$AMD shipped Helios to OAI and Anthropic in September.
Then Lisa Su went on a mission to TSMC and SK Hynix and Samsung to tell them she needed more capacity.
Does anyone think she did this on spec?
She just dreamed up the added capacity that she needed? Or did the customers tell her they wanted more capacity.
Then Sam went to Samsung and told them, some companies (with motorcycle jackets) were building chips their customers didn’t order.
Who do you think that is?
Lisa or Jensen?
Mfer is double fisting.
Lisa Su delivered the hammer to my man.
He has been in full freak-out panic mode since MI455X and Helios were announced last year.
Now that it is shipping, he has totally burned out.
Soon everyone will understand this.
$AMD
17
1,269
Neoclouds that were going to be forced to go with Nvidia are now going to be forced to go with TCO.
They are no longer obligated to pay the Nvidia tax.
They don’t get the money from Nvidia, but their capacity in the pipeline will generate substantial revenue.
When they deploy that revenue, they will have a fiduciary responsibility to maximize their infrastructure around TCO.
When you provide the compute with the lowest TCO, you win this business.
That’s $AMD
1
21
1,280
MLPerf has a new “Endpoints” submission program that allows submissions on a rolling basis starting from October 12, 2026.
AMD ran testing on MI455X in October and announced the results last week for Deepseek R1.
If AMD officially submitted MLPerf in October, the results could be announced soon after the review process has been completed.
Results are provided every 1st and 3rd Wednesday of the month.
The next window is this Wednesday, November 18th.
They ran the test.
They say it is in accordance with MLPerf.
I imagine they would have submitted it for review.
If they submitted it for review in October when they ran the test, it could likely come out this week.
The test will show 21,333 tokens/sec/GPU
Vs Rubin at 16,435 tokens/sec/GPU.
I don’t know why this would get glossed over.
AMD Claims Its New AI Chip Beats Nvidia’s Latest by 30% — But AMD Ran the Tests
On November 12, AMD published benchmark results showing its new Instinct MI455X accelerator processing 21,333 tokens per second running DeepSeek-R1 in the MLPerf Inference Offline configuration
192
$AMD
If MI555X is (2) MI455X on the same package with HBM 4E instead of HBM 4,
This will dust Rubin Ultra by a mile.
I would find a honeypot and get ousted too if this was on the horizon.
Mama Mia. 🤯🤯🤯🤯
Good thing $AMD is bringing Mext to market.
Now when you buy AMD EPYC CPUs, you only need about half the DRAM to end up with 95% of the performance.
You can install twice the CPUs on the same memory footprint.
teapawt.org/liwei_twcapital/status…
Intel CEO Lip-Bu Tan says memory prices are up six to seven times and AI now needs as many CPUs as GPUs
"The memory becomes a constraint, next year it's going to be even worse"
"The prices have gone through the roof... Yes, it's six, seven times"
"The ratio, GPU to CPU is 8 to 1. And now it's kind of parity"
He also says the gap between AI capex and the revenue AI generates is still very big ($INTC $NVDA $MU)
Agents are turning the CPU back into a growth story, and memory is the new bottleneck
via @YahooFinance, featuring Lip-Bu Tan
This would be good for $AMD
Definitely less cutthroat supply chain games.
Not really a visionary, milk toast, bland, cuck.
He can manage Nvidia’s decline.
EXCLUSIVE FT BREAKING NEWS ‼️
Arm boss Rene Haas approached in Nvidia CEO search
Full Story: 👇👇👇
Nvidia’s board has approached Arm chief executive Rene Haas in its search for a potential successor to Jensen Huang, according to people familiar with the matter.
Haas, who has led the SoftBank-backed UK chip company since 2022, worked at Nvidiafor about seven years more than a decade ago, and still has a strong relationship with both Huang and some members of the company’s board.
Haas is among a number of candidates to take over from Huang at Nvidia, and there is no guarantee he will be offered the job or accept it, the sources said.
Colette Kress, Nvidia’s chief financial officer, is serving as the company’s interim CEO.
Arm declined to comment. Nvidia did not immediately respond to a request for comment.
$NVDA
10
1,125
Mext is an under appreciated advantage for $AMD, it allows Hyperscalers to greatly save on DRAM without sacrificing performance.
In many workloads this translates to half the DRAM requirement.
In a memory shortage era this makes AMD CPUs value more than double the competition.
It’s amazing what you can achieve when you aren’t constantly scheming to subvert government restrictions or navigating endless affairs.
Holy moly. AI itself is making computing infrastructure dramatically more efficient.
AMD just found a way to run servers with 50% less RAM while retaining 95% of performance. 🤯
AMD's new MEXT technology uses AI-powered predictive memory to move rarely accessed data from expensive DRAM to cheaper NVMe flash storage, then predictively restores memory pages before applications need them.
In AMD's tests, this preserved 95% of Synopsys PrimeTime performance with half the physical DRAM, while its Spark SQL benchmark retained 99%.
This new technology is already running on 145 AMD servers, with plans to scale to 2,000 servers by the end of 2027.
What happens when all the benchmarks have MI455X in the lead?
What happens when headlines get splashed all over X about Helios taking a commanding lead?
With no more Jensen to backstop Nvidia chips and grab the entire silicon supply chain capacity, what happens then?
The GPUs are better
The CPUs are better
The AI racks are better
Software is no longer a thing.
$AMD
Sounds like Rubin is a disaster so Nvidia is going to dogfood the market with more Blackwell until the end of time.
4D chess.
Maybe just make a smart decision, instead of continually fucking up.
Re: the rumor that NVIDIA cut orders
One thing I’ve been able to confirm is that Rubin has been delayed again. I’m told motherboard warpage caused the latest delay and that the HBM issue was resolved in October. (As it happens, Korean media reported today that SK hynix’s HBM4 yield had reached 70%.) As a result, volume shipments of Rubin have slipped to 2027, and I’m told NVIDIA has increased its Blackwell orders in the meantime.
Since posting about the rumored order cuts, I’ve spoken with several funds and supply chain contacts. One rumor making the rounds is that CCL orders were cut by 30%. One source I spoke with said that claim was false.
I hope this helps. Thanks.
1
3
333
Logically, they probably resolved the issue with HBM yields on the nerfing. Now they need to assemble the GPUs with the newly qualified HBM and that is causing the delay.
It seems like they held on as long as they could to try and get that last contract signed with Elon before the de-spec. Then a month to make it seem like they didn’t know.
Real slimy, dirtbag operation, get the contract signed then break out the rug pull.
It will be hilarious to watch these chips fail at unprecedented rates in space. It seems like an amazing waste of money to launch a rack, then watch the GPUs gradually go dark. They have enough trouble with reliability on earth, now you are going to launch overclocked, fragile, warping, HBM yield issues into space?
teapawt.org/0xmitsurii/status/2001…
I guess the nerfing Nvidia did to Rubin didn’t do the trick.
They are really in for some shit now. How low can they go?
AMD doesn’t seem to have yield issues at 7.6 Gbps, Nvidia could try that. That will give them 15.5 TB/s bandwidth to 23.3 TB/s for MI455X.
Basically 66% performance on memory bound workloads, so inference. The future of compute.
Is anyone paying attention out there?
$AMD
1
241
I guess the nerfing Nvidia did to Rubin didn’t do the trick.
They are really in for some shit now. How low can they go?
AMD doesn’t seem to have yield issues at 7.6 Gbps, Nvidia could try that. That will give them 15.5 TB/s bandwidth to 23.3 TB/s for MI455X.
Basically 66% performance on memory bound workloads, so inference. The future of compute.
Is anyone paying attention out there?
$AMD
Taiwan’s KYEC lowered its Q4 guidance due to delays affecting a major GPU, believed to be NVIDIA’s Rubin, caused by HBM4 yield issues at a memory supplier.
finance.technews.tw/2026/11/…
AMD has its own proprietary packaging technology that only they will be able to deploy. EFB
This allows them to package more compute into a single GPU.
This leads to more compute per watt.
If something isn’t getting installed, it is an ASIC.
Limitations would only make AMD’s compute more valuable.
$AMD
Every so often Wallstreet wants to shuffle money around.
So they clobber the best ideas and prop up the shittiest ideas in an attempt to move money from the shittiest to the best.
AI is going away and we are all going back to using software, got it.
Some congressman said something. Good luck.
Okay AI is totally finished, now what?
Stupid never lasts, there are many more sunny days ahead.
$AMD
12
1,031
Everyone seems convinced that there will be less datacenters in the future. I guess Democrats in congress will be able to limit international datacenters too.
So now we should only install the most efficient GPUs on the market. Anything else will not provide the compute per limited watts.
$AMD
3
390
Nvidia gets roughly $500,000 per CoWoS wafer in annual revenue.
If you believe AMD gets half that at $250,000 then 530,000 wafers would be $132B in revenues for 2027.
Today we get our first report with Helios shipping.
Every quarter afterthis quarter we get more Helios.
$AMD
I am going to look back with fond memories on the shares I bought today when it is July 2027 and Helios is banging out billions in revenues while MI555X starts its ramp.
$AMD
It has been 508 days since $AMD introduced Helios at Advancing AI in 2025.
Tomorrow we see the first revenue hit the books.
Then every quarter after this day we see the revenue go up at an amazing rate for years.
We have come a long way, but we have only now reached the starting line.
Time to run.
What do hyperscalers think about Nvidia dropping the performance on them by 17% at the last hour?
Are they in love with Nvidia? They waited until after all of them had no alternative but to move forward, and then they dropped the specs.
They had to know this was going to happen, months ago. People might have made different choices months ago.
Nvidia is a piece of shit company.
They are really trying to piss the customers off intentionally now. Let’s see how that works out for them.
1
8
1,038
Everyone questions why Nvidia is so cheap. $NVDA forward PE is 16.
Investors are not as stupid as many think, how sustainable is growth expansion when you don’t have the performance lead?
Nvidia can’t just buy all their own chips, at some point customers get a say in which rack gets ordered.

















