Interesting if true - that Moonshot can train a ~3T SOTA model on only 20K NVIDIA GPUs, while others like Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs).
I guess at least partly a reflection of all the optimizations in the Kimi 3 architecture.
In the recent leaked DeepSeek investor meeting, they also mentioned only having a 20K GPU cluster (unclear if NVIDIA, or Huawei).
> A person familiar with Moonshot’s procurement strategy confirmed that the company does indeed have a channel for accessing Blackwell processors via Southeast Asia. They didn’t specify whether this was a rental channel — which is legal, in most cases — or direct purchases, which are a breach of US regulations. The Information reported this week that Moonshot is seeking additional Blackwell processors to train its next model.
> The 20,000 chips Moonshot accesses via Alibaba, meanwhile, are from Nvidia’s earlier generation of Hopper products, the people familiar with the agreement said.
So the 20k GPUs from Alibaba is only a lower bound on how many you need to train a model like Kimi K3.
The advantage of having more GPUs in any case is not so much that you can train bigger models, but that the turnaround time is faster, so you can run more experiments to dial in training choices. It's entirely possible that Musk has more than enough compute, but can't hire the talent to run all those experiments. (That would also explain why he has excess capacity he can rent to Google.)
From a superficial look I think the submitted title ("Moonshot built on 20k Nvidia chip cluster from Alibaba") might have been misleading. Article says "That hardware forms a substantial chunk of the overall computing capacity Moonshot uses for its Kimi models", implying that other hardware was also used.
(We've changed the title to what the article says now.)
It's not that surprising to me. Most of the innovation in Chinese models has been in efficiency gains and optimizations. K3 coming from the factory in MXFP4 weights is a pretty relevant factor. Big performance gap probably also due to Moonshot doing QAT. Throw in the fact that Musk has easier access to compute, and I think you have your answer on the disparity.
Do we believe them? It seems there’s no reason with the resources OpenAI and Anthropic have they wouldn’t be incentivised to build equally efficient pipelines? I think it’s mostly the fine tuning with distillation that gives Chinese labs their big advantage. However I’m pretty sure all the main labs are poisoning the output now when they detect Chinese activity so while K3 looks frontier on tests when you use it it’s miles off Sol and Fable.
everything about the current 'a.i' models - east vs west points the wastefulness of western labs - whether in terms of gpu clusters being ran, cost of inference.
the only way is down for the massive valuations and 'a.i' revenue projections.
> west points the wastefulness of western labs whether in terms of gpu clusters being ran, cost of inference.
Can’t possibly be intentional media strategy by a geopolitical target, right? If its going to negatively impact valuations and revenue of major rivals, seems like a desirable strategy?
We just saw OpenAI take the steps to significantly lower the cost of one of their models, which confirms that at least one western lab has a large margin on inference, not a large inference cost.
Meanwhile, we don’t know how many experiments these east/west labs are performing relative to each other. We also know that many western labs have a whole portfolio of models too, which is product breadth not necessarily waste.
The wastefulness is akin to having multiple private jets on standby, or workers coming in to a fishbowl office to be gawked at, or massive datacenters to be used as social "proof".
Its all a fucking capitalistic farce to display to other rich elite that "Look at how much clout I have! I can make these peons dance around and do my bidding! Im a slave-owner!"
He noted that that Silicon Valley doesnt really want to SOLVE problems. They want to find already-solved problems with problem matching. And of course, we just throw more people and more compute instead of optimization and understanding.
The Chinese are being actively constrained with bullshit politics around a second Red Scare moment. And, well, they're winning. A lot.
> Struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs).
This feels like a very American way of designing things - just throw more horse power at it, bigger is better! The rest of the world is usually a bit more resource constrained and efficient at using those resources.
See also Mustangs vs German sports cars, giant American fridges, giant American suburban McMansions vs livable cities etc etc.
> Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs)
Are you sure they are using all of their compute on training? Didn't they rent out a ton to other AI companies?
It's obviously a question - did they just train for 10x as long due to having 10x fewer GPUs, but then that spoils the claim that they distilled Fable which was only recently introduced. Now doubt they did use some training data generated from older US models though.
This is just not how it works - it is perfectly plausible (in fact the most likely) that pretraining was well finished by the time Fable was released. A bit of extra distillation takes far less compute.
Moreover, it’s very plausible (and expected) to use multiple clusters and GPU types for RL rollouts which could very well not be included in this count.
Relatedly I also wonder how much the orchestrated distillation bot nets that Anthropic uses against China are real. Not saying it does not happen but I have never seen dating showing how pervasive it is and I always imagine other western labs would absolutely be doing the same.
I’m still baffled that OpenAI can complain about this with a straight face while these models are literally trained on everything regardless of copyright or permission.
It kind of confirms a hypothesis I have that the next phase of AI development will be about getting smaller (in terms of model size and compute), because smaller is more capital efficient for training (allowing faster iteration and more iteration cycles for a given amount of capital), allows for denser inference (more inference for a given amount of compute hardware), and allows for more edge inference applications.
The goal will be to develop smaller models with more efficient architectures, that have similar or even better performance than larger models.
Yes, continual learning is currently a hot topic since this is what will allow an "AI intern or new employee" to learn on the job rather than be stuck on groundhog day, or pre-trained for every eventuality - needing to anticipate all the quirks and proprietary knowledge of every customer!
Continual learning tends to imply individualized models, else there is no data privacy (the secrets learned on the job at your company now being available to your competition), which really turns the current AI business model of a single centralized model served to everyone on it's head. If every customer has a different model that essentially means the end of batch processing with the same weights loaded into the GPU.
The direction this suggests is a move away from centrally served common models to locally served individual ones, which generally requires them to be smaller, even if some larger companies may be willing to invest in beefier hardware.
I think this is at least in part why the AI companies are trying NOT to implement true continual learning and see if they can instead finesse it by implementing continual compacted(?) memorization instead, since then it's "just" additional context that needs to be recalled and fed into every request, not weights that need feeding into the GPU. I don't think memorization is any substitute for learning, especially learning of practiced skills, but since it's far easier to implement, and non-disruptive to the cloud-based API business model, this is what we will see first.
The recent news of NVIDIA' investment in Sutskever's SSI has a tiny hint of this also, talking about SSI advising on NVIDIA's future architectural direction - apparently pushing it in a different direction than current (cloud-based, pre-trained) models. NVIDIA may be quite happy to see a move towards local models.
The middle ground would be continual learning for a group of users. For a company with a couple thousand employees, one model learning on all your employees might be viable. With some filtering for information that is supposed to stay compartmentalized (HR data, stuff under NDA, core technolgy). But this only really works if you bring enough volume to make it economical to effectively server you a whole other model
It feels like we're not really bottlenecked by model performance now anyway. Most cases I find where a model is being an idiot can be remedied by rethinking the context and tooling available to that model.
If they got a lot cheaper and only a little dumber, and we got a little smarter about how we use them, they could appear to the bystander as much more useful than they are.
It begs an interesting question as well: are frontier labs stuck in the incumbent phase of the "innovator's dilemma," where there's such intense pressure to maintain the flavor/character of their widely-loved existing models, that they dare not invest their resources into radical reimagination of their approach towards architectures for smaller models? A company like Moonshot does not have this cultural limitation.
The frontier labs would be well served in carving out 20k sub-clusters and giving research teams carte blanche in building things with radically different architectures - with full permission to distill whatever they want from the flagship models. We'd expect to see more product lines that feel "different" from the flagship models if this were already being done.
Eventually, multi-modality might even get offloaded as workflows, which might allow models to do this at a fraction of the data and compute required.
I am adding multi-modality to https://github.com/guilt/TinyToT, and I see that dis-aggregating capabilities, very similar to how our own sensory organs work, seems to be paying off quite well.
This is likely what Google has been doing, as it suites them best to have small fast models rather than hulking slow giants. Lower intelligence but way more ability to serve. Anthropic would probably need a datacenter the size of a small country to serve Fable on a Google search/services scale.
Yes but that's because of the scaling laws for transistors. ML models seem to get better the bigger they are. If you want to compress the world's information, you need to look at all the information in the world.
My thesis is still that model building has no moat. Folks continue to migrate around between the big labs. There is a lot of value in having good taste around the harness and how the models are used. The medium to long term winners will be the folks that control the compute.
You can only control the compute if it is not commoditized, which at the current moment, it is not, leading to industry players having 75-80% margins.
Someone will find a way to make cheaper compute, and since nothing fundamental changed there (LLM didn't change how silicon were made), that's bound to happen.
If Chinese companies are able to just reliably rent the damn GPUs then I really don't see how these restrictions are meant to be effective in the slightest. I guess it makes it less convenient and supply less certain, and retains some optionality for cutting off supply later?
If Alibaba can't bring the chips into China, but can buy a whole bunch in Thailand or Singapore or whereever (or secure exclusive rights via JV partners where relevant) and then just provide them as a service to its customers in China - what is the point? I'm sure many customers would actually prefer such an arrangement.
I would not minimize the extent to which AI progress depends on things like the effort that goes into training material acquisition and data labeling. In those areas, improvement is a direct function of how much you invest.
I wonder how much spending is motivated by the phenomenon of sudden emergent performance in LLMs. Clearly some people who are smarter than me expect something like emergent AGI, or at least they think the odds justify spending whatever it takes to see if that would happen.
That leaves a lot of room for efficient aggressive followers.
Huawei A910 is supposed to have native support for MXFP4. The problem is that huawei doesn't have the capacity to build more GPUs right now. Those yield numbers must be really terrible.
Personally, I think 20K nvidias is a stop gap solution because they really don't have the capacity to serve their models to earn any money right now.
A person familiar with Moonshot’s procurement strategy confirmed that the company does indeed have a channel for accessing Blackwell processors via Southeast Asia. They didn’t specify whether this was a rental channel...
They mean that K3's native floating point is MXFP4, a newer standard that's not supported in Hopper, an older nvidia chipset, so it's impossible the GPUs in question were Hopper (the article specifically mentions H200s), which is correct.
Probably either the author or the people they interviewed got the specific GPU details mixed up or wrong. Maybe it was B200s.
It's also built on the back of downloading all the content of the top ten sites on the list, the fact that all the independent content sites with their "pre-war" steel haven't been acquired at top dollar by U.S. companies shows the existential threat / stop China posturing is a total joke.
Interesting if true - that Moonshot can train a ~3T SOTA model on only 20K NVIDIA GPUs, while others like Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs).
I guess at least partly a reflection of all the optimizations in the Kimi 3 architecture.
In the recent leaked DeepSeek investor meeting, they also mentioned only having a 20K GPU cluster (unclear if NVIDIA, or Huawei).
From the article:
> A person familiar with Moonshot’s procurement strategy confirmed that the company does indeed have a channel for accessing Blackwell processors via Southeast Asia. They didn’t specify whether this was a rental channel — which is legal, in most cases — or direct purchases, which are a breach of US regulations. The Information reported this week that Moonshot is seeking additional Blackwell processors to train its next model.
> The 20,000 chips Moonshot accesses via Alibaba, meanwhile, are from Nvidia’s earlier generation of Hopper products, the people familiar with the agreement said.
So the 20k GPUs from Alibaba is only a lower bound on how many you need to train a model like Kimi K3.
The advantage of having more GPUs in any case is not so much that you can train bigger models, but that the turnaround time is faster, so you can run more experiments to dial in training choices. It's entirely possible that Musk has more than enough compute, but can't hire the talent to run all those experiments. (That would also explain why he has excess capacity he can rent to Google.)
> they also mentioned only having a 20K GPU cluster (unclear if NVIDIA, or Huawei).
A few quotes from the transcript:
> Our current computing capacity is approximately 20,000 H-equivalent units, most of which have just arrived within the past month or two
> Regarding the Huawei 950, Huawei currently provides us with 16,000 SIM cards
> A Huawei 950 [cluster] with 16,000 cards is equivalent to only a B-series card [cluster] with 4,000 cards.
From a superficial look I think the submitted title ("Moonshot built on 20k Nvidia chip cluster from Alibaba") might have been misleading. Article says "That hardware forms a substantial chunk of the overall computing capacity Moonshot uses for its Kimi models", implying that other hardware was also used.
(We've changed the title to what the article says now.)
It's not that surprising to me. Most of the innovation in Chinese models has been in efficiency gains and optimizations. K3 coming from the factory in MXFP4 weights is a pretty relevant factor. Big performance gap probably also due to Moonshot doing QAT. Throw in the fact that Musk has easier access to compute, and I think you have your answer on the disparity.
But perf improvements not only mean you can run the same thing on cheaper hardware, you can also run more capable models in the same hardware
In this sense, any advance in intelligence is a performance improvement and vice versa
Do we believe them? It seems there’s no reason with the resources OpenAI and Anthropic have they wouldn’t be incentivised to build equally efficient pipelines? I think it’s mostly the fine tuning with distillation that gives Chinese labs their big advantage. However I’m pretty sure all the main labs are poisoning the output now when they detect Chinese activity so while K3 looks frontier on tests when you use it it’s miles off Sol and Fable.
everything about the current 'a.i' models - east vs west points the wastefulness of western labs - whether in terms of gpu clusters being ran, cost of inference.
the only way is down for the massive valuations and 'a.i' revenue projections.
> west points the wastefulness of western labs whether in terms of gpu clusters being ran, cost of inference.
Can’t possibly be intentional media strategy by a geopolitical target, right? If its going to negatively impact valuations and revenue of major rivals, seems like a desirable strategy?
We just saw OpenAI take the steps to significantly lower the cost of one of their models, which confirms that at least one western lab has a large margin on inference, not a large inference cost.
Meanwhile, we don’t know how many experiments these east/west labs are performing relative to each other. We also know that many western labs have a whole portfolio of models too, which is product breadth not necessarily waste.
The wastefulness is akin to having multiple private jets on standby, or workers coming in to a fishbowl office to be gawked at, or massive datacenters to be used as social "proof".
Its all a fucking capitalistic farce to display to other rich elite that "Look at how much clout I have! I can make these peons dance around and do my bidding! Im a slave-owner!"
https://infosec.exchange/@david_chisnall/116991627711001827
He noted that that Silicon Valley doesnt really want to SOLVE problems. They want to find already-solved problems with problem matching. And of course, we just throw more people and more compute instead of optimization and understanding.
The Chinese are being actively constrained with bullshit politics around a second Red Scare moment. And, well, they're winning. A lot.
> Struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs).
This feels like a very American way of designing things - just throw more horse power at it, bigger is better! The rest of the world is usually a bit more resource constrained and efficient at using those resources.
See also Mustangs vs German sports cars, giant American fridges, giant American suburban McMansions vs livable cities etc etc.
One does wonder whether there has been an expiration of the actual weights of Opus at one point.
> Musk (who freely admits to distilling OpenAI's models) struggle to build a competitive 1T model (Grok 4.5) with massively more compute (Colossus-1 100-250K GPUs, Colossus-2 500K+ GPUs)
Are you sure they are using all of their compute on training? Didn't they rent out a ton to other AI companies?
The renting out is a pretty recent development.
Genuine question. Is there a time factor part of that equation?
It's obviously a question - did they just train for 10x as long due to having 10x fewer GPUs, but then that spoils the claim that they distilled Fable which was only recently introduced. Now doubt they did use some training data generated from older US models though.
This is just not how it works - it is perfectly plausible (in fact the most likely) that pretraining was well finished by the time Fable was released. A bit of extra distillation takes far less compute.
Moreover, it’s very plausible (and expected) to use multiple clusters and GPU types for RL rollouts which could very well not be included in this count.
No part of this pipeline is fixed in stone.
Relatedly I also wonder how much the orchestrated distillation bot nets that Anthropic uses against China are real. Not saying it does not happen but I have never seen dating showing how pervasive it is and I always imagine other western labs would absolutely be doing the same.
> Musk (who freely admits to distilling OpenAI's models)
Source?
https://www.forbes.com/sites/antoniopequenoiv/2026/04/30/elo...
Sounds fair, since Sam can run a for profit NGO
I’m still baffled that OpenAI can complain about this with a straight face while these models are literally trained on everything regardless of copyright or permission.
It kind of confirms a hypothesis I have that the next phase of AI development will be about getting smaller (in terms of model size and compute), because smaller is more capital efficient for training (allowing faster iteration and more iteration cycles for a given amount of capital), allows for denser inference (more inference for a given amount of compute hardware), and allows for more edge inference applications.
The goal will be to develop smaller models with more efficient architectures, that have similar or even better performance than larger models.
Yes, continual learning is currently a hot topic since this is what will allow an "AI intern or new employee" to learn on the job rather than be stuck on groundhog day, or pre-trained for every eventuality - needing to anticipate all the quirks and proprietary knowledge of every customer!
Continual learning tends to imply individualized models, else there is no data privacy (the secrets learned on the job at your company now being available to your competition), which really turns the current AI business model of a single centralized model served to everyone on it's head. If every customer has a different model that essentially means the end of batch processing with the same weights loaded into the GPU.
The direction this suggests is a move away from centrally served common models to locally served individual ones, which generally requires them to be smaller, even if some larger companies may be willing to invest in beefier hardware.
I think this is at least in part why the AI companies are trying NOT to implement true continual learning and see if they can instead finesse it by implementing continual compacted(?) memorization instead, since then it's "just" additional context that needs to be recalled and fed into every request, not weights that need feeding into the GPU. I don't think memorization is any substitute for learning, especially learning of practiced skills, but since it's far easier to implement, and non-disruptive to the cloud-based API business model, this is what we will see first.
The recent news of NVIDIA' investment in Sutskever's SSI has a tiny hint of this also, talking about SSI advising on NVIDIA's future architectural direction - apparently pushing it in a different direction than current (cloud-based, pre-trained) models. NVIDIA may be quite happy to see a move towards local models.
The middle ground would be continual learning for a group of users. For a company with a couple thousand employees, one model learning on all your employees might be viable. With some filtering for information that is supposed to stay compartmentalized (HR data, stuff under NDA, core technolgy). But this only really works if you bring enough volume to make it economical to effectively server you a whole other model
It feels like we're not really bottlenecked by model performance now anyway. Most cases I find where a model is being an idiot can be remedied by rethinking the context and tooling available to that model.
If they got a lot cheaper and only a little dumber, and we got a little smarter about how we use them, they could appear to the bystander as much more useful than they are.
It begs an interesting question as well: are frontier labs stuck in the incumbent phase of the "innovator's dilemma," where there's such intense pressure to maintain the flavor/character of their widely-loved existing models, that they dare not invest their resources into radical reimagination of their approach towards architectures for smaller models? A company like Moonshot does not have this cultural limitation.
The frontier labs would be well served in carving out 20k sub-clusters and giving research teams carte blanche in building things with radically different architectures - with full permission to distill whatever they want from the flagship models. We'd expect to see more product lines that feel "different" from the flagship models if this were already being done.
Eventually, multi-modality might even get offloaded as workflows, which might allow models to do this at a fraction of the data and compute required.
I am adding multi-modality to https://github.com/guilt/TinyToT, and I see that dis-aggregating capabilities, very similar to how our own sensory organs work, seems to be paying off quite well.
This is likely what Google has been doing, as it suites them best to have small fast models rather than hulking slow giants. Lower intelligence but way more ability to serve. Anthropic would probably need a datacenter the size of a small country to serve Fable on a Google search/services scale.
Getting smaller is how the last revolution in computing happened.
A VAX 11/780 was good, but an 80386 was a lot better, since the latter could run on 3 AA batteries and the former needed 6,000 watts of 3 phase.
Yes but that's because of the scaling laws for transistors. ML models seem to get better the bigger they are. If you want to compress the world's information, you need to look at all the information in the world.
My thesis is still that model building has no moat. Folks continue to migrate around between the big labs. There is a lot of value in having good taste around the harness and how the models are used. The medium to long term winners will be the folks that control the compute.
You can only control the compute if it is not commoditized, which at the current moment, it is not, leading to industry players having 75-80% margins.
Someone will find a way to make cheaper compute, and since nothing fundamental changed there (LLM didn't change how silicon were made), that's bound to happen.
https://archive.is/vgmMB
Thank you
If Chinese companies are able to just reliably rent the damn GPUs then I really don't see how these restrictions are meant to be effective in the slightest. I guess it makes it less convenient and supply less certain, and retains some optionality for cutting off supply later?
If Alibaba can't bring the chips into China, but can buy a whole bunch in Thailand or Singapore or whereever (or secure exclusive rights via JV partners where relevant) and then just provide them as a service to its customers in China - what is the point? I'm sure many customers would actually prefer such an arrangement.
"what is the point?" There is no point, just zero-sum thinking which isn't compatible with reality.
I read this as if they purchased it from Alibaba the marketplace.....
"2.8TB model, fast shipping!"
I would not minimize the extent to which AI progress depends on things like the effort that goes into training material acquisition and data labeling. In those areas, improvement is a direct function of how much you invest.
I wonder how much spending is motivated by the phenomenon of sudden emergent performance in LLMs. Clearly some people who are smarter than me expect something like emergent AGI, or at least they think the odds justify spending whatever it takes to see if that would happen.
That leaves a lot of room for efficient aggressive followers.
K3 is natively trained to mxfp4, if they cannot get a hold of Blackwell chip, it is meaningless. Hopper does not do native 4-bit floating math.
Either they have Blackwell with native 4-bit floating math, or they use have Chinese domestic NPU that support mxfp4 natively.
The article’s statement does not make sense.
Huawei A910 is supposed to have native support for MXFP4. The problem is that huawei doesn't have the capacity to build more GPUs right now. Those yield numbers must be really terrible.
Personally, I think 20K nvidias is a stop gap solution because they really don't have the capacity to serve their models to earn any money right now.
> Either they have Blackwell
From the article:
A person familiar with Moonshot’s procurement strategy confirmed that the company does indeed have a channel for accessing Blackwell processors via Southeast Asia. They didn’t specify whether this was a rental channel...
What do you mean by "it's meaningless" can you expand?
They mean that K3's native floating point is MXFP4, a newer standard that's not supported in Hopper, an older nvidia chipset, so it's impossible the GPUs in question were Hopper (the article specifically mentions H200s), which is correct.
Probably either the author or the people they interviewed got the specific GPU details mixed up or wrong. Maybe it was B200s.
Oh, okay, thank you for the explanation. I was under the impression that we, the United States, were banning exports of such GPUs.
https://archive.md/vgmMB
It's also built on the back of downloading all the content of the top ten sites on the list, the fact that all the independent content sites with their "pre-war" steel haven't been acquired at top dollar by U.S. companies shows the existential threat / stop China posturing is a total joke.
https://archive.is/AvWWX
No one is using that dataset for any serious models. It’s way too small and old.