Pppp Dz

[Lucas Patiri @lucaspatiri_]

https://t.co/CZueUct2Eb

[Miles Ma @miles_mazy]

https://t.co/ECB42N6gYS

[VoxCat @VoxcatAI]

如果你真的看懂这篇文章，并且举一反三，那么市面上 90% 的 GPT 生图 Skill，都可以直接取消收藏了。

[Florian Darroman @floriandarroman]

Must read.

[三个三 @3three_AI]

https://t.co/VJUxgayNuw

[李岳 @liyue_ai]

https://t.co/wRDYvz1peQ

[Pppp Dz @pppp_dz]

@batatalal @JackieNate1 Thanks to Ponte Vecchio, lessons learned!

you can't expect to mix & match feveros with emib in there as well, too much

[Pppp Dz @pppp_dz]

@IntelProMUltra apple could be another customer under this model

[Jukan @jukan05]

Wells Fargo noted in its NVIDIA GTC 2026 preview report that NVIDIA’s Feynman GPU could potentially adopt Intel’s EMIB-T packaging, which could translate into an approximately $1 billion annual revenue opportunity for Intel Foundry.

$INTC

[Pppp Dz @pppp_dz]

@Xolani859243340 Because China doesn’t need starlink, every places you visit has 5G station in it. There is absolutely 0 revenue likely to be made anyway

[TBPN @tbpn]

FULL INTERVIEW: @dylan522p joins TBPN at the Cisco AI summit to discuss data centers in space, the limits of today’s AI hardware, and how chips, power, and geopolitics will shape the future of AI infra.

0:00 Space data centers and space tourism
01:03 AI chips in space and servicing challenges
02:39 Tesla FSD vs GPU reliability
05:18 Nvidia, Groq, and specialized AI chips
06:41 Cerebras and speed-first inference
09:55 xAI rumors and space compute bets
11:00 Starlink power and scaling limits
15:38 TSMC vs energy bottlenecks
27:02 Chips to China and geopolitics
30:32 Claude code and vibe coding
36:44 Meta, wearables, and the AI bull case

[Pppp Dz @pppp_dz]

@jukan05 Because Apple is a direct competitor, and that closed relationship that has locked in Qualcomm won’t happen here for Apple. Apple may go to Intel if capacity constrained or want to diversify which now they do. They go to Intel, which despite being a foundry,never sell beyond chip

[PyQuant News 🐍 @pyquantnews]

Algorithmic trading is the domain of secretive hedge funds and banks.

Python unlocked these secrets for everyone (even Goldman Sachs has an open-source tool).

Use the same tools the professionals use.

Here are 17 Python libraries that open the black box:

[Pppp Dz @pppp_dz]

@MCH2024 @insane_analyst If they can really outperform 18a in terms of both yield and performance, then Qualcomm will have no reason to be still be using TSMC, they will be forced to use Samsung as their primary foundry.

[Pppp Dz @pppp_dz]

@MCH2024 @insane_analyst what a joke? It’s better than 18a, and so it can attract only bitcoin miners? Lol

[Pppp Dz @pppp_dz]

@nav09284172 @Alex_Intel_ @jukan05 Not really @Alex_Intel_ have said clearly, the hyperscalers will pay whatever it is to bring up emergent fab capacity online, it doesn’t matter how the rest of the OEM want to work. Us government and hyperscalers will want to spend all they had to buy up capacity at Intel

[Pppp Dz @pppp_dz]

@BenBajarin @benitoz but TSMC never competes with its own customers. intel, being an IDM, will definitely do this deal

[Pppp Dz @pppp_dz]

@diamondrapids @WuMacus @Silicon_Fly @eding421 @Alex_Intel_ Before that, the trump administration need to get rid of pro TSMC voices. David Sachs is probably pro TSMC and lutnick is pro intel. What has happened in Biden admin that gave TSM the preferential treatment shall not happened again

[Pppp Dz @pppp_dz]

但是这里有一个问题就是过去任何一代m系列都是单工艺。 除非未来m系列转向chiplet。假如这条真如所说intel负责低阶m，tsm负责pro，max，跟ultra的话。这意味着intel理论上也可以做高档假如良率许可的情况下。因为是agnostic的，就像intel现在的cpu core tiles

[老猎户 @yao30059829]

Intc口径是2027实现foundry业务收支平衡，且只需“低到中个位数十亿美元”的外部收入的外部收入即可实现这一目标。
假定为50亿。Apple的M芯片+Google TPU封装几乎可以锁定这一目标。如图。
这将帮助intc在2027达成约65B的营收，EPS 2.7–2.9。
对应股价区间为30-70（现金流折现至当下）。
30已经是铁底。

[Pppp Dz @pppp_dz]

@FinansMerak137 What are you talking about? Aws has been a ifs packaging customer for many years now