Nvidia reportedly considered another copper-based design, called NVL72x2, as an alternative to Kyber. The system would have placed two Oberon racks back-to-back to expand the size of the NVLink scale-up domain without using optical interconnects. However, SemiAnalysis says customers rejected the unusual design and operational requirements, but does not specify their individual objections that could include serviceability, cooling, cabling, and data-center layout.
Meanwhile, the planned NVL576 rack scale solution that was supposed to combine eight Oberon racks interconnected using co-packaged optics between NVSwitches has also been postponed, or shipped in relatively small quantities because of 'ongoing CPO challenges,' SemiAnalysis claims.
The existence of the planned NVL576 configuration suggests that Nvidia had been developing some form of CPO-enabled NVSwitch connectivity for the Rubin generation. In theory, similar optical switch-to-switch connectivity could potentially be used to join smaller GPU groups into an NVL144 system and bypass Kyber's problematic copper midplane. However, the available information does not clearly indicate whether the CPO technology intended for NVL576 could reproduce Kyber's topology, bandwidth, and latency characteristics, or whether it was sufficiently mature for high-volume deployments by potential NVL144 customers.
The reported Kyber delay comes on the heels of another report saying that Nvidia had canceled quad-compute-chiplet version of its Rubin Ultra in favor or a dual-compute-chiplet design that is projected to deliver 2X lower performance. With Kyber NVL144 delayed and NVL72x2 cancelled, Nvidia will only be able to offer 72-way scale-up systems till sometimes in 2028, meaning that AMD and Google may end up with more competitive scale-up systems in 2027 – 2028. AMD's Mega Pod based on the Verano CPUs and Instinct MI500-series accelerators, is expected to pack up to 256 accelerators . Google's TPU 8i can provide roughly 1,024–1,152 accelerators within one low-latency domain , whereas the TPU 8t goes much further and can get to 9,600 chip packages per domain.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Anton Shilov is a contributing writer at Tom\u2019s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends. ","collapsible":{"enabled":true,"maxHeight":250,"readMoreText":"Read more","readLessText":"Read less"}}), "https://slice.vanilla.futurecdn.net/13-4-25/js/authorBio.js"); } else { console.error('%c FTE ','background: #9306F9; color: #ffffff','no lazy slice hydration function available'); } Anton Shilov Social Links Navigation Contributing Writer Anton Shilov is a contributing writer at Tom’s Hardware. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.
hotaru251 given the amount of $ spent on ram(of all types nvidia uses) if Jensen wasn't actually scared of the bubble popping he'd of invested in his own fab for memory by now as would have made them (Nvidia) more $ to make and produce it themselves & then also have that product for even after bubble pops. Reply
alan.campbell99 Giant amounts of Vera Rubin?! Most sold Blackwells are still yet to be lit up so are companies going to spend yet more money for Vera Rubins to join them in storage waiting for the 'warm shells' to be finished? Not sure why Jensen doesn't get asked about this or why it's not circular financing when he's funding customers buying some of his GPUs? Reply
3tank Oh- Im glad to hear the thing i can't buy that's used for bad things against me that I never asked for and preventing me from getting the things i do want to buy is hitting its production milestones. Reply
usertests hotaru251 said: if Jensen wasn't actually scared of the bubble popping he'd of invested in his own fab for memory by now They aren't going from being a fabless chip design house to having their own memory fab within a few years. Even if they fronted the cash for a brand new fab by Samsung/SK/Micron and signed an exclusivity deal, it would still take years to break ground. alan.campbell99 said: Giant amounts of Vera Rubin?! Most sold Blackwells are still yet to be lit up so are companies going to spend yet more money for Vera Rubins to join them in storage waiting for the 'warm shells' to be finished? Not sure why Jensen doesn't get asked about this or why it's not circular financing when he's funding customers buying some of his GPUs? Who knows, but one thing that could be done is to swap Blackwell cards or racks for Vera Rubin. So if you already have the datacenter facility and power budget, you get the shiny new card. The old Blackwell gets sold to feed orphans. 3tank said: Oh- Im glad to hear the thing i can't buy that's used for bad things against me that I never asked for and preventing me from getting the things i do want to buy is hitting its production milestones. Be less poor, and you can buy it. And the more you buy, the more you save! Reply
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/nvidias-huang-vows-to-deliver-giant-amounts-of-vera-rubin-company-says-that-our-roadmap-is-intact#main
- https://www.tomshardware.com/subscription
- US gov't allows Chinese telecom giant ZTE to purchase Nvidia H200 AI chips — firm joins Alibaba, Tencent, and ByteDance in access to Hopper tech
- Modder successfully runs Counter-Strike clone at 60 FPS on the original Sony PSP — created his own Rust-based 3D engine to power 480 x 272 gameplay, also works
- Developer successfully ports Linux to 1994 Sega 32X — Genesis and MegaDrive expansion runs open-source OS on paltry 23MHz processors and 256KB of RAM
- How Open Models Are Driving AI Research
- Drop's $130 desktop PC speakers are now just $23 — save a massive 82% on these dual-orientation BMR1 V2 speakers
Informational only. No financial advice. Do your own research.