Hacker Newsnew | past | comments | ask | show | jobs | submit | raincole's commentslogin

Not just DRAM market, but the GPU market soon. China is the great equalizer of the world.

> Not just DRAM market, but the GPU market soon

China doesn't have EUV fabs... They've pushed DUV impressively far... but until they get EUV working industrially (and reasonable timelines are at least 2-4 years for that) it shouldn't be possible for them to compete for that market.

> China is the great equalizer of the world.

China is hardly an egalitarian society...


The future is here now, it’s just not evenly distributed. China will mass produce something to the point that it is widely distributed. That is how China acts as a great equalizer on a global scale.

Another way China is a great equalizer is their willingness to do business with anyone that can pay.


There are no egalitarian societies. Societies in the west favour the super rich and believing anything else is simply delusional. Sorry to burst the bubble.

Not an egalitarian society, but their companies have a honey-badger like mentality from what I have read, where they ruthlessly reduce costs and margin down past where non-Chinese companies cannot compete.

How much does the node size matter for dram? My understanding was that it’s been marginal gains on sram since about 7nm TSMC. I would naively expect the capacitor size requirements not to shrink as well as logic, does the smaller transistor make up for the lower capacitance, or do they have to run at higher frequencies and refresh more frequently?

No one says that China is an egalitarian society lmao. Where did you get the idea?

Not really sure if I'd describe what China has done to various manufacturing sectors as "equalizing".

As someone who doesn't know jj and read this article, it does sound like `git rebase -i` to me. I'm sure that if I actually spent time learning jj I'd know the difference though.

When it comes to something with random results (unfortunately that's what LLMs are), people will think the odds are rigged against them.

It's a good thing that hype-chasers are cancelling though. So we can use the services with a reasonable latency.


Wow... what happened to Fortune? I thought they're far above this kind of clickbait.

Now use common accounting standard and amortize the cost.

Oh it doesn't fit the narrative. Never mind then.


The depreciation is also insane and thus to sustain operations and improve, the spend will keep going.

I assume you are saying it would look less ridiculous? By how much?

If OpenAI and Anthropic adopted GAAP nobody would be able to invest in them it would be so bad

Yeah, institutional investors who plowed billions into them are unsophisticated rubes who got hoodwinked because they don't get GAAP. And it's not like both OpenAI and Anthropic are both going to IPO soon which would require disclosures far beyond GAAP numbers. /s

I take it you weren't alive in 2008 if you say that sarcastically

You should do a Michael Burry and go short. Clearly you seem to think you are much better informed, put your money where your mouth is.

Shorting isn't the magic trick people seem to think it is, but yes I'm heavily hedged for OpenAI's IPO failing and taking Oracle down with it

In what ways do common accounting standards and amortizing the costs (this is tricky for ai and the current batch of gpus I hear!) change the data presented here? Does it detract from the point? Completely contradict it?

You can turn your drive-by dismissal into something really informative if you want to.


First of all, the whole website is based on what the CEOs said they're going to spend. Not the actual money spent. So there is no real 'data' presented here or to contradict.

Second, even if you take CEOs' words at face value, they didn't distinguish the capex for hardware, electricity, software and salary. You can make up whatever the percentage for hardware and the depreciation rate you believe and fit an arbitrary narrative.


In other words, PG is so rich now that he forgets the existence of less fortunate citizens.

They're selling at a loss (obviously).

But why not? Gaining market share at a loss isn't the US's patent.


They haven't raised enough money to be selling at a loss. And selling at a loss to gain market share in an industry with zero switching friction between sellers is not a strategy. That doesn't make sense.

Loss leading only works when

- it leads to a situation that allows you to prevent competitors from selling to your customers (gilded age railroad and pipeline industries are great examples). Then you can eventually raise prices and not lose back any market share.

- or when it allows you to remarket to customers and make back the difference (selling a single console at a loss to sell a whole library of high margin videos games, or selling jet engines at a loss to lock in 30-year maintenance contracts).


Yeah, cool theory, but they are selling at a loss. We know that because their model is open and available on other providers too. No other provider even sells a quantitized version of DeepSeek V4 Pro at that price.

Also, in case of LLM, market share = more people uploading their whole codebase/legal documents/unfinished books/literally everything to your servers for you to use in future training. So the incentive to sell at a loss is much stronger than other kinds of service.


We are missing the fact that they have created their GPU's that are now just 4-5 years behind. And considering it's China, which does everything-hardware at insane scale, and efficiency, my guess is that they are at step-1 now... gain market share at loss, and at the same time, gradually, start plugging their in-house cards to power these models to gauge their performance on real workloads.

Once they cross a certain threshold, nVidia can say goodbye to it's monopolisitic profit margins of over 70%.

GPU infra capex is the biggest spend for the inference providers as of now, power, second biggest.

China has already cracked the power part, they are now close to cracking the GPU part.


Didn’t the DeepSeek team release a paper documenting inference improvements that showed they were still making a profit even under heavy discount? Why would it be impossible for them to make a profit now, with a new model and more research?

Before DeepSeek, no one sold cheap tokens anyways and then DS showed the profit margins.


they might have trained the model with fancy optimisations that only they can unlock

Maybe Anthropics efforts to thwart deepseek from distilling their model is bearing fruit.

So their strategy now is to try get as much raw content for their inference. You're being "paid", via discount, for your use


> So their strategy now is to try get as much raw content for their inference. You're being "paid", via discount, for your use

There is an implicit social contract, and for many it might work out well:

We use your data to improve the model. You get to use the improved model for affordable prices and (the important part): you get _the model_.


From Antropics own report:

"DeepSeek

Scale: Over 150,000 exchanges"

Doesn't sound like much of distilling. Maybe they are runnung benchmarks?


Proof?

Renters will always pay one way or another. You can name it wealth tax or property tax or house tax. It doesn't matter -- the result will be higher rent.

Economy's doing good? Rent increases because people can afford to pay more, and demand is always very high.

Economy's doing badly? Rent increases because costs and taxes for the landlord are higher.

UBI? Rent increases because people have more disposable income.

---

The secret to the infinite money glitch is to maintain a much lower supply than the demand, and to concentrate jobs within a small area


The point of my post was to indicate that just because you're a home owner (rather than a renter) does not make you special.

Property taxes are not wealth taxes, but fees for services rendered by the local government. Both home owners and renters (may) benefit from those fees.


Humans work like that too. If you're not comfortable with Claude involves in every step (for whatever reason) then just use different providers for each.

DeepSeek is very likely selling tokens at a loss. There're many cloud providers that provide you with DeepSeek V4 Pro via API, and those services at least twice as expensive as DeepSeek itself.

^Sorry for this understatement. DeepSeek is actually selling tokens at a far cheaper price than my previous comment implied.

DeepSeek V4 Pro price on OpenRouter:

deepseek: $0.435 / $0.87

baidu/fp8: $1.521 / $3.042

novita/fp8: $1.64 / $3.38

Yup. DeepSeek either has next-generation hardware that somehow no one else has access to, or they're selling at a loss.


> next-generation hardware that somehow no one else has access to

not necessarily 'next-gen', but they've optimized for the Huawei Ascend 950 right? not a lot of those outside china at least


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: