AMD signs AI chip-supply deal with OpenAI, gives it option to take a 10% stake (opens in new tab)

Fully anecdotal, but nearly every knowledge worker I know has Copilot or ChatGPT or some kind of corporate LLM subscription included in their job now. I don’t think the majority use it very well, but I know at least a handful who do.

Corporate America seems ready to spend real money on AI, at least for now. This money being spent today doesn’t come close to recouping the investment OpenAI et al have made, but the trickle’s begun. The money’s not all imaginary.

But also under the assumption that the taxpayers will bail them out and foot the bill when things go wrong. Privatize the gains, socialize the losses. This is completely on the assumption that the US government would be able to do this, despite the bond market, inflation, debt that would constrain any response.

Does anyone here who lived through the dotcom bubble feel like we're in a similar situation with AI? I am excited about AI and recognize its potential. However, the increasing number of business deals like this one is raising my concerns that we may be heading toward a harsh bubble.

It's probably closer to 5x (80% gross margin) than 20x.

>> Nah, nothing in the current market is based on fundamentals.

Sure it is. Stock prices are based on supply and demand. Oh, did you mean company fundamentals? Hmm.

BTW that demand is all coming from 401k contributions. The powers that be are terrified of a downturn causing too many job losses tanking the whole thing.

> stock price, and AMD's will go down

stonks goes up

Imagine if the answer to the Fermi paradox ends up being not "they destroyed themselves through war" but "they destroyed themselves trying to find meaning in money and fake thinking machines".

Sounds like you are talking about AGI or superintelligence or something? Because productive AI tools are already here and have been for a few years. AI has already "worked out". Nothing speculative about that.

If a company gave away some portion of is assets, shouldn't its market cap go _down_ by that amount not up? That's how stocks trade ex dividend e.g.

The stock popped quite a bit today, so investors seem to think this is a good deal for AMD. I tend to agree, may finally enable AMD to get a foot in the door in the whole large scale AI market.

Can someone with banking / deal-making experience explain to me how this warrant works? My point: If OpenAI exercises the warrant and buys 10% of AMD for 1 penny per share, where does AMD get the shares? Or will they do a secondary offering and dilute existing shares? Or do they take from treasury holdings (of shares)? Or does AMD coordinate a buyback of shares, then deliver to OpenAI?

I’ve seen a lot of confusion about this type of deal recently: notably, it is often taken to imply something more or less shady is going on. I’m not sure when such arrangements became a thing, but I know equity stakes have long been an important part of enterprise SaaS deals. The reasoning is relatively straight-forward: if a large client commits to a vendor in a way that holds some risk to the former and will materially impact the latter’s business — and especially if the client’s support of said vendor will directly or indirectly benefit their competitors who might also use this vendor — an equity stake is a way to offset risk with upside.

You can see this play out in the history of OpenAI. NVIDIA supported them from an early stage and in exchange received OpenAI equity to offset the risk. Now from a position of relative strength, OpenAI has become concerned about vendor lock-in and so is rationally exploring AMD. Yet, because any such deal will materially impact AMD’s stock price and there is risk both of losing time trying to train with new chips as well as of benefiting competitors if they work with AMD to improve their hardware offerings/APIs, it is reasonable to ask for equity upside. So, for the same reasons (increase in stock price and enterprise client who will help improve their product offering) only without risk, is it understandable why AMD would want to offer equity on such favorable terms.

TLDR; My sense is that the sudden skepticism towards this relatively common enterprise deal structure seems to derive from the understandable interest in identifying signs of an AI bubble. Such a bubble may (and indeed almost certainly does) exist, but I don’t think this is evidence thereof.

----

EDIT: I'm just clarifying something I saw in a lot of responses. My only point is that it is important to try and empirically tease out what represents: (1) a circular deal in which vendors facing the limits of growth are subsidizing vulnerable clients; versus (2) a risk-hedging deal in which a non-market leading vendor offers upside to a market leading client.

I believe the recent Oracle and NVIDIA deals are cases of (1) that provide evidence of an AI bubble, but that this AMD deal is most likely a case of (2) that provides no further evidence.

>Couldn't they just buy the AMD chips if they're good enough?

OpenAI would presumably need to raise money to buy the AMD chips.

The "genius" of this deal is that AMD is "giving away" 10% of the company (at $0.01/share) to OpenAI. Then OpenAI will presumably turn around and sell those shares (or borrow against them) to raise enough money to purchase the AMD GPUs.

> allows OpenAI to buy 160 million shares at 1 cents a share.

> I think at least part of the 10% is if AMD stock reaches 600.

AMD market cap today is $350B (at $200/share).

AMD would need to 3x their market cap ($1,000B) to be at $600/share.

Which would mean that OpenAI could gain $100B in AMD stock, for the minuscule cost of only $1.6 million (160 million shares at 1 cent each).

--

Sam is spinning the world on his finger tip with these deals he's crafting.

> I also don't get why there commiting so much to the future, are they sure of the quality of the products and their demand that much?

It's one big game of musical chairs, and everyone can hear the phonograph slowing down.

OpenAI is making these desperation plays because they've ran out of hype. GPT-5 "bombed", the wider public doesn't believe AI is going to keep getting exponentially better anymore. They're out of options to generate new hype beyond spewing ever larger numbers into the news cycle.

AMD is making this desperation play because soon, once the AI bubble pops, there'll be a flood of cheap unused GPUs & GPU compute. Nobody's going to be buying their new cards when you can get Nvidia's prior gen for pennies on the dollar.

This is too top of the top to ignore. Everyone can see the scam now, it's a joke

> Obviously, for the stock price to go up money needs to come from somewhere.

Not convinced that’s true anymore in current climate. Bigger numbers announcements and AI Pixie dust works too apparently lol

> A tax on the mistaken believe that NVidia has an monopoly on putting transitions in a particular configuration which they obviously don't

NVIDIA doesn't place transistors in particular configurations. Foundries do that for them. And it is currently common sense that the software is the moat, not the hardware design.

Good luck changing the ecosystem to use AMD.

I figure these GPUs are typically around 1kW (unclear if 6GW is including overhead like cooling, which might double(?) the power), so in the range of 3-6M GPUs.

If these are somewhere in the range of $10-30k (who knows what current or future models are contemplated), that's $30-180B. So clearly the low end doesn't make sense for the 'rebate', but at the high end a ~17% discount doesn't seem unreasonable.

> They say that Open ai committed to buying six gigawatts of AMD chips. Anybody know how to convert that into money?

MI350 spec sheet says it's 1000 watts typical. So we're talking about something on the scale of a couple million chips.

Microsoft seems to have made a good deal: it got the models AND profits.

Also microsoft is pushing copilot to office and I think it will sell. Since they sell to general B2B and not only to the peogrammer niche.

AMD is trying to buy market share by donating 10% equity. I also think it is crazy

I didn't mean to imply that there was a winner or loser. Just that AMD was subsidizing it's GPUs with equity.

I think there are logical reasons for both companies to agree to this deal. AMD is trying to break CUDA dominance. OpenAI is getting extremely cheap compute for expansion and they'd also benefit from the Nvidia monopoly falling if that ever happens.

OpenAI isn't publicly listed so it's hard to tell how this affects them from a "share price" concept. However for a company that's not public and only has capital from financing rounds and revenue, this gives OpenAI a lot more flexibility for the future and hedges risk while maximizing upside.

Its called hedging your risk: making sure Nvidia isn't the only game in town.

To paraphrase Dire Straits - The money from nothing and the chips for free.

OpenAI and AMD both have strong incentives to chip away at Nvidia's dominant position, and the best way to do that is by releasing software tools to the public.

Ask yourself: What would first-class PyTorch/Triton and Jax support for AMD hardware do to Nvidia's dominance and pricing power?

Today, yes, it's true that AMD hardware can be competitive mainly for inference.

The story here is about the future. Over time, OpenAI would benefit if AMD hardware becomes competitive for training too.

Nvidia currently gets to charge whatever the market can bear on its dominant training hardware.

If AMD hardware becomes a real alternative for training, Nvidia will be forced to compete on price.

Larger memory, weaker comms. You can optimize for this by doing things like increasing batch size/data parallelism vs sharding schemes with more comms.

At scale training won’t be able to avoid comms entirely, while many models can fit in a single MI300 for serving.

> OpenAI

> "If they even share openly"

That fact that this even needs to be pointed out but is normalized in the AI industry. Sadly, in the history of Open AI, they've been anything but open.

For inference there are many solutions like groq. The margins there will be small.

The fat margins is in training, in NVidia hardware.

The option to buy shares seems pretty well locked down with multiple vesting conditions So pretty unlikely that this ends up in a “for nothing” space

https://ir.amd.com/news-events/press-releases/detail/1260/am...

> AMD just gave away $160m for nothing in return.

Where'd you get this number? The equity is worth closer to 34bil if fully realized?

No... because the warrants can only be exercised if the share price is at certain points. If AMD goes down, the warrants can't be exercised at all. E.g. we know at least one share price points is when AMD shares are worth $600/share.

I think OpenAI will try and continue this elsewhere, which would be pretty worrying. It lets them not give up any equity, just use their name to pump stocks and earn capital.

Nothing in return? AMD stock is up 30% today on this news.

if you buy all the chips, no one else can build agi. sama probably

Michael Lewis has written a bunch of books about financial crises and their origins. He wrote The Big Short for example.

If you’re into tech & startups, he wrote Going Infinite about FTX, and Flash Boys about high frequency trading.

Pretty much. Strip away the AI narrative and there’s not much else driving growth right now - manufacturing’s flat, consumer spending’s slowing, and traditional tech margins are shrinking.

This artificial boom in “AI infrastructure” is basically the last engine keeping the charts pointing up. When that music stops, there’s nothing underneath it but leverage and power bills.

Reminds me of PGs 2008 essay passage:

>Sam Altman has it. You could parachute him into an island full of cannibals and come back in 5 years and he'd be the king. If you're Sam Altman, you don't have to be profitable to convey to investors that you'll succeed with or without them.

Looks like he's doing well wheeling and dealing us cannibals.

It's valued from secondary markets based on how much it can raise from private investments. These get reported out eventually which juices the hype further in a bit of a vicious cycle. this cycle also has quite the incestuous component too with how every major company seems to have their hands in each other's pants.

Sale of shares on secondary market at that valuation.

https://www.reuters.com/technology/openai-hits-500-billion-v...

Are consumer cards also benefiting from the improvements or only datacenters?

AMD already make GPUs that are competitive with Nvidia at every tier except the very top end.

It’s the positioning and pricing that isn’t competitive.

NVidia is terrible value for consumers when comparing their own offering from previous years. The graph from https://gamersnexus.net/gpus/great-nvidia-switcheroo-gpu-shr... shows how low/mid segments become weaker and weaker with each generation.

Meanwhile AMD doesn't try to disturb the market, they keep their pricing at "price of similar NVidia card - $50". They aren't going to gain market share that way.

“He’s my cuz so I’ll let him win the biggest business battle there is.”

I dunno, doesn’t have the ring if credibility to it.

So the “we’re so successful, we don’t need any more business, so we’re just cruising”, strategy eh?

Doesn’t resonate for me as how most American companies do business.

This isn't stated in any of their press releases.

They are getting a meaningless deal unless they can nearly triple their market value from here. Sam isn't buying AMD shares for more than a penny, what should we value it at?

>Existing AMD shareholders are getting a great deal; their shares are worth 30% more today and will be worth 5X more if OpenAI gets to use their option. Yes, there is some dilution for existing shareholders, but only after a 5X gain.

So if I read this right then an existing shareholder benefits up to the point where AMD stock reaches the $600/shr level and after that it becomes a dilution.

For the maximum benefit to existing shareholders, the stock price must get near $600/shr and if that looks unlikely they should consider other investments on less speculative terms. This whole AI thing feels like I'm watching the soapy fluid flow along the bubble exterior to form a droplet of soap on the bottom, thinning and weakening the bubble until it pops due to fluid film rupture at the top of the bubble as the droplet leaks away due to gravity.

Not all of the $100 billion are purchases by OpenAI ("AMD expects to receive more than $100 billion in new revenue over four years from OpenAI and other customers").

For some reason the OpenAI portion of this deal is quoted in gigawatts rather than number of MI450s purchased, which makes it hard to tell how much of that $100 billion is from OpenAI. It's probably around $80 billion.

Just the 1st tranche of shares would be like $6 billion at $220 so if they borrow against that they can fund it. If the hype continues they keep it going

Just enjoy it if you can’t avoid it

Alternatively, they're just doing a clips episode based on 2000.

Trading any of this, the only question that you need to answer is:

What is more painful to you, missing some gains on the way up or holding positions that have flipped from positive to negative very rapidly?

Markets like this, the moves will be violent in both directions, not just on the way up.

The fact you are even writing this usually means you should sell immediately.

> Cheap cloud gaming.

> Those GPU's will finally push pixels again :)

Have they ever? I wonder if, say, an H100 even supports graphics APIs.

Cloud gaming sucks due to latency, not price.

They won’t. High end AI GPUs get stomped by mid tier gaming GPUs on pixels

There will definitely be takers though. Can see scientific community for example loving some cheap GPUs

What a mess. When they have to write about how “circular financing” became highly regulated in the future, I hope they know that the problem was not confusing in the big picture.

Maybe I can get quoted in a sidebar:

“Hey Econ students, we normal people could see that the AI market was extremely gamed. We can see the investment feedback loop. It is just that the organizations continuing it have control over so much money, they don’t have to stop and ask society in general for permission to continue. Really this is a symptom of the wealth inequality crisis that they covered chapter…”

Oh, we saw it before. 1998-2000. Cisco, Sun, et al, "investing" in the dotcoms so those dotcoms could turn around and use the funds to buy servers and routers from them.

The real, sustained value creation for AMD is if their GPUs can be competitive with Nvidia for inference or training. The market is currently betting that AMD will get that "spillover" effect from working with OpenAI to get their chips + software in a state that openAI can use at scale.

So day 1 valuation is mostly hype, but that hype may translate into real value later.

Nvidia owns a piece of OAI, and now Nvidia would own a piece of AMD. I am just waiting for Samsung type of circular ownership structures here in the US.

Can you explain why you think that is? The SP500 has never been more top heavy and when those 7 or so all AI associated incest stocks falter I can't see anyway it's not going to be a bloodbath of the decade long type.

Based on the recently released graph of how people are using chatgpt. ~80% of use cases (practical guidance, seeking information, writing) could presumably run on a local model.

My assumption is that most users won't actually care if the LLM is in the cloud or device. That said, quite a few folks have iPhones and Apple's only way into the AI race is to go to it's strength, 1B+ hardware devices that they design the silicon for. They will produce a phone that runs a local LLM and market it as private and secure. People upgrade every couple of years (lose or breaks) so this will drive adoption. I'm not saying people will vibe code on their iphones.

Price, for one. I don't mind running a local model at half the speed if all it costs is electricity.

A local model basically allows me to experiment with running an agent 24x7, 365 days a year with continuous prompting.

SaaS won't be able to match that.

The models running on $50k GPUs will get better but the models running on commodity hardware will hit an inflection point where they are good enough for most use cases.

If I had to guess I would say that's probably 10 or 15 years away for desktop class hardware and longer for mobile (maybe another 10 years).

Maybe the frontier models of 2040 are being used for more advanced things like medical research and not generating CRUD apps or photos of kittens. That would mean that the average person is likely using the commodity models that are either free or extremely cheap to use.

ok, you can technically upload all your photos to Google cloud for all the same semantic labeling features as iOS Photos app, but having local, always available and fast local inferencing is arguably more useful and valuable to the end user.

Trying to compare RAM size and CPU cores is so yesterday. Apple owns the entire stack they can make anything fit into their core if they so desire.

> What's the benefit to running LLMs locally?

At work:

That I don't rent $30,000 a month of PTUs from Microsoft. That I can put more restricted data classifications into it.

> LLM inferencing isn't particularly constrained by Internet latency

But user experience is

The data you need is mostly not remote. A friend works at a software development company, they can use LLMs, but only local ones (local as in their datacenter) and it can only be trained on their code base). Customer service LLMs need to be trained on in-house material, not generic Internet sources.

The general advantage is that you know that you're not leaking information, because there's nowhere to leak it to. You know the exact input, because you provided it. You also get the benefit of being able to have on device encryption, the data is no good in the datacenter if it's encrypted.

You're right. It's not really money printing, at least not directly. Only banks can print money.

What these recent deals do is inflate asset prices by making (future) revenues appear higher or perhaps just more certain than they really are.

Assets can be used as collateral for loans. If someone were to use their AMD shares as collateral for a loan at a commercial bank (not a margin loan), that would be money printing and you could print more of it today than before the deal was announced.

> because it’s backed by something

If I finance a car backed by hopes and dreams, I'd be driving real good.

Backed by what?

AMD issues new shares and gets a penny (read: effectively zero) back for them.

ALL ELSE BEING EQUAL this means everyone holding AMD has 10% of their equity/value taken away and handed to OpenAI.

But all else is not equal. OpenAI only gets the shares if they buy AMD GPUs. The intent is that this offsets the dilution by making AMD overall more valuable. (This is why the stock price jumped on the announcement) It's a GPU subsidy paid for by AMD's shareholders rather than AMD itself.

The real risk is that this further entangles AMD in the AI bubble. OpenAI already has enormous datacenter construction obligations. The likelihood of them failing to meet these new obligations, and thus this deal falling through or otherwise not materialising, is pretty high. If the AI bubble goes *POP*, AMD will be hurting a lot more than before this deal.

It isn’t printing money (money supply was not increased, and no one has access to more money due to the contract signed between the two parties), but I assume it makes the poster feel good to be outraged and express it by writing something like that.

It is dilution. 160 million shares of AMD for 1 cent will be added. The idea is that the value increase of the remaining shares enabled by the OpenAI cooperation will be stronger than the decrease caused by dilution.

if only they could also anticipate what will happen once the tide is falling