need to rely on google for high performance AI stuff.

All of this protects google's search business a lot.

Don't get me wrong, on the small stuff google is happy to let their people use 10% time to come up with a cool app which they'll kill after a couple of years, but for their big bets, every single time they've gone after something they have a lot to show for it where it counts to them.

Yeah, and Google has cared deeply about AI as a long term play since before they were public. And have been continuously invested there over the long haul.

The small stuff that they kill is just that--small stuff that was never important to them strategically.

I mean, sure, don't heavily invest (your attention, time, business focus, whatever) in something that is likely to be small to Google, unless you want to learn from their prototypes, while they do.

But to pretend that Google isn't capable of sustained intense strategic focus is to ignore what's clearly visible.

I suppose you could argue that Amazon does have one special thing going for it here, idle compute resources in AWS. However that is not the sort of thing that requires "AI talent" to make use of.

They also have made pretty big investments in cloud VMs with GPUs attached, so they are making money off the AI craze regardless

The largest LLMs are mostly going to be running in the cloud, so the general purpose cloud providers (Amazon, Microsoft, Google) are presumably going to be in the business of serving models, but that doesn't necessarily mean they need to build the models themselves.

LLMs look to be shaping up as an interchangeable commodity as training datasets, at least for general purpose use, converge to the limits of the available data, so access to customers seems just as important, if not more, than the models themselves. It seems it just takes money to build a SOTA LLM, but the cloud providers have more of a moat, so customer access is perhaps the harder part.

Amazon do of course have a close relationship with Anthropic both for training and serving models, which seems like a natural fit given the whole picture of who's in bed with who, especially as Anthropic and Amazon are both focused on business customers.

Microsoft is building it's own in-house LLM's based on OpenAI's IP. Google builds it's own models.

Sure, but you can also sell something without having built it yourself, just as Microsoft Copilot supports OpenAI and Anthropic models.

It doesn't have to be either/or of course - a cloud provider may well support a range of models, some developed in house and some not.

Vertical integration - a cloud provider building everything they sell - isn't necessarily the most logical business model. Sometimes it makes more sense to buy from a supplier, giving up a bit of margin, than build yourself.

I'm just an observer. Microsoft has invested billions in OpenAI and can access their IP as a result. It might even be possible MS hopes that OpenAI fails and doesn't allow them to restructure to continue to acquire outside funding. You can go directly to the announcement of their in-house model offerings and they are clearly using this as a recruiting tool for talent. Whether it makes sense for the cloud providers to build their own models is not for me to say, but they may not have a choice given how quickly OpenAI/Anthropic are burning cash. If those two fail then they're essentially ceding the market to Google.

> But actually, Amazon, Apple etc aren't natural homes for this, they don't need to burn money to chase it.

I really liked the concept of Apple Intelligence with everything happening all on device, both process and data with minimal reliance off device to deliver the intelligence. It’s been disappointing that it hasn’t come to fruition yet. I am still hopeful the vapor materializes soon. Personally I wouldn’t mind seeing them burning a bit more to make it happen.

It will likely occur, just maybe not this year or next. If we look over the last eighty years of computing, the trend has been smaller and more powerful computers. No reason to think this won’t occur with running inference on larger models.

I generally agree with you, although Amazon is really paranoid about being behind here.

On the last earnings call the CEO gave a long rambling defensive response to an analyst question on why they’re behind. Reports from the inside also say that leaders are in full blown panic mode, pressing teams to come up with AI offerings even though Amazon really doesn’t have any recognized AI leaders in leadership roles and the best talent in tech is increasingly leaving or steering clear of Amazon.

I agree they should just focus on what they’re good at, which is logistics and fundamental “boring” compute infrastructure things. However leadership there though is just all over the map trying to convince folks their not behind vs just focusing on strengths.

Which bar raiser is going to raise the bar first??

Doesn't Amazon have a huge lead just because of AWS? Every other player is scrambling for hardware/electricity while Amazon has been building out data centers for the last 20 years.

In theory they should, but it’s increasingly looking like they’re struggling to attract/retain the right talent to take advantage of that position. On paper they should be wiping the floor with others in this space. In practice they’re getting their *ss kicked and in a panic on what to do.

> Doesn't Amazon have a huge lead just because of AWS?

They have huge exposure because of AWS; if the way people use computing shifts, and AWS isn't well-configured for AI workloads, then AWS has a lot to lose.

> Every other player is scrambling for hardware/electricity while Amazon has been building out data centers for the last 20 years.

Microsoft and Google have also been building out data centers for quite a while, but also haven't sat out the AI talent wars the way Amazon has.

> AWS isn't well-configured for AI workloads

What does that mean? Not enough GPUs?

A few things, including:

1. Price-performance has struggled to stay competitive. There’s some supply-demand forces at play, but the top companies consistently seem to strike better deals elsewhere.

2. The way AWS is architected, especially on networking, isn’t ideal for AI. They’ve dug their heels on in their own networking protocols despite struggling to compete on performance. I personally know of several workloads that left AWS because they couldn’t compete on networking performance.

3. Struggling on the managed side. On paper a service like Bedrock should be great but in practice it’s been a hot mess. I’d love to use Anthropic via Bedrock, but it’s just much more reliable when going direct. AWS has never been great at these sort of managed services at scale and they’re again struggling here.

My understanding is that they fell behind on offering the latest gen Nvidia hardware (Blackwell/Blackwell Ultra) due to their focus on internally developed ASICs (Trainium/Inferentia gen 2).

> Zuckerberg rushing into every new fad with billions of dollars has somehow tricked people into thinking that's what big tech is about and all of them should be shovelling money into this.

Zuckerberg failed every single fad he tried.

He's becoming more irrelevant every year and only the company's spoils from the past (earned not less by enabling, for example, a genocide to be committed in Myanmar https://www.pbs.org/newshour/world/amnesty-report-finds-face...) help carry them through to the series of disastrous idiotic decision Zuck is inflicting on them.

- VR with Oculus. It never caught on, for most people who own one, it's just gathering dust.

- Metaverse. They actually spend billions on that? https://www.youtube.com/watch?v=SAL2JZxpoGY

- LLAMA is absolute trash, a dumpster fire in the world of LLMs

Zuck is now trying to jump again on the LLM bandwagon and he's trying to...buy his way in with ridiculous pay packages: https://www.nytimes.com/2025/07/31/technology/ai-researchers.... Why is he so wrong to do that, you might ask?

He is doing it at the worst possible moment: LLMs are stagnating and even far better players than Meta like Anthropic and OpenAI can't produce anything worth writing about.

ChatGPT5 was a flop, Anthropic are struggling financially and are lowering token limits and preparing users for cranking up prices, going 180 on their promises not to use chat data for training, and Zuck, in his infinite wisdom, decides to hire top AI talent for premium price at a rapidly cooling market? You can't make up stuff like that.

It would appear that apart from being an ass kisser to Trump, Zuck shares another thing with the orange man-child running the US: a total inability to make good, or even sane deals. Fingers crossed that Meta goes bankrupt just like Trump's 6 banrkruptcies and then Zuck can focus on his MMA career.

meta made $62 billion dollars last year. Mark burns all this money because his one and only priority is making sure his company doesnt become an also ran. The money means nothing to him

Yet his company and him are becoming rapidly irrelevant.

I've been taking heat for years for making fun of the metaverse. I had hopeful digital landlords explain to me that theyll be charging rent in there! Who looked at that project and thought it was worth anything?

> I've been taking heat for years for making fun of the metaverse

I don't know in what circles you're hanging out, I don't know a single person who believed in the metaverse

> I don't know in what circles you're hanging out, I don't know a single person who believed in the metaverse

Oh please, the world was full of hype journalists wanting to sound like they get it and they are in it, whatever next trash Facebook throws their way.

The same way folks nowadays pretend like the LLMs are the next coming of Jesus, it's the same hype as the scrum crowd, the same as crypto, nfts, web3. Always ass kissers who cant think for themselves and have to jump on some bandwagon to feign competence.

Look at what the idiots at Forbes wrote: https://www.forbes.com/councils/forbestechcouncil/2023/02/27...

They are still very influential, despite having shit takes loke that.

Accenture still think the Meta is groundbreaking: https://www.accenture.com/us-en/insights/metaverse

What a bunch of losers!

71% of executives seemed to be very excited about it: https://www.weforum.org/stories/2022/04/metaverse-will-be-go...

Executives (like Zuck) are famous for being rather stupid so if they are claiming something, you bet its not gonna happen.

Apparently, "The metaverse is slowly becoming the new generation’s digital engagement platform, but it’s making changes across enterprises, too."

https://www.softserveinc.com/en-us/blog/the-promise-of-the-m...

i don't care about virtual real estate, but VR mini golf sure is fun!

Zuckerberg knows what he's doing, his net worth is >250 billion dollars now.

Go all in the new fad, investors pile up on your stock, dump, repeat...

> Zuckerberg knows what he's doing, his net worth is >250 billion dollars now.

Does he have this net worth because what he is doing or despite what he is doing? :-)

Correlation does not imply causation. Attribution is hard.

Zuckerberg bought Whatsapp and Instagram. For normal people, those replaced 90% of the internet here in Argentina

(The other 10% is mostly Google Maps and MercadoLibre.)

But that didn't require deep insight. Both were already really popular and clearly a threat to Facebook. WhatsApp was huge in Europe before they bought (possibly other places as well).

Buying competition is par for the course for near-monopolies in their niches. As long as the scale differences in value are still very large, you can avoid competition relatively cheaply, while the acquired still walk away with a lot of money.

Why does investing in AI require deep insight? ChatGPT is already huge, significantly bigger than Whatsapp when the deal was done. And while OpenAI is not for sale, he figured that their employees are. Also not to mention, investors are very positive for AI.

So far there hasn't been a transformative use case for LLMs besides the straightforward chat interface (Or some adjacent derivative). Cursor and IDE extensions are nice, but not something that generates billions in revenue.

This means there's two avenues:

1. Get a team of researchers to improve the quality of the models themselves to provide a _better_ chat interface

2. Get a lot of engineers to work LLMs into a useful product besides a chat interface.

I don't think that either of these options are going to pan out. For (1), the consumer market has been saturated. Laymen are already impressed enough by inference quality, there's little ground to be gained here besides a super AGI terminator Jarvis.

I think there's something to be had with agentic interfaces now and in the future, but they would need to have the same punching power to the public that GPT3 did when it came out to justify the billions in expenditure, which I don't think it will.

I think these companies might be able to break even if they can automate enough jobs, but... I'm not so sure.

> Cursor and IDE extensions are nice, but not something that generates billions in revenue.

I mean Cursor is already at $500 million ARR...

How many software engineers are there in the world? How many are going to stop using it when model providers start increasing token cost on their APIs?

I could see the increased productivity of using Cursor indirectly generating a lot more value per engineer, but... I wouldn't put my money on it being worth it overall, and neither should investors chasing the Nvidia returns bag.

Whatsapp had $10M revenue when it was acquired[1]. Lots of so called "chatgpt wrappers" has more revenue than that. While in hindsight Whatsapp acquisition at $19B seems no brainer, no concrete metric pointed to that compared to him investing $19B in AI now.

[1]: https://www.sec.gov/Archives/edgar/data/1326801/000132680114...

Dude Zuckerberg bought whatsapp because FB Messenger was losing market share... nothing to do with Whatsapps revenue! Rather Zuckerbergs fear of FB products being displaced.

Pretty sure everyone was balking at the purchase prices at the time

In the UK it was an obvious purchase - whatsapp was the main mode of communicaton on a phone. Nobody used Messenger for instance.

I'd argue that Meta's income derives in no small part from their best in class ad targeting.

Being on the forefront of

(1) creating a personalized, per user data profile for ad-targeting is very much their core business. An LLM can do a very good job of synthesizing all the data they have on someone to try predicting things people will be interested in.

(2) by offering a free "ask me anything" service from meta.ai which is tied directly to their real-world human user account. They gather an even more robust user profile.

This isn't in-my-opinion simply throwing billions at a problem willy nilly. Figuring out how to apply this to their vast reams of existing customer data economically is going to directly impact their bottom line.

>An LLM can do a very good job of synthesizing all the data they have on someone to try predicting things people will be interested in.

Is synthesizing the right word here?

I think is absolutely is, LOL. Though a "very good job of synthesizing" might not actually good for much...

5 minutes on facebook being force-fed mesopotamian alien conspiracies is all you'll need to experience to fully understand just how BADLY they need some kind of intelligence for their content/advertising targeting, artificial or not...

Obviously one is a very bad sample, but why are the ads I see on FB so badly targetted?

You probably don't spend enough time on their sites to have a good ad targeting model of you developed. The closer you are to normal users, with hundreds of hours of usage and many ad clicks, the more accurate the ads will be for you.

You mean the closer I am to the top of the bell curve, the more your ads "shooting from the hip" will land? Who would've thunk it?!

Did you block their tracking across the whole damn internet, by any chance?

Same terrible experience for me while I was on FB. I was spending a lot of time there and I do shop a lot online. They couldn’t come with relevant ad targeting for me. For my wife they started to show relevant ads AFTER she went to settings and manually selected areas she is interested in. This is not an advanced technology everyone claim FB has.

Instagram has killer ad targeting; no wonder all these direct-to-consumer brands flock there. FB not so much I agree.

People look at all the chaos in their AI lab but ignore the fact that they yet again beat on earnings substantially and directly cited better ad targeting as the reason for that. Building an LLM is nice for them, but applying AI to their core business is what really matters financially, and that seems like it's going just fine.

> But actually, Amazon, Apple etc aren't natural homes for this, they don't need to burn money to chase it.

Amazon is the biggest investor of AI of any company. They've already spent over $100b YTD on capex for AI infrastructure.

To do what for that money? Write summaries of product reviews? If they wanted to do something useful, they'd use the LLM to figure out which reviews are for a different product than what is currently being displayed.

"useful" means different things

This is "shovels" they rent out; very different then research investment.

Zuckerbergs AI "strategy" seems to be to make it easy for people to generate AI slop and share it on FB thus keeping them active on the platform. Or to give people AI "friends" to interact with on FB, thus keeping them on the platform and looking at ads. It's horrifying but it does make business sense (IMHO) at least at first glance.

Except that Amazon's AWS business is severely threatened by the rise of alternative cloud providers who offer much more AI-friendly environments. It's not an existential topic yet, but could easily turn into one.

Microsoft has the pleasure of letting you pay for your own hosted GPT models, Mixtral, etc

Microsoft's in a sweet spot. Apple's another interesting one, you can run local LLM models on your Mac really nicely. Are they going to outcompete an Nvidia GPU? Maybe not yet, but they're fast enough as-is.

I think this analysis is a bit shallow with regard to Metas product portfolio and how AI fits in.

Much more than the others, metter runs a content business. Gen AI aides in content generation so it behooves them to research it. Even before the current explosion of chatbots, meta was putting this stuff into their VR framework. It's used for their headset tracking and speech to text is helpful for controlling a headset without a physical keyboard.

You're making it sound like they'll follow anything that walks by but I do think it's more strategic than that.

They have out sensors though for any AGI, because AGI could subvert buisness fields and expertise moats. Thats what most AI teams are- vanity projects and a few experts calming the higher ups every now and then with a "its still just autocompletion on steroids, it can not yet do work and research alone."

By that logic, a social media company shouldn't rush into it either, but they did anyway.

Exactly. Being a tech company doesn't mean you need to do everything any more than just because you're a family doctor you also should do trauma surgery, dentistry, and botox injections. Pick a lane, be an expert in it.