[HN Gopher] Microsoft CTO says he wants to swap most AMD and Nvi...
       ___________________________________________________________________
        
       Microsoft CTO says he wants to swap most AMD and Nvidia GPUs for
       homemade chips
        
       Author : fork-bomber
       Score  : 168 points
       Date   : 2025-10-03 14:48 UTC (8 hours ago)
        
 (HTM) web link (www.cnbc.com)
 (TXT) w3m dump (www.cnbc.com)
        
       | okokwhatever wrote:
       | somebody wants to buy NVDA cheaper... ;)
        
         | synergy20 wrote:
         | plus someone has no leverage whatsoever other than talking
        
           | ambicapter wrote:
           | Microsoft, famously resource-poor.
        
             | fidotron wrote:
             | They have so much money it is harmful to their ability to
             | execute.
             | 
             | Just look at the implosion of the XBox business.
        
               | fennecbutt wrote:
               | Granted, if everyone had done what the highly paid
               | executives had told them to do, xbox would never have
               | existed.
               | 
               | And I'm guessing that the decline is due to executive
               | meddling.
               | 
               | What is it that executives do again? Beyond collecting
               | many millions of dollars a year, that is.
        
               | fidotron wrote:
               | They sit around fantasizing about buying Nintendo because
               | that would be the top thing they could achieve in their
               | careers.
        
       | rjbwork wrote:
       | Guess MSFT needs somewhere else AI adjacent to funnel money into
       | to produce the illusion of growth and future cash flow in this
       | bubblified environment.
        
         | balls187 wrote:
         | > produce the illusion of growth and future cash flow in this
         | bubblified environment.
         | 
         | I was ranting about this to my friends; Wallstreet is now
         | banking on Tech firms to produce the illusion of growth and
         | returns, rather than repackaging and selling subprime
         | mortgages.
         | 
         | The tech sector seems to have a never ending supply of things
         | to spur investment and growth: cloud computing, saas, mobile,
         | social media, IoT, crypto, Metaverse, and now AI.
         | 
         | Some useful, some not so much.
         | 
         | Tech firms have a lot of pressure to produce growth, it's
         | filled with very smart people, and wields influence on public
         | policy. The flip side is the mortage crisis, at least before it
         | collapsed, got more Americans into home ownership (even if they
         | weren't ready for it). I'm not sure the tech sectors meteoric
         | rise has been as helpful (sentiment of locals in US tech hubs
         | suggests a overall feeling of dissatisfaction with tech)
        
       | giancarlostoro wrote:
       | So similar to Apple Silicon. If this means they'll be on par with
       | Apple Silicon I'm okay with this, I'm surprised they didn't do
       | this sooner for their Surface devices.
       | 
       | Oh right, for their data centers. I could see this being useful
       | there too, brings costs down lower.
        
         | CharlesW wrote:
         | > _So similar to Apple Silicon._
         | 
         | Yes, in the sense that this is at least partially inspired by
         | Apple's vertical integration playbook, which has now been
         | extended to their own data centers based on custom Apple
         | Silicon1 and a built-for-purpose, hardened edition of Darwin2.
         | 
         | 1 https://security.apple.com/blog/private-cloud-compute/ 2
         | https://en.wikipedia.org/wiki/Darwin_(operating_system)
        
           | giancarlostoro wrote:
           | Yeah, its interesting, years ago I never thought Apple nor
           | Microsoft would do this, but also Google has done this on
           | their cloud as well, so it makes sense.
        
           | georgeburdell wrote:
           | Vertical integration only works if your internal teams can
           | stay in the race at each level well enough to keep the stack
           | competitive as a whole. Microsoft can't attract the same
           | level of talent as Apple because their pay is close to the
           | industry median
        
       | quadrature wrote:
       | Not suprising that the hyperscalers will make this decision for
       | inference and maybe even a large chunk of training. I wonder if
       | it will spur nvidia to work on an inference only accelerator.
        
         | edude03 wrote:
         | > I wonder if it will spur nvidia to work on an inference only
         | accelerator.
         | 
         | Arguably that's a GPU? Other than (currently) exotic ways to
         | run LLMs like photonics or giant SRAM tiles there isn't a
         | device that's better at inference than GPUs and they have the
         | benefit that they can be used for training as well. You need
         | the same amount of memory and the same ability to do math as
         | fast as possible whether its inference or training.
        
           | CharlesW wrote:
           | > _Arguably that 's a GPU?_
           | 
           | Yes, and to _@quadrature_ 's point, NVIDIA _is_ creating GPUs
           | explicitly focused on inference, like the Rubin CPX:
           | https://www.tomshardware.com/pc-components/gpus/nvidias-
           | new-...
           | 
           |  _" ...the company announced its approach to solving that
           | problem with its Rubin CPX-- Content Phase aXcelerator --
           | that will sit next to Rubin GPUs and Vera CPUs to accelerate
           | specific workloads."_
        
             | edude03 wrote:
             | Yeah, I'm probably splitting hairs here but as far as I
             | understand (and honestly maybe I don't understand) - Rubin
             | CPX is "just" a normal GPU with GDDR instead of HBM.
             | 
             | In fact - I'd say we're looking at this backwards - GPUs
             | used to be the thing that did math fast and put the result
             | into a buffer where something else could draw it to a
             | screen. Now a "GPU" is still a thing that does math fast,
             | but now sometimes, you don't include the hardware to put
             | the pixels on a screen.
             | 
             | So maybe - CPX is "just" a GPU but with more generic naming
             | that aligns with its use cases.
        
               | bonestamp2 wrote:
               | There are some inference chips that are fundamentally
               | different from GPUs. For example, one of the guys who
               | designed Google's original TPU left and started a company
               | (with some other engineers) called groq ai (not to be
               | confused with grok ai). They make a chip that is quite
               | different from a GPU and provides several advantages for
               | inference over traditional GPUs:
               | 
               | https://www.cdotrends.com/story/3823/groq-ai-chip-
               | delivers-b...
        
           | AzN1337c0d3r wrote:
           | I would submit Google's TPUs are not GPUs.
           | 
           | Similarly, Tenstorrent seems to be building something that
           | you could consider "better", at least insofar that the goal
           | is to be open.
        
           | quadrature wrote:
           | I'm not very well versed, but i believe that training
           | requires more memory to store intermediate computations so
           | that you can calculate gradients for each layer.
        
           | conradev wrote:
           | They're already optimizing GPU die area for LLM inference
           | over other pursuits: the FP64 units in the latest Blackwell
           | GPUs were greatly reduced and FP4 was added
        
           | imtringued wrote:
           | The AMD NPU has more than 2x the performance per watt versus
           | basically any Nvidia GPU. Nvidia isn't leading because they
           | are power efficient.
           | 
           | And no, the NPU isn't a GPU.
        
           | nsteel wrote:
           | Isn't Etched's Soho ASIC claimed to be much better than a
           | GPU?
           | 
           | https://www.etched.com/announcing-etched
        
       | hkt wrote:
       | Even just saying this applies downward pressure on pricing:
       | NVIDIA has an enormous amount of market power (~"excess" profit)
       | right now and there aren't enough near competitors to drive that
       | down. The only thing that will work is their biggest _consumers_
       | investing, or threatening to invest, if their prices are too
       | high.
       | 
       | Long term, I wonder if we're exiting the "platform compute" era,
       | for want of a better term. By that I mean compute which can run
       | more or less any operating system, software, etc. If everyone is
       | siloed into their own vertically integrated hardware+operating
       | system stack, the results will be awful for free software.
        
         | startupsfail wrote:
         | In that case, it's great that Microsoft is building their
         | silicon. Keeps NVIDIA in check, otherwise these profits would
         | evaporate into nonsense and NVIDIA would lose the AI industry
         | to competition from China. Which, depending if AGI/ASI is
         | possible or not, may or may not be a great move.
        
       | harrall wrote:
       | Google has been using its own TPU silicon for machine learning
       | since 2015.
       | 
       | I think they do all deep learning for Gemini on ther own silicon.
       | 
       | But they also invented AI as we know it when they introduced
       | transformer architecture and they've been more invested in
       | machine learning than most companies for a very long time.
        
         | chrismustcode wrote:
         | I thought they use GPU for learning and TPU for inference, I'm
         | open to been corrected.
        
           | cendyne wrote:
           | I've heard its a mixture because they can't source enough in-
           | house compute
        
           | xnx wrote:
           | Some details here:
           | https://news.ycombinator.com/item?id=42392310
        
           | surajrmal wrote:
           | The first tpu they made was inference only. Everything since
           | has been used for training. I think that means they weren't
           | using it for training in 2015 but rather 2017 based on
           | Wikipedia.
        
             | lokar wrote:
             | The first TPU they *announced" was for inference
        
           | dekhn wrote:
           | no. for internal training most work is done on TPUs, which
           | have been explicitly designed for high performance training.
        
         | buildbot wrote:
         | Not that it matters, but Microsoft has been doing AI
         | accelerators for a bit too - project Brainwave has been around
         | since 2018 - https://blogs.microsoft.com/ai/build-2018-project-
         | brainwave/
        
           | interestpiqued wrote:
           | Yeah I worked in the hardware org around this time. We got
           | moved from under Xbox org to azure and our main work became
           | AI related accelerators
        
             | buildbot wrote:
             | Very cool! Catapult/Brainwave is what got me into hardware
             | & ML stuff :)
        
         | rjzzleep wrote:
         | The first revisions were stuff made by qualcomm right? I don't
         | think we have much data on how much customizations they make
         | and where they their IP from, but given how much of the Tensor
         | cores comes from Samsung I think it's safe to say to assume
         | that there is a decent amount coming from some of the big
         | vendors.
        
           | Moto7451 wrote:
           | To be fair that's a pretty good approach if you look at
           | Apple's progression from assembled IPs in the first iPhone
           | CPU to the A and M series.
        
             | rjzzleep wrote:
             | Yeah but at least when it comes to mobile CPUs Apple seemed
             | vastly more competent in how they approached it.
        
             | osivertsson wrote:
             | Apple generally tries to erase info about acquisitions from
             | their official company story, they want it to look like
             | internal Apple innovation.
             | 
             | When it comes to CPUs they bought P.A. Semi back in 2008
             | and got a lot of smart people with decades of relevant
             | experience that were doing cutting-edge stuff at the time.
             | 
             | This was immensely important to be able to deliver current
             | Apple CPUs.
        
               | foobiekr wrote:
               | PA Semi was kind of an interesting acquisition. The team
               | was definitely very skilled, but there are always
               | gotchas. Before the acquisition happened we were on their
               | receiving end of their dual core PPC and it was not great
               | at all. We had a lot of issues with board bringup, power,
               | and heat. More errata than I've ever seen. We eventually
               | had to went with x86 for the project instead, which was
               | more performant and certainly a lot easier overall at the
               | time.
               | 
               | I had previously encountered some of that team with the
               | SiByte MIPS in an embedded context, I know they were
               | highly skilled, they had tons of pedigree, but PA Semi
               | itself was a strange beast.
        
           | conradev wrote:
           | For TPUs I believe it is Broadcom:
           | https://www.theregister.com/2023/09/22/google_broadcom_tpus/
           | 
           | Not sure about the mobile SoCs
        
             | foobiekr wrote:
             | Broadcom fills the same role for Google TPU that Marvell
             | fills for Trainium @ amazon.
        
           | kccqzy wrote:
           | It's "made by" TSMC as usual. Their customization comes from
           | identifying which compute operations that want optimized in
           | hardware and do it themselves. And then they buy non-compute
           | IP like HBM from Broadcom. And Broadcom also does things like
           | physical design.
        
         | Der_Einzige wrote:
         | I'm 99.999% sure that the claim of "all deep learning for
         | Gemini on their own silicon" is not true.
         | 
         | Maybe if you restrict it similarly to the Deepseek paper to
         | "Gemini uses TPU for the final successful training run and for
         | scaled inference" you might be correct, but there's no way that
         | GPUs aren't involved for at minimum comparability and more
         | rapid iteration reasons during the extremely buggy and error
         | prone point of getting to the final training run. Certainly the
         | theoretical and algorithmic innovations that are often being
         | done at Google and do make their way into Gemini also sometimes
         | using Nvidia GPUs.
         | 
         | GCP has a lot of, likely on the order of at least 1 million
         | GPUs in their fleet today (I'm likely underestimating). Some of
         | that is used internally and is made available to their
         | engineering staff. What constitutes "deep learning for gemini"
         | is very up to interpretation.
        
           | imtringued wrote:
           | Google is the creator of JAX and XLA. Maybe the developer
           | laptops have Nvidia GPUs and they do some testing there, but
           | for Google there is literally no point in bothering with
           | CUDA, pytorch or any other ecosystem strongly focused on
           | Nvidia GPUs.
           | 
           | In my experience JAX is way more flexible than pytorch the
           | moment you want to do things that aren't training ML models.
           | E.g. you want to build an optimizer that uses the derivative
           | of your model with respect to the input.
        
             | kingstnap wrote:
             | Honestly Pytorch is weird imo, I'm surprised people love it
             | so much.
             | 
             | Loss.backward? Tensor.grad? Optimizer.zero grad()? With
             | torch.no_grad()?
             | 
             | What is with all these objects holding pointers to stuff?
             | An ndarray is a pointer to memory and a shape my dudes. A
             | gradient is the change in a scalar function w.r.t to some
             | inputs.
        
           | kccqzy wrote:
           | You seem to think GPUs are better than TPUs for rapid
           | iteration. Why is that? There's no inherent reason why one is
           | more suited to rapid iteration than another; it's entirely a
           | matter of developer tooling and infrastructure. And Google
           | famously has excellent tooling. And furthermore, the tooling
           | Google exposes to the outside world is usually poorer than
           | the tooling used internally by Googlers.
        
             | bjourne wrote:
             | It was a few years ago since I last played with Googles hw,
             | but iirc TPUs were inflexible and very fast. Worked well
             | for linear and convolutional layers but could not
             | accelerate certain LSTM configurations. For such networks
             | GPUs were faster. It wouldn't surprise me the least if TPU
             | hardware support lagged behind what the latest and greatest
             | LLMs require for training.
        
           | summerlight wrote:
           | Internally, TPU is much cheaper for the same amount of
           | compute compared to GPU, so I don't see much reasons why they
           | need to use GPU. Probably >99% of compute budgets are spent
           | on TPU. It might be true if you say these <1% still counts,
           | but I guess it is pretty safe to say all of its meaningful
           | production workload are running on TPU. It is simply too
           | expensive to run a meaningful amount of compute on non-TPU.
           | 
           | Just to clarify, TPU has been in development for a decade and
           | it is quite mature these days. Years ago internal consumers
           | had to accept the CPU/GPU and TPU duality but I think this
           | case is getting rarer. I guess this is even more true for
           | DeepMind since itself owns a ML infra team. They likely be
           | able to fix most of the issues with a high priority.
        
           | owkman wrote:
           | That's a strange position to take with such high certainty.
           | Google has been talking about training on TPUs for a long
           | time. Many ex and current employees have been on the record
           | talking about how much nicer the Google internal training
           | infra using TPUs is. GPU is an afterthought for Google's
           | inference and non-existent in training.
        
       | leshokunin wrote:
       | Honestly this would be great for competition. Would love to see
       | them impish in that direction.
        
       | bee_rider wrote:
       | The most important note is:
       | 
       | > The software titan is rather late to the custom silicon party.
       | While Amazon and Google have been building custom CPUs and AI
       | accelerators for years, Microsoft only revealed its Maia AI
       | accelerators in late 2023.
       | 
       | They are too late for now, they realistically hardware takes a
       | couple generations to become a serious contender and by the time
       | Microsoft has a chance to learn from their hardware mistakes the
       | "AI" bubble will have popped.
       | 
       | But, there will probably be some little LLM tools that do end up
       | having practical value; maybe there will be a happy line-crossing
       | point for MS and they'll have cheap in-house compute when the
       | models actually need to be able to turn a profit.
        
         | surajrmal wrote:
         | At this point it will take a lot of investment to catch up.
         | Google relies heavily on specialized interconnects to build
         | massive tpu clusters. It's more than just designing a chip
         | these days. Folks who work on interconnects are a lot more rare
         | than engineers who can design chips.
        
         | alephnerd wrote:
         | > hardware takes a couple generations to become a serious
         | contender
         | 
         | Not really and for the same reason Chinese players like Biren
         | are leapfrogging - much of the workload profile in AI/ML is
         | "embarrassingly parallel", thus reducing the need for
         | individual ASICs to be bleeding edge performant.
         | 
         | If you are able to negotiate competitive fabrication and energy
         | supply deals, you can mass produce your way into providing
         | "good enough" performance.
         | 
         | Finally, the persona who cares about hardware performance in
         | training isn't in the market for cloud offered services.
        
           | kenjackson wrote:
           | And current LLM architectures affinitize differently to HW
           | than DNNs even just a decade ago. If you have the money and
           | technical expertise (both of which I assume MS has access to)
           | then a late start might actually be beneficial.
        
           | rcxdude wrote:
           | As I understood it the main bottleneck is interconnects,
           | anyhow. It's more difficult to keep the ALUs fed than it is
           | to make them fast enough, especially once your model can't
           | fit in one die/PCB. And that's in principle a much trickier
           | part of the design, so I don't really know how that shakes
           | out (is there a good enough design that you can just buy as a
           | block?)
        
             | alephnerd wrote:
             | The bet I'm seeing is to try and invest in custom ASICs to
             | become integrated as part of an SoC to solve that
             | interconnect bottleneck.
             | 
             | It's largely a solved problem based on Google/Broadcom's
             | TPU work - almost everyone is working with Broadcom to
             | design their own custom ASIC and SoC.
        
         | interestpiqued wrote:
         | Most of the big players started working on hardware for this
         | stuff in 2018/2019. I worked at MSFT silicon org during this
         | time. Meta was also hiring my coworkers for similar projects. I
         | left a few years ago and don't know current state but they
         | already have some generations under their belt
        
       | cjbgkagh wrote:
       | I guess Microsoft's investment into Graphcore didn't pay off. Not
       | sure what they're planning but more of that isn't going to cut
       | it. At the time (late 2019) I was arguing for either a GPU
       | approach or specialized architecture targeting transformers.
       | 
       | There was a split at MS where the 'Next Gen' bayesian was being
       | done in the US and the frequentist work was being shipped off to
       | China. Chris Bishop was promoted to head of MSR Cambridge which
       | didn't help.
       | 
       | Microsoft really is an institutionally stupid organization so I
       | have no idea on which direction they actually go. My best guess
       | is that it's all talk.
        
         | Den_VR wrote:
         | Microsoft lacks the credibility and track record for this to be
         | anything but talk. Hardware doesn't simply go from zero to
         | gigawatts of infrastructure on talk. Even Apple is better
         | positioned for such a thing.
        
           | pjmlp wrote:
           | Microsoft has plenty of home grown hardware on Azure, some of
           | which even has firmware written in Rust nowadays.
        
             | buildbot wrote:
             | Yep, tons of decent homegrown stuff in Azure. And for a
             | long time - see the Catapult FPGA cards.
        
         | migueldeicaza wrote:
         | They do have such a dedicated chip, the MAIA 100 chip which is
         | an in-house chip, and it is a chip that was designed in the era
         | of transformers, and this is what is being discussed in the
         | interview.
        
           | cjbgkagh wrote:
           | I missed that, it's been a few years since I've paid
           | attention to MS hardware and it is very possible that my
           | thoughts are out of date. I left MS with a rather bad taste
           | in my mouth. I'm checking out the info on that chip and what
           | I am seeing is a little light on details. Just TPUs and fast
           | interconnects.
           | 
           | What I've found; MIAI 200 the next version is having issues
           | due to brain drain, and MIAI 300 is to be an entirely new
           | architecture so the status for that is rather uncertain.
           | 
           | I think a big reason MS invested so heavily into OpenAI was
           | to have a marquee customer push cultural change through the
           | org, which was a necessary decision. If that eventually
           | yields in a useful chip I will be impressed, I hope it does.
        
       | latchkey wrote:
       | It always falls back on the software. AMD is behind, not because
       | the hardware is bad, but because their software historically has
       | played second fiddle to their hardware. The CUDA moat is real.
       | 
       | So, unless they also solve that issue with their own hardware,
       | then it will be like the TPU, which is limited to usage primarily
       | at Google, or within very specific use cases.
       | 
       | There are only so many super talented software engineers to go
       | around. If you're going to become an expert in something, you're
       | going to pick what everyone else is using first.
        
         | amelius wrote:
         | > The CUDA moat is real.
         | 
         | I don't know. The transformer architecture uses only a limited
         | number of primitives. Once you have ported those to your new
         | architecture, you're good to go.
         | 
         | Also, Google has been using TPUs for a long time now, and
         | __they__ never hit a brick wall for a lack of CUDA.
        
           | latchkey wrote:
           | It is beyond porting, it is mentality of developers. Change
           | is expensive and I'm not just talking about $ value.
           | 
           | > Also, Google has been using TPUs for a long time now, and
           | __they__ never hit a brick wall for a lack of CUDA.
           | 
           | That's exactly what I'm saying. __they__ is the keyword.
        
             | amelius wrote:
             | Not sure what you mean. Google is a big company. Their TPUs
             | have many users internally.
        
               | latchkey wrote:
               | Very few developers outside of Google have ever written
               | code for a TPU. In a similar way, far fewer have written
               | code for AMD, compared to NVIDIA.
               | 
               | If you're going to design a custom chip and deploy it in
               | your data centers, you're also committing to hiring and
               | training developers to build for it.
               | 
               | That's a kind of moat, but with private chips. While you
               | solve one problem (getting the compute you want), you
               | create another: supporting and maintaining that ecosystem
               | long term.
               | 
               | NVIDIA was successful because they got their hardware
               | into developers hands, which created a feedback loop,
               | developers asked for fixes/features, NVIDIA built them,
               | the software stack improved, and the hardware evolved
               | alongside it. That developer flywheel is what made CUDA
               | dominant and is extremely hard to replicate because the
               | shortage of talented developers is real.
        
               | amelius wrote:
               | I mean it's all true to some extent. But that doesn't
               | mean implementing the few primitives to get transformers
               | running requires CUDA, or that it's an impossible task.
               | Remember, we're talking about >$1B companies here who can
               | easily assemble teams of 10s-100s of developers.
               | 
               | You can compare CUDA to the first PC OS, DOS 1.0. Sure,
               | DOS was viewed as a moat at the time, but it didn't keep
               | others from kicking its ass.
        
               | latchkey wrote:
               | > _You can compare CUDA to the first PC OS, DOS 1.0._
               | 
               | Sorry, I don't understand this comparison at all. CUDA
               | isn't some first version of an OS, not even close. It's
               | been developed for almost 20 years now. Bucketloads of
               | documentation, software and utility have been created
               | around it. It won't have its ass kicked by any stretch of
               | imagination.
        
               | amelius wrote:
               | Yes, CUDA has a history. And it shows. CUDA has very bad
               | integration with the OS for example. It's time some other
               | company (Microsoft sounds like a good contender) showed
               | them how you do this the right way.
               | 
               | Anyway, this all distracts from the fact that you don't
               | need an entire "OS" just to run some arithmetic
               | primitives to get transformers running.
        
               | latchkey wrote:
               | > _CUDA has very bad integration with the OS for
               | example._
               | 
               | If you want to cherry pick anything, you can. But in my
               | eyes, you're just solidifying my point. Software is
               | critical. Minimizing the surface is obviously a good
               | thing (tinygrad for example), but you're still going to
               | need people who are willing and able to write the code.
        
               | amelius wrote:
               | OK, but Microsoft is a software company ...
        
               | VirusNewbie wrote:
               | >Very few developers outside of Google have ever written
               | code for a TPU.
               | 
               | Anthropic?
        
         | imtringued wrote:
         | Internal ASICs are a completely different market. You know your
         | workloads and there is a finite number of them. It's as if you
         | had to build a web browser, normally an impossible task, except
         | it only needs to work with your company website, which only
         | uses 1% of all of the features a browser offers.
        
           | latchkey wrote:
           | Very true indeed. I'm not arguing against that at all.
        
         | migueldeicaza wrote:
         | The CUDA moat is real for general purpose computing and for
         | researchers that want a swiss army knife, but when it comes to
         | well known deployments, for either training or inference, the
         | amount of stuff that you need from a chip is quite limited.
         | 
         | You do not need most of CUDA, or most of the GPU functionality,
         | so dedicated chips make sense. It was great to see this theory
         | put to the test in the original llama.cpp stack which showed
         | just what you needed, the tiny llama.c that really shows how
         | little was actually needed and more recently how a small team
         | of engineers at Apple put together MLX.
        
           | latchkey wrote:
           | Absolutely agreed on the need for just specific parts of the
           | chip and tailoring to that. My point is bigger than that.
           | Even if you build a specific chip, you still need engineers
           | who understand the full picture.
        
       | outside1234 wrote:
       | For GPUs at least this is pretty obvious. For CPUs it is less
       | clear to me that they can do it more efficiently.
        
         | sgerenser wrote:
         | When Microsoft talks about "making their own CPUs," they just
         | mean putting together a large number of off-the-shelf Arm
         | Neoverse cores into their own SoC, not designing a fully custom
         | CPU. This is the same thing that Google and Amazon are doing as
         | well.
        
       | qwertytyyuu wrote:
       | better late than never to get into to game... right? right....?
        
         | johncolanduoni wrote:
         | Just like mobile!
        
       | alephnerd wrote:
       | I've mentioned this before on HN [0][1].
       | 
       | The name of the game has been custom SoCs and ASICs for a couple
       | years now, because inference and model training is an
       | "embarrassingly parallel" problem, and models that are optimized
       | for older hardware can provide similar gains to models that are
       | run on unoptimized but more performant hardware.
       | 
       | Same reason H100s remain a mainstay in the industry today, as
       | their performance profile is well understood now.
       | 
       | [0] - https://news.ycombinator.com/item?id=45275413
       | 
       | [1] - https://news.ycombinator.com/item?id=43383418
        
         | philipwhiuk wrote:
         | > The name of the game has been custom SoCs and ASICs for a
         | couple years now, because inference and model training is an
         | "embarrassingly parallel" problem, and models that are
         | optimized for older hardware can provide similar gains to
         | models that are run on unoptimized but more performant
         | hardware.
         | 
         | Is anyone else getting crypto flashbacks?
        
           | kcb wrote:
           | The difference is crypto wasn't memory and throughput
           | dependent. That's why a small asic on a USB stick could
           | outperform a GPU.
        
             | alephnerd wrote:
             | One thing to point out - "ASIC" is more of a business term
             | than a technical term.
             | 
             | The teams that work on custom ASIC design at (eg.) Broadcom
             | for Microsoft are basically designing custom GPUs for MS,
             | but these will only meet the requirements that Microsoft
             | lays out, and Microsoft would have full insight and
             | visibility into the entire architecture.
        
       | croisillon wrote:
       | homemade chips is probably a lot of fun but buying regular Lay's
       | is so much easier
        
       | babuloseo wrote:
       | The current M$ sure is doing a great job at making people move to
       | alternatives.
        
         | bongodongobob wrote:
         | People, sure, but that's not their target demographic. It's
         | businesses and they aren't moving away from MS anytime soon.
        
       | floxy wrote:
       | On a slightly different tangent, is anyone working on analog
       | machine learning ASICs? Sub-threshold CMOS or something? I mean
       | even at the research level? Using a handful of transistor for an
       | analog multiplier. And get all of the crazy fascinating
       | translinear stuff of Barrie Gilbert fame.
       | 
       | https://www.electronicdesign.com/technologies/analog/article...
       | 
       | https://www.analog.com/en/resources/analog-dialogue/articles...
       | 
       | http://madvlsi.olin.edu/bminch/talks/090402_atact.pdf
        
         | nickpsecurity wrote:
         | A bunch of people. Just type these terms into DuckDuckGo:
         | 
         | analog neural network hardware
         | 
         | physical neural network hardware
         | 
         | Put "this paper" after each one to get academic research. Try
         | it with and without that phrase. Also, add "survey" to the next
         | iteration.
         | 
         | The papers that pop up will have the internal jargon the
         | researchers use to describe their work. You can further search
         | with it.
         | 
         | The "this paper," "survey," and internal jargon in various
         | combinations are how I find most CompSci things I share.
        
           | accrual wrote:
           | Thanks for these helpful search terms!
        
           | tovej wrote:
           | Are you aware of dblp.org?
        
         | xadhominemx wrote:
         | For large models, the bottlenecks are memory bandwidth,
         | network, and power consumption by the DAC/ADC arrays
         | 
         | It's never come even close to penciling out in practice.
         | 
         | For small models there are people working on this implemented
         | in flash memory eg Mythic.
        
       | improgrammer007 wrote:
       | For those who don't know Msft is working on
       | https://azure.microsoft.com/en-us/blog/azure-maia-for-the-er...
        
         | hulitu wrote:
         | > For those who don't know Msft is working on
         | 
         | We do know: ads, spyware and rounding corners of UI elements.
         | 
         | If their processor work like their software, i really feel pity
         | for people who use it.
        
       | snowwrestler wrote:
       | Made where? Isn't foundry capacity the limiting factor on chips
       | for AI right now?
        
         | Jyaif wrote:
         | By cutting the middle man out, MS could pay TSMC more than
         | nvidia per wafer and still save money.
        
           | fishmicrowaver wrote:
           | This is the whole game right here.
        
           | nsteel wrote:
           | They can even pay Broadcom to be a lower-level middle man
           | instead. Despite the BRCM tax, it'll still be way cheaper
           | than going to Nvidia.
        
         | migueldeicaza wrote:
         | TSMC manufactures the MAIA 100:
         | 
         | https://azure.microsoft.com/en-us/blog/azure-maia-for-the-er...
        
       | blibble wrote:
       | well yeah, I can't imagine sending all your shareholders money to
       | nvidia to produce slop no-one is willing to pay for is going down
       | too well
        
       | sleepybrett wrote:
       | Microsoft just can't stop following apple's lead.
        
       | pjmlp wrote:
       | Doesn't come as a surprise, I imagine they would also build on
       | top of Direct Compute, or something else they can think of.
        
       | alecco wrote:
       | For many years, every few months Microsoft and Meta say they are
       | going to do AI hardware. But nothing tangible is delivered.
        
         | Herring wrote:
         | Yeah they could just be negotiating a better deal w nvidia.
        
       | jtfrench wrote:
       | "Microsoft Silicon", coming up.
       | 
       | Is it practical for them to buy an existing chip maker? Or would
       | they just go home-grown?
       | 
       | - As of today, Nvidia's market cap is a whopping 4.51 trillion
       | USD compared to Microsoft's 3.85 trillion USD, so that might not
       | work.
       | 
       | - AMD's market cap is 266.49 billion USD, which is more in reach.
        
         | floxy wrote:
         | Intel is almost a bargain at $175 billion
        
           | j_walter wrote:
           | They were more of a bargain 3 months ago at $80B.
        
         | hulitu wrote:
         | > "Microsoft Silicon", coming up.
         | 
         | Will they equip the new Microsoft Vacuum Cleaner with it ? /s
        
         | arisAlexis wrote:
         | They could buy with peanut money Cerebras
        
       | foobarbecue wrote:
       | Weird use of "homemade"! I guess they mean "in-house"?
        
         | ravenstine wrote:
         | Just like how mama used to make 'em!
        
           | brnt wrote:
           | Moms secret ingredient to her AI was Nvidia!
        
           | smallmancontrov wrote:
           | It's a delightful coincidence of history that the "What's
           | Jensen been cooking" pandemic gag happened on the generation
           | that would wake up the AIs.
           | 
           | https://www.youtube.com/watch?v=So7TNRhIYJ8
        
       | john01dav wrote:
       | I don't like this trend where all the big tech companies are
       | bringing hardware in-house, because it makes it unavailable to
       | everyone else. I'd rather not have it be the case that everyone
       | who isn't big tech either pays the Nvidia tax or deals with
       | comparatively worse hardware, while each big tech has their own.
       | If these big tech companies also sold their in-house chips, then
       | that would address this problem. I like what Ampere is doing in
       | this space.
        
         | Etheryte wrote:
         | Another way to look at it would be that if all the big players
         | stop buying up all the GPUs, prices will come back down for
         | regular consumers, making it available for everyone else.
        
           | jlarocco wrote:
           | But there's a real risk that in the long term all of the
           | "serious" AI hardware research will get done inside a few big
           | companies, essentially shutting out smaller players with a
           | hardware moat.
           | 
           | Unless they start selling the hardware, but in the current AI
           | market nobody would do that because it's their special sauce.
           | 
           | On the other hand, maybe it's no different than any other
           | hardware, and other makers will catch up eventually.
        
             | tyre wrote:
             | And some of those people will leave to start their own
             | companies or join competitor chipmakers. Innovation will
             | continue.
        
       | dang wrote:
       | Url changed from
       | https://www.theregister.com/2025/10/02/microsoft_maia_dc/, which
       | points to this.
       | 
       | Submitters: " _Please submit the original source. If a post
       | reports on something found on another site, submit the latter._ "
       | - https://news.ycombinator.com/newsguidelines.html
        
       | eptcyka wrote:
       | I do too.
        
       | yalogin wrote:
       | The big data center owners will want to build their own hardware,
       | that is a no brainer for them.
        
       ___________________________________________________________________
       (page generated 2025-10-03 23:01 UTC)