Array ( [content] => [params] => Array ( [0] => /forum/threads/openai-inks-amd-chips-deal-worth-tens-of-billions-of-dollars.23740/page-2 ) [addOns] => Array ( [DL6/MLTP] => 13 [Hampel/TimeZoneDebug] => 1000070 [SV/ChangePostDate] => 2010200 [SemiWiki/Newsletter] => 1000010 [SemiWiki/WPMenu] => 1000010 [SemiWiki/XPressExtend] => 1000010 [ThemeHouse/XLink] => 1000970 [ThemeHouse/XPress] => 1010570 [XF] => 2030770 [XFI] => 1060170 ) [wordpress] => /var/www/html )
You mean Elon as well?![]()
So many companies' stocks now depend on Sam Altman actually buying what he signed those MoUs for.
Oracle
AMD
NVIDIA
Microsoft
Forgetting a few more.
We seem to be converging to a singularity - OpenAI.
It's kind of fun here:
Sept 2025:
Nvidia invests (up to) $100B in OpenAI as they buy GPUs and build out data centers together
Oct 2025:
OpenAI invests up to $60B in AMD as they sell GPUs / build out data centers together
The cynic in me also can't help but wonder "Is OpenAI a money laundering scheme (illegal export sales) for Nvidia?"![]()
the AI frenzy will be pushed a bit further with this for sure. Some say OpenAI has already become too big to fail. What do you guys think about Elon's xAI effort as a comparison?So many companies' stocks now depend on Sam Altman actually buying what he signed those MoUs for.
Oracle
AMD
NVIDIA
Microsoft
Forgetting a few more.
We seem to be converging to a singularity - OpenAI.
On this topic: HBM 4E memory employs a separate logic die. Those die are being made at Foundries. Samsung is using their own Foundry while Hynix is using TSMC.Samsung
SK hynix
If this continues, we won't have enough production facilities...I bet all the customers need to prepay more to secure production capacity from TSMC.
On this topic: HBM 4E memory employs a separate logic die. Those die are being made at Foundries. Samsung is using their own Foundry while Hynix is using TSMC.
AI Overview
Logic dies for HBM4 memory are being fabbed by different companies, with Samsung Electronics using its own 4nm foundry process and TSMC producing base dies on its 12nm 12FFC+ and 5nm N5 processes for SK Hynix. Some manufacturers are also exploring innovative designs, such as SK Hynix's goal to integrate logic and memory on a single die for HBM4 to improve performance and reduce complexity.
the AI frenzy will be pushed a bit further with this for sure. Some say OpenAI has already become too big to fail. What do you guys think about Elon's xAI effort as a comparison?
Is it just me or do some of these creative financing arrangements (the CoreWeave/nVidia circular relationship being just one example and this feeling in some ways similar) display symptoms of a positive feedback system ? [OpenAI can't have all the $100bns they claim to be investing/spending/whatever]. And if so, should we really be expecting a stable result ?
Intel is doing the base dies at their fab as well for their GPUs iircIn addition to SK hynix, I believe Micron also provides HBM memory to TSMC’s customers for packaging at TSMC.
I also believe Samsung, a latecomer to HBM3, is doing (or will do) the same, working with TSMC for integration. I’m not sure how far their collaboration will go.
Stratechery has a great substack article analyzing OpenAI recent move.
![]()
OpenAI’s Windows Play
OpenAI is making a play to be the Windows of AI: the all-encompassing platform that controls both hardware supplier and software developers.stratechery.com
View attachment 3713