Editor at https://xitter.cf/t.co/lkBc9TVsoG
Joined June 2008
- Tweets15.4K
- Following318
- Followers17.6K
- Likes4.2K
An EUV laser prototype (the FlintNano LPP-OXXO from Langdao Sliex) was showcased at a trade show in China. However, its performance parameters are still far from meeting the requirements of modern lithography systems.
hardwareluxx.de/index.php/ne…
.@JensenHuang reportedly criticized Germany for overthinking rather than acting decisively, arguing that the country is failing to capitalize on its technological expertise and talent.
Fireside Talk with @_FriedrichMerz and Jensen: youtube.com/live/PCroCEAGQ7g
One thing to add to this back-and-forth and Jeff Fischer's statement: While we were waiting for NVIDIA to publish a statement last week, someone from NVIDIA told me that he doesn't think there will be an official statement on this.
I'm shocked (irony off)
SHOCKING: We have confirmed that multiple Neoclouds have been refused access to GTC Berlin due to them adding AMD GPUs, TPUs, and other chips to their cloud offerings
Note that this doesn't apply to hyperscalers, as they have more buying power. But for neoclouds, executives feel NVIDIA has been using high-pressure tactics to keep them NVIDIA-only.
As such, neocloud executives are starting to consider offering TPUs or AMD GPUs quietly to avoid getting pressured by NVIDIA and to avoid the feeling that NVIDIA will retaliate.
xitter.cf/SemiAnalysis_/status/2…
And, as I do every three months: my chart showing the TSMC revenue by technology.
N2 and N3 taking 41 % now, N5 is shrinking from 33 to 28 %. TSMC exposed 4.476 million wafers over the past three months (+10% YoY).
My review of the GB10/N1X aka @NVIDIARTXSpark with 128 GB unified memory inside the @Microsoft @surface Laptop Ultra - from Gaming to Agentic and behond.
hardwareluxx.de/index.php/ar…
Bolt Graphics Zeus test chip. RISC-V based GPU architecture with a focus on Path Tracing and FP64. A0 silicon is a 3 mm² chip with 1.4 watts build on TSMC 12 nm.
Its still a long way until planned general availability in 2028 can be called.
Our test chip is up and running!!!!!
We’ll be sharing more later tonight at the party 😄 Thanks to everyone here for all the support since we announced. 💛 Now we celebrate!
You can read the full press release here: prnewswire.com/news-releases…
Ah and for the final chip @boltgraphics is going with TSMC's N5P, to avoid competing for wafer allocation.
Am I seeing this correctly? In the technical specifications for the M26-NVLS-ATX09, Decenta also lists the AI computing power for the CPU, NPU, and GPU. Does that refer to Nova Lake? It implies a boost for CPU and NPU performance.
hardwareluxx.de/index.php/ne…
With the rumor that surfaced over the weekend that NVIDIA had marked the GeForce RTX 5090 as end-of-life, I reached out to NVIDIA for an official statement. So far, nothing but silence!
Not to mention that the GeForce RTX 5090 is unattractive to gamers anyway due to its price…
Intel GPU D-series, code-named Crescent Island
- 480 GB LPDDR5X memory
- 1.5 TB/s of memory bandwidth
- 1.3 petaflops of FP8
- 350 W TDP
hardwareluxx.de/index.php/ne…
Microsoft Maia 300 chiplet configuration:
- 1x Main Compute Die (TSMC N2P)
- 2x I/O Die
- 4x Memory Die
- 8x HBM4 Stacks (288 GB)
- TDP: 950 W
AMD Helios: Hyperscaler Disaggregated Agentic Rack
AMD Helion: All-in-One Agentic Rack
Different hyperscale configurations: Personal, Enterprise, Technical Agent
Different enterprise configurations: Agentic Node, Inference Node, General Purpose Node
hardwareluxx.de/index.php/ne…
AMDs statement on the VBIOS/firmware update for Radeon RX 7000 Series.
hardwareluxx.de/index.php/ne…
🤖 Made with AI
GlobalFoundries did announced a manufacturing agreement with TSMC. They plan to establish a US based supply of silicon interposers for TSMC’s CoWoS advanced packaging ecosystem. GF Malta fab will get a capacity extension.
hardwareluxx.de/index.php/ne…
