I just read a claim on Twitter that the reason these companies (Google and Amazon as well as OpenAI) are using Broadcom isn't just for design expertise, but because Broadcom have allocation agreements in place with TSMC and the memory manufacturers.
alephnerd 28 minutes ago [-]
Most design partners have allocation agreements. The thing is Broadcom is an absolute GIANT in the ASIC design space, and it's closest competitor Marvell is a fraction of it's size.
There are a lot of large tech companies that most of HN has never heard about that completely dominate entire segments.
a_conservative 1 hours ago [-]
I recently put 2+2 together.
Broadcom has become wealthy by being Google's TPU hardware partner, including sharing their TSMC capacity with Google, and evidently now they are doing the same thing with OpenAI. What a brilliant way to take advantage of the AI gold rush!
I wish they weren't using their piles of money to extort money out of the software industry like they are with VMWare and Bitnami.
alephnerd 41 minutes ago [-]
> Broadcom has become wealthy by being Google's TPU hardware partner...
Kinda, but not exactly.
Broadcom cornered the enterprise infra and security market in the late 2010s after acquiring CA Technologies, BMC, Symantec, and VMWare and were able to make a strong cybersecurity story during the late 2010s cybersecurity and SaaS boom.
That gave them plenty of cashflow that helped subsidize their hardware business when hardware was not viewed as hot as it is today.
Additionally, Broadcom is GCP's marquee customer and has been for a little under a decade so they were able to make a sweetheart deal where all that software businesses at Broadcom would be exclusively using GCP and in return GCP would working with Broadcom to design it's silicon and source infra needed for their DC buildouts.
Ironically, the DoJ blocking Broadcom's acquisition of Qualcomm was the best thing it ever could have done for Broadcom, because it gave Broadcom the dry powder to dominate the Enterprise SaaS and build a strong niche in the cybersecurity space.
> piles of money to extort money out of the software industry
From personal experience, executives and leadership who started off in the electronics and hardware industry are much more vicious and cutthroat than their peers who started in software.
Working in an industry that historically had to deal with high commodification, low margins, and long tail development leads to leadership that can execute. Additionally, no one climbs the leadership ladder without having spent years as a line-level engineer.
maz1b 2 hours ago [-]
Pretty huge move. Google and their TPUs are looking infinitely more prescient as I think they are on their 7th generation, along with the offshoots it inspired like the LPU and even others, perhaps like Cerebras and their Wafer Scale Engine.
However, based off first impressions, it seems like this is meant for inference side, and not training, which is also an interesting choice.
skeledrew 56 minutes ago [-]
Training is pretty much a 1x cost, and efficiency there is already on the way down with architectural improvements. Inference though is an ongoing cost which over time takes orders of magnitude more resources, so focusing on making that far more efficient means way greater gains over time.
zer00eyz 55 minutes ago [-]
> early testing shows that Jalapeño will deliver performance per watt substantially better than current state-of-the-art
We're starting to see what really matters here, and though this is hand wavy the TPU makes similar claims.
I think googles memo about having no moat still stands (see: https://newsletter.semianalysis.com/p/google-we-have-no-moat... if you are unaware). It kind of makes sense that all of this is looking more like 60's to 90's IBM, DEC, Cray, Sun and the hardware race that happened then. History doesn't repeat but it often rhymes and I suspect that these efforts will follow the same trajectory.
forrestthewoods 1 hours ago [-]
Inference costs are higher than training now. I think.
Nvidia is king of general purpose training chips. But inferences can be specialized.
v5v3 60 minutes ago [-]
>designed for initial deployment by the end of 2026 and expanding in the years ahead,
So after the IPO and will be featured heavily in the IPO sales brochure as a future promise?
I'm sceptical over any pre-IPO announcements.
kilroy123 3 hours ago [-]
I hope to see something like this, but in a small form factor like the NVIDIA spark.
I want a super fast LLM that is Opus 4.6+, like, in ability.
smith7018 6 minutes ago [-]
Unfortunately Sam Altman won't be the one to deliver us at-home hardware that can run Opus-level models
wmf 38 minutes ago [-]
Memory bandwidth is the bottleneck in the Spark. If you replace the SoC with an optimized ASIC but keep the same 256-bit LPDDR5 the performance will be the same. You can increase performance by using wider memory but that's also more expensive.
theowaway213456 43 minutes ago [-]
This seems like more competition for Cerebras? Am I understanding correctly?
HarHarVeryFunny 10 minutes ago [-]
This is just an uncut wafer - I don't think it's intended to be wafer-scale chip.
Cerebras etch memory onto the wafer alongside the processing elements, but AFAIK OpenAI are going to be using HBM memory and a conventional chiplet design.
38 minutes ago [-]
dadoum 2 hours ago [-]
> May we scale smoothly, exponentially and uneventfully through A[SI]
That sentence sounds weird to me. I can't really put my finger on why, maybe the combination of adverbs, or just the fact of writing the desire of scaling as a company so directly. It feels (to me) like openly claiming their selfish goals. Or maybe I am just misinterpreting and they are referring to the whole humanity as "We" (but knowing Broadcom and in a lesser extent OpenAI doings, I am not convinced).
satvikpendem 1 hours ago [-]
I'm assuming they used LLMs to (help humans) do custom circuit design. Even pre LLM there were various computer optimizations that didn't require humans like genetic algorithms. It'd be cool to see a paper on how they did it.
Legend2440 1 hours ago [-]
The only surprising thing about this is that they didn't do it three years ago.
gravypod 1 hours ago [-]
I wonder how close OpenAI is getting to using the memory they purchased. Are they planning to stack a huge amount of HBM2 into these chips?
wmf 26 minutes ago [-]
I assume OpenAI has been buying memory and "giving" it to Nvidia in exchange for a discount.
43 minutes ago [-]
qsxfthnkp2322 2 hours ago [-]
aw shucks nvda has some spicy competition
Make sure you all use that fancy ñ
boarush 1 hours ago [-]
They don't have true competition, what they lose out on is market share with hyperscalers, since OpenAI would have no plans to share inference hardware with any other company right now. Plus, I don't know how does NVIDIA's investment equation pans out long terms given OpenAI will be investing in more purpose built inference stack for the future.
fennecbutt 51 minutes ago [-]
I mean I'd love to be able to buy something like the 17k tps taalas chip as a pcie or m.2.
Imagine when we can roar along at that speed, low power. Can just have the model reason for a while about anything and everything. It reminds me of the "race to idle" for mcus etc.
ipdashc 28 minutes ago [-]
> 17k tps taalas chip
It's odd to me that I haven't heard anything about this approach (baking LLMs/weights into silicon directly) since. It seems almost common-sense that we're going to end up there eventually. And it feels like that point is drawing ever closer now that model capabilities, if not quite plateauing out, are at least getting to a "good enough" point for a LOT of use cases.
I wonder if it's being worked on in secret, if there's something about it that makes it infeasible, or if companies are really too nervous to lock in one model like that because the next one down the line could be a huge improvement. Re. infeasability, I have heard that the Taalas demonstration chip ran Llama 3.1 8B (a pretty horrible model) and that even that took a massive amount of transistors / die area. So it might just be the case that the good models are too big to fit on silicon?
wmf 23 minutes ago [-]
Good models will require multiple Taalas chips but Groq and Cerebras also require a lot of chips and that hasn't stopped them.
MichaelNolan 29 minutes ago [-]
The current taalas chip is for a 3.1B param model. I’m hope so much that they can get that up to the 30B range. Just imagine Gemma 4 or Qwen 3.6 at 17k tps.
jabedude 1 hours ago [-]
how much does this chip help with inference speed?
wmf 37 minutes ago [-]
It's probably the same speed but cheaper.
fibonacci112358 3 hours ago [-]
So this is where all the memory they bought is going to.
babelfish 1 hours ago [-]
that's not really how it works
jerojero 2 hours ago [-]
One thing I don't like about California based companies is how cringe the names always are.
"Jalapeño" is such a bad name, having an "ñ" already makes it difficult and annoying to deal with in so many little ways. Good luck with that.
But also, theres the sort of "yes lets use Mexican related things because we're California" thought that I just really hate. I don't know, its like corporate Memphis to me. You see a product like this, you know it's an uppity califonia based firm that came up with it.
thewebguyd 1 hours ago [-]
No worse, I suppose, than, the obsession with Lord of the Rings that the authoritarian surveillance companies have. Palantir, Anduril. Then we have the not defense/surveillance ones: Mithril, Valar, Narya, Erebor
skeledrew 55 minutes ago [-]
What kinds of names would you suggest?
thewebguyd 50 minutes ago [-]
None, probably. Just saying Jalapeño is no worse than any other non-descriptive company name. Although at least Palantir and Anduril are aptly named for what they do. The VC firms less so.
utopiah 41 minutes ago [-]
Strawberry was too complicated as a codename.
qsxfthnkp2322 2 hours ago [-]
Jalapeño
Jalapeño
Jalapeño
Really has a… ring to it
Rendered at 17:04:05 GMT+0000 (Coordinated Universal Time) with Vercel.
1. https://www.investing.com/news/stock-market-news/openai-unve...
There are a lot of large tech companies that most of HN has never heard about that completely dominate entire segments.
Broadcom has become wealthy by being Google's TPU hardware partner, including sharing their TSMC capacity with Google, and evidently now they are doing the same thing with OpenAI. What a brilliant way to take advantage of the AI gold rush!
I wish they weren't using their piles of money to extort money out of the software industry like they are with VMWare and Bitnami.
Kinda, but not exactly.
Broadcom cornered the enterprise infra and security market in the late 2010s after acquiring CA Technologies, BMC, Symantec, and VMWare and were able to make a strong cybersecurity story during the late 2010s cybersecurity and SaaS boom.
That gave them plenty of cashflow that helped subsidize their hardware business when hardware was not viewed as hot as it is today.
Additionally, Broadcom is GCP's marquee customer and has been for a little under a decade so they were able to make a sweetheart deal where all that software businesses at Broadcom would be exclusively using GCP and in return GCP would working with Broadcom to design it's silicon and source infra needed for their DC buildouts.
Ironically, the DoJ blocking Broadcom's acquisition of Qualcomm was the best thing it ever could have done for Broadcom, because it gave Broadcom the dry powder to dominate the Enterprise SaaS and build a strong niche in the cybersecurity space.
> piles of money to extort money out of the software industry
From personal experience, executives and leadership who started off in the electronics and hardware industry are much more vicious and cutthroat than their peers who started in software.
Working in an industry that historically had to deal with high commodification, low margins, and long tail development leads to leadership that can execute. Additionally, no one climbs the leadership ladder without having spent years as a line-level engineer.
However, based off first impressions, it seems like this is meant for inference side, and not training, which is also an interesting choice.
We're starting to see what really matters here, and though this is hand wavy the TPU makes similar claims.
I think googles memo about having no moat still stands (see: https://newsletter.semianalysis.com/p/google-we-have-no-moat... if you are unaware). It kind of makes sense that all of this is looking more like 60's to 90's IBM, DEC, Cray, Sun and the hardware race that happened then. History doesn't repeat but it often rhymes and I suspect that these efforts will follow the same trajectory.
Nvidia is king of general purpose training chips. But inferences can be specialized.
So after the IPO and will be featured heavily in the IPO sales brochure as a future promise?
I'm sceptical over any pre-IPO announcements.
I want a super fast LLM that is Opus 4.6+, like, in ability.
Cerebras etch memory onto the wafer alongside the processing elements, but AFAIK OpenAI are going to be using HBM memory and a conventional chiplet design.
That sentence sounds weird to me. I can't really put my finger on why, maybe the combination of adverbs, or just the fact of writing the desire of scaling as a company so directly. It feels (to me) like openly claiming their selfish goals. Or maybe I am just misinterpreting and they are referring to the whole humanity as "We" (but knowing Broadcom and in a lesser extent OpenAI doings, I am not convinced).
Make sure you all use that fancy ñ
Imagine when we can roar along at that speed, low power. Can just have the model reason for a while about anything and everything. It reminds me of the "race to idle" for mcus etc.
It's odd to me that I haven't heard anything about this approach (baking LLMs/weights into silicon directly) since. It seems almost common-sense that we're going to end up there eventually. And it feels like that point is drawing ever closer now that model capabilities, if not quite plateauing out, are at least getting to a "good enough" point for a LOT of use cases.
I wonder if it's being worked on in secret, if there's something about it that makes it infeasible, or if companies are really too nervous to lock in one model like that because the next one down the line could be a huge improvement. Re. infeasability, I have heard that the Taalas demonstration chip ran Llama 3.1 8B (a pretty horrible model) and that even that took a massive amount of transistors / die area. So it might just be the case that the good models are too big to fit on silicon?
"Jalapeño" is such a bad name, having an "ñ" already makes it difficult and annoying to deal with in so many little ways. Good luck with that.
But also, theres the sort of "yes lets use Mexican related things because we're California" thought that I just really hate. I don't know, its like corporate Memphis to me. You see a product like this, you know it's an uppity califonia based firm that came up with it.
Jalapeño
Jalapeño
Really has a… ring to it