#verarubin — Public Fediverse posts

TSMC shows smaller, faster chips without a pricey new tool from ASML

By Stephen Nellis and Max A. Cherney SANTA CLARA, California, April 22 (Reuters) – Taiwan Semiconductor Manufacturing Co…
#Netherlands #Nederland #NL #Europe #Europa #EU #ASML #California #EUV #GordonMoore #Nvidia #SANTACLARA #TaiwanSemiconductorManufacturingCo #TSMC #VeraRubin
https://www.europesays.com/netherlands/5120/

Professor Andjelka B. Kovačević’s video presentation for the IAU Symposium 404 is now online - “Multiscale Coherence for Astrobiology and Technosignatures with the Vera C. Rubin Observatory LSST” (Kovačević, Mason, Ciprijanović et al.).

https://www.youtube.com/watch?v=C23gezrXAWk #LSST #VeraRubin #Observatory #technosignatures #Astronomy #AndjelkaKovacevic

BERNSTEIN BETS BIG ON RUBIN: NVIDIA PRICE TARGET JUMP

Bernstein raises Nvidia's price target to $300, citing the new Vera Rubin AI platform expected in late 2026. See how it impacts chip performance and costs.

#Nvidia #VeraRubin #AI #TechStocks #Bernstein

https://newsletter.tf/nvidia-vera-rubin-platform-price-target-300/

Bernstein analysts predict Nvidia's stock could reach $300, driven by the upcoming Vera Rubin AI platform. This new platform is expected to be 5 times faster for inference tasks.

#Nvidia #VeraRubin #AI #TechStocks #Bernstein
https://newsletter.tf/nvidia-vera-rubin-platform-price-target-300/

https://winbuzzer.com/2026/04/15/microsoft-takes-over-stargate-data-center-openai-norway-xcxwbn/

Microsoft Takes Over Second Stargate Site as OpenAI Retreats

#AI #Microsoft #OpenAI #StargateProject #DataCenters #VeraRubin #Nscale #AIInfrastructure #BigTech #Europe #Norway

Meta、CoreWeaveと210億ドルのAIインフラ契約を締結 | Data Center Café https://www.yayafa.com/2778007/ #AgenticAi #AI #aiインフラ #ArtificialGeneralIntelligence #ArtificialIntelligence #coreweave #LLAMA #Meta #MetaAI #NVIDIA #VeraRubin #エージェント型AI #人工知能 #汎用人工知能 #資金調達

I primi dati dell'Osservatorio Rubin rivelano oltre 11.000 nuovi asteroidi

#asteroidi #veraRubin @astronomia

Gli scienziati del Rubin Observatory hanno inviato al IAU Minor Planet Center un insieme senza precedenti di rilevazioni di asteroidi, che include centinaia di mondi lontani oltre Nettuno e 33 asteroidi vicini alla Terra precedentemente sconosciuti.

https://umbertogaetani.substack.com/p/i-primi-dati-dellosservatorio-rubin

Price of #Nvidia's #VeraRubin #NVL72 racks skyrockets to $8.8M apiece, but #server makers' margins will be tight — Nvidia is moving closer to shipping entire full-scale systems
#Blackwell #NVL72 #rackscale system costs $2.8 – $3.4M for an #AI training and #HPC NVL72 #GB200 and $6M to $6.5M for an AI inference NVL72 #GB300
Vera Rubin NVL72 #VR200 systems are currently quoted at $5M - $7M per unit.
Nvidia has never confirmed the list prices of its NVL72 or #NVL144 products.
https://www.tomshardware.com/tech-industry/artificial-intelligence/price-of-nvidias-vera-rubin-nvl72-racks-skyrockets-to-as-much-as-usd8-8-million-apiece-but-server-makers-margins-will-be-tight-nvidia-is-moving-closer-to-shipping-entire-full-scale-systems

Price of #Nvidia's #VeraRubin #NVL72 racks skyrockets to $8.8M apiece, but #server makers' margins will be tight — Nvidia is moving closer to shipping entire full-scale systems
#Blackwell #NVL72 #rackscale system costs $2.8 – $3.4M for an #AI training and #HPC NVL72 #GB200 and $6M to $6.5M for an AI inference NVL72 #GB300
Vera Rubin NVL72 #VR200 systems are currently quoted at $5M - $7M per unit.
Nvidia has never confirmed the list prices of its NVL72 or #NVL144 products.
https://www.tomshardware.com/tech-industry/artificial-intelligence/price-of-nvidias-vera-rubin-nvl72-racks-skyrockets-to-as-much-as-usd8-8-million-apiece-but-server-makers-margins-will-be-tight-nvidia-is-moving-closer-to-shipping-entire-full-scale-systems

Price of #Nvidia's #VeraRubin #NVL72 racks skyrockets to $8.8M apiece, but #server makers' margins will be tight — Nvidia is moving closer to shipping entire full-scale systems
#Blackwell #NVL72 #rackscale system costs $2.8 – $3.4M for an #AI training and #HPC NVL72 #GB200 and $6M to $6.5M for an AI inference NVL72 #GB300
Vera Rubin NVL72 #VR200 systems are currently quoted at $5M - $7M per unit.
Nvidia has never confirmed the list prices of its NVL72 or #NVL144 products.
https://www.tomshardware.com/tech-industry/artificial-intelligence/price-of-nvidias-vera-rubin-nvl72-racks-skyrockets-to-as-much-as-usd8-8-million-apiece-but-server-makers-margins-will-be-tight-nvidia-is-moving-closer-to-shipping-entire-full-scale-systems

Price of #Nvidia's #VeraRubin #NVL72 racks skyrockets to $8.8M apiece, but #server makers' margins will be tight — Nvidia is moving closer to shipping entire full-scale systems
#Blackwell #NVL72 #rackscale system costs $2.8 – $3.4M for an #AI training and #HPC NVL72 #GB200 and $6M to $6.5M for an AI inference NVL72 #GB300
Vera Rubin NVL72 #VR200 systems are currently quoted at $5M - $7M per unit.
Nvidia has never confirmed the list prices of its NVL72 or #NVL144 products.
https://www.tomshardware.com/tech-industry/artificial-intelligence/price-of-nvidias-vera-rubin-nvl72-racks-skyrockets-to-as-much-as-usd8-8-million-apiece-but-server-makers-margins-will-be-tight-nvidia-is-moving-closer-to-shipping-entire-full-scale-systems

Price of #Nvidia's #VeraRubin #NVL72 racks skyrockets to $8.8M apiece, but #server makers' margins will be tight — Nvidia is moving closer to shipping entire full-scale systems
#Blackwell #NVL72 #rackscale system costs $2.8 – $3.4M for an #AI training and #HPC NVL72 #GB200 and $6M to $6.5M for an AI inference NVL72 #GB300
Vera Rubin NVL72 #VR200 systems are currently quoted at $5M - $7M per unit.
Nvidia has never confirmed the list prices of its NVL72 or #NVL144 products.
https://www.tomshardware.com/tech-industry/artificial-intelligence/price-of-nvidias-vera-rubin-nvl72-racks-skyrockets-to-as-much-as-usd8-8-million-apiece-but-server-makers-margins-will-be-tight-nvidia-is-moving-closer-to-shipping-entire-full-scale-systems

https://winbuzzer.com/2026/03/18/microsoft-foundry-agent-service-vera-rubin-nvl72-gtc-2026-xcxwbn/

Microsoft First to Power On NVIDIA Vera Rubin NVL72 GPUs

#AI #Azure #Microsoft #NVIDIA #CloudComputing #GPUs #VeraRubin #FoundryAgentService #AIAgents #AgenticAI #BigTech #Datacenters #AIInfrastructure

https://winbuzzer.com/2026/03/18/microsoft-foundry-agent-service-vera-rubin-nvl72-gtc-2026-xcxwbn/

Microsoft First to Power On NVIDIA Vera Rubin NVL72 GPUs

#AI #Azure #Microsoft #NVIDIA #CloudComputing #GPUs #VeraRubin #FoundryAgentService #AIAgents #AgenticAI #BigTech #Datacenters #AIInfrastructure

https://winbuzzer.com/2026/03/18/microsoft-foundry-agent-service-vera-rubin-nvl72-gtc-2026-xcxwbn/

Microsoft First to Power On NVIDIA Vera Rubin NVL72 GPUs

#AI #Azure #Microsoft #NVIDIA #CloudComputing #GPUs #VeraRubin #FoundryAgentService #AIAgents #AgenticAI #BigTech #Datacenters #AIInfrastructure

https://winbuzzer.com/2026/03/18/microsoft-foundry-agent-service-vera-rubin-nvl72-gtc-2026-xcxwbn/

Microsoft First to Power On NVIDIA Vera Rubin NVL72 GPUs

#AI #Azure #Microsoft #NVIDIA #CloudComputing #GPUs #VeraRubin #FoundryAgentService #AIAgents #AgenticAI #BigTech #Datacenters #AIInfrastructure

https://winbuzzer.com/2026/03/18/microsoft-foundry-agent-service-vera-rubin-nvl72-gtc-2026-xcxwbn/

Microsoft First to Power On NVIDIA Vera Rubin NVL72 GPUs

#AI #Azure #Microsoft #NVIDIA #CloudComputing #GPUs #VeraRubin #FoundryAgentService #AIAgents #AgenticAI #BigTech #Datacenters #AIInfrastructure

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

NVIDIA’s new Vera Rubin platform brings together specialized chips (Vera CPUs, Rubin GPUs, Groq LPUs, and BlueField-4 DPUs) into coordinated, rack-scale systems designed for real-time AI.

The big shift: AI isn’t just about training models anymore — it’s about orchestrating entire systems to power intelligent, autonomous agents in real time.
https://www.buysellram.com/blog/the-agentic-ai-era-how-nvidia-rubin-vera-cpu-groq-3-lpus-bluefield-4-redefine-the-inference-factory/
#NVIDIAGTC #AgenticAI #VeraRubin #DataCenter #GPU #InferenceFactory #AIInfrastructure #Groq #NVIDIA #NVLink #AIHardware #technology

https://winbuzzer.com/2026/03/17/samsung-hbm4e-nvidia-gtc-2026-sk-hynix-supply-race-xcxwbn/

Samsung Unveils HBM4E AI Memory Chips at GTC 2026 in SK Hynix Supply Race

#AI #NVIDIA #Samsung #SKHynix #Semiconductors #Memory Groq #BigTech #Chipmakers #Inference #VeraRubin #HBM4 #HBM4E #GTC2026

https://winbuzzer.com/2026/03/17/samsung-hbm4e-nvidia-gtc-2026-sk-hynix-supply-race-xcxwbn/

Samsung Unveils HBM4E AI Memory Chips at GTC 2026 in SK Hynix Supply Race

#AI #NVIDIA #Samsung #SKHynix #Semiconductors #Memory Groq #BigTech #Chipmakers #Inference #VeraRubin #HBM4 #HBM4E #GTC2026

https://winbuzzer.com/2026/03/17/samsung-hbm4e-nvidia-gtc-2026-sk-hynix-supply-race-xcxwbn/

Samsung Unveils HBM4E AI Memory Chips at GTC 2026 in SK Hynix Supply Race

#AI #NVIDIA #Samsung #SKHynix #Semiconductors #Memory Groq #BigTech #Chipmakers #Inference #VeraRubin #HBM4 #HBM4E #GTC2026

https://winbuzzer.com/2026/03/17/samsung-hbm4e-nvidia-gtc-2026-sk-hynix-supply-race-xcxwbn/

Samsung Unveils HBM4E AI Memory Chips at GTC 2026 in SK Hynix Supply Race

#AI #NVIDIA #Samsung #SKHynix #Semiconductors #Memory Groq #BigTech #Chipmakers #Inference #VeraRubin #HBM4 #HBM4E #GTC2026

https://winbuzzer.com/2026/03/17/samsung-hbm4e-nvidia-gtc-2026-sk-hynix-supply-race-xcxwbn/

Samsung Unveils HBM4E AI Memory Chips at GTC 2026 in SK Hynix Supply Race

#AI #NVIDIA #Samsung #SKHynix #Semiconductors #Memory Groq #BigTech #Chipmakers #Inference #VeraRubin #HBM4 #HBM4E #GTC2026

https://winbuzzer.com/2026/03/17/nvidia-vera-rubin-space-1-orbital-ai-data-centers-xcxwbn/

Nvidia Unveils Space-1 Chip for Orbital AI Data Centers

#NVIDIA #AI #VeraRubinSpace1 #DataCenters #GPUs #Semiconductors #JensenHuang #SpaceTech #AIChips #AIInfrastructure #VeraRubin #PlanetLabs #Chipmakers #AIPerformance #OrbitalComputing

At #Nvidia’s GTC 2026, CEO #JensenHuang announced expectations for $1 trillion in orders for #Blackwell and #VeraRubin chips through 2027, driven by surging demand for #AI processing power. Nvidia also unveiled the #Groq3 Language Processing Unit, designed to enhance its AI capabilities, and provided a glimpse of the Kyber architecture, expected in 2027. https://www.cnbc.com/2026/03/16/nvidia-gtc-2026-ceo-jensen-huang-keynote-blackwell-vera-rubin.html?eicker.news #tech #media #news

via #Microsoft : Microsoft at NVIDIA GTC: New solutions for Microsoft Foundry, Azure AI infrastructure and Physical AI

https://ift.tt/vjnXd7P
#Microsoft #NVIDIA #GTC2026 #Foundry #AzureAI #AIinfrastructure #PhysicalAI #NVIDIAOmniverse #Nemotron #VeraRubin #NVL72 #AzureDatacen…

https://winbuzzer.com/2026/03/16/hon-hai-q4-profit-miss-ai-server-growth-2026-xcxwbn/

Nvida Partner Foxconn's Q4 Profit Misses but AI Server Boom Fuels 2026 Outlook

#AI #Foxconn #NVIDIA #BigTech #Semiconductors #DataCenters #CloudComputing #Manufacturing #Server #StargateProject #SupplyChain #YoungLiu #VeraRubin

Jensen Huang’s message to the DRAM market is clear: "Whatever you can build, we will take." With the Vera Rubin platform on the horizon and HBM4 demand hitting fever pitch, we are entering a structural "AI Memory Supercycle" that will redefine data center ROI through 2027.

https://www.buysellram.com/blog/dram-shortage-wont-end-in-2026-why-nvidia-just-told-fabs-to-build-more-well-buy-it-all/

#ArtificialIntelligence #DRAM #DRAMShortage #HBM #Semiconductors #DataCenter #AIInfrastructure #NVIDIA #VeraRubin #JensenHuang #TechStrategy #GPU #RAM #technology

https://winbuzzer.com/2026/02/18/meta-deepens-nvidia-partnership-multi-billion-chip-deal-xcxwbn/

Meta to Deploy Millions of Nvidia Chips in Multi-Billion Deal

#AI #AIChips #MetaInc #NVIDIA #BigTech #Semiconductors #Processors #AIInfrastructure #DataCenters #NvidiaBlackwell #VeraRubin

https://www.europesays.com/pl/191972/ NVIDIA Vera Rubin Superchip – omówienie platformy AI o pięciokrotnie wyższej wydajności niż Grace Blackwell #ai #Biznes #bluefield‑4 #budowa #Business #CES2026 #connectx‑9 #cpu #GPU #hbm4 #informacje #lpddr5x #nvidia #nvl72 #nvlink 6 #omówienie #opis #PL #Poland #Polish #Polska #Polski #premiera #specyfikacja #superchip #SztucznaInteligencja #targi #VeraRubin

#Nvidia’s #VeraRubin Platform Obsoletes Current AI Iron Six Months Ahead Of Launch
Nvidia’s top brass says Vera-Rubin #NVL72 #rackscale system, with 72 GPU sockets and 36 CPU sockets as well as NVSwitch fabric linking the, delivers a 10X reduction in #inference cost per token for MoE #AI models and 4X reduction in number of GPUs (but maybe not a 4X reduction in cost, be careful) to train those models compared to the prior generation of #GraceBlackwell NVL72 systems.
https://www.nextplatform.com/2026/01/05/nvidias-vera-rubin-platform-obsoletes-current-ai-iron-six-months-ahead-of-launch/

NVIDIA unveils its next-generation Vera Rubin superchip at CES 2026, entering full production and promising to accelerate AI adoption by slashing costs and boosting efficiency for major cloud providers.
#YonhapInfomax #NVIDIA #VeraRubin #Superchip #AICostReduction #CloudProviders #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=98596

NVIDIA unveils its next-generation Vera Rubin superchip at CES 2026, entering full production and promising to accelerate AI adoption by slashing costs and boosting efficiency for major cloud providers.
#YonhapInfomax #NVIDIA #VeraRubin #Superchip #AICostReduction #CloudProviders #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=98596

NVIDIA unveils its next-generation Vera Rubin superchip at CES 2026, entering full production and promising to accelerate AI adoption by slashing costs and boosting efficiency for major cloud providers.
#YonhapInfomax #NVIDIA #VeraRubin #Superchip #AICostReduction #CloudProviders #Economics #FinancialMarkets #Banking #Securities #Bonds #StockMarket
https://en.infomaxai.com/news/articleView.html?idxno=98596

Sojusz tytanów AI. Microsoft, Nvidia i Anthropic wchodzą w układ za 30 mld dolarów. Koniec przewagi OpenAI?

Mamy strategiczne trzęsienie ziemi w Dolinie Krzemowej. Microsoft, dotychczas kojarzony jako główny patron OpenAI, właśnie ogłosił potężny, trójstronny sojusz z firmą Anthropic (twórcami Claude’a) oraz Nvidią.

Stawką jest 30 miliardów dolarów, dostęp do najnowszych chipów i… obecność modelu Claude w usługach Microsoftu.

Jak donosi serwis AI News, umowa ta sygnalizuje koniec ery „zależności od jednego modelu”. Microsoft oficjalnie dywersyfikuje swoje portfolio, otwierając się na największego konkurenta ChataGPT.

30 miliardów na infrastrukturę przyszłości

Skala tego porozumienia jest gigantyczna. Anthropic zobowiązał się do zakupu mocy obliczeniowej w chmurze Azure za kwotę 30 miliardów dolarów.

W zamian otrzyma priorytetowy dostęp do najnowszego sprzętu. Plan zakłada wykorzystanie systemów NVIDIA Grace Blackwell, a w dalszej perspektywie – nowej architektury Vera Rubin. Jensen Huang, CEO Nvidii, zapowiada „skokową poprawę prędkości”, co ma kluczowe znaczenie dla obniżenia kosztów działania sztucznej inteligencji.

Claude wchodzi do Copilota

Dla zwykłego użytkownika najważniejsza jest inna informacja. Satya Nadella, CEO Microsoftu, potwierdził, że jest to integracja obustronna. Oznacza to, że modele Anthropic (rodzina Claude) zostaną zintegrowane z produktami Microsoftu, w tym z rodziną Copilot.

To rewolucja. Użytkownicy ekosystemu Microsoft 365 nie będą już skazani wyłącznie na silniki GPT od OpenAI. Będą mogli korzystać z Claude’a – modelu, który wielu programistów i twórców uważa za lepszy w kodowaniu i kreatywnym pisaniu.

Dlaczego to ważne? „Agentic AI” i koszty myślenia

Sojusz kładzie ogromny nacisk na „sztuczną inteligencję agentową” (Agentic AI), którą Microsoft epatuje choćby na najnowszym wydarzeniu Microsoft Ignite 2025, o którym pisaliśmy. Jensen Huang podkreślił, że protokół MCP (Model Context Protocol) od Anthropic zrewolucjonizował ten obszar, a inżynierowie Nvidii już używają Claude Code do refaktoryzacji starych baz kodu.

Raport zwraca też uwagę na zmianę w ekonomii AI. Koszty przesuwają się z trenowania na wnioskowanie (inference). Nowe modele „myślą” dłużej przed udzieleniem odpowiedzi (test-time scaling), co sprawia, że koszty nie są już płaską stawką za token, ale zależą od „głębokości myślenia” danego modelu.

Dla Anthropic to rozwiązanie problemu dotarcia do klienta korporacyjnego – „wskakują na plecy” Microsoftu i zyskują dostęp do milionów firm. Dla Microsoftu to bezpiecznik – jeśli OpenAI powinie się noga, mają pod skrzydłami ich najgroźniejszego rywala. Trudno odmówić Microsoftowi strategicznego podejścia.

Microsoft Ignite 2025: koniec gadania, czas na działanie. Windows w chmurze w małym pudełku i AI, które widzi Twój ekran

#anthropic #azure #claude #graceBlackwell #microsoft #news #nvidia #openai #sojuszAi #veraRubin

Cooling system for a single #Nvidia #Blackwell Ultra #NVL72 rack costs a staggering $50,000 — set to increase to $56,000 with next-generation #NVL144 racks
Thermals will grow further in the #VeraRubin NVL144 platform that will rely on hotter Vera CPUs, Rubin GPUs (up to 1,800W per unit), and next-generation NVSwitch 6.0 ASICs, bringing the cost of the rack-scale cooling system to $55,710, up 17% compared to the cooling system for the GB300, according to Morgan Stanley.
https://www.tomshardware.com/pc-components/cooling/cooling-system-for-a-single-nvidia-blackwell-ultra-nvl72-rack-costs-a-staggering-usd50-000-set-to-increase-to-usd56-000-with-next-generation-nvl144-racks

https://www.europesays.com/pl/74217/ NVIDIA Vera Rubin Superchip – prezentacja prototypu wielkości płyty głównej, który wejdzie do produkcji już w przyszłym roku #ai #AkceleratorAI #arm #cpu #GPU #GtcWashington #JensenHuang #konferencja #Nauka #NaukaITechnika #NaukaTechnika #nvidia #PL #Poland #Polish #Polska #Polski #prezentacja #prototyp #Science #ScienceAndTechnology #ScienceTechnology #serwer #superchip #SztucznaInteligencja #Technika #Technology #VeraRubin

Nowy „układ nerwowy” dla AI. Meta i Oracle stawiają na NVIDIA Spectrum-X

Meta i Oracle ogłosiły, że wybrały platformę sieciową NVIDIA Spectrum-X do modernizacji swoich centrów danych przystosowanych do zadań sztucznej inteligencji.

To strategiczna decyzja, która ma na celu sprostanie gigantycznym wymaganiom obliczeniowym modeli AI nowej generacji, operujących na bilionach parametrów.

Podczas gdy większość uwagi w wyścigu AI skupia się na mocy obliczeniowej procesorów GPU, prawdziwym wąskim gardłem w budowie tzw. „fabryk AI” staje się sieć, która musi połączyć miliony tych procesorów. Według NVIDII, platforma Spectrum-X została zaprojektowana specjalnie, by pełnić funkcję „układu nerwowego” w tych systemach, zapewniając bezprecedensową wydajność.

Koniec Messengera na Macu. Meta wygasza aplikację i prosi o zapisanie historii czatów

Tradycyjny Ethernet, używany w wciąż w wielu miejscach świata, także w centrach danych, często osiąga efektywną przepustowość na poziomie około 60% z powodu kolizji i zatorów. Platforma Spectrum-X, dzięki technologiom takim jak adaptacyjny routing i kontrola przeciążeń oparta na telemetrii, ma osiągać efektywność na poziomie 95%. Pozwala to na znacznie szybsze trenowanie modeli AI i jednoczesne uruchamianie wielu zadań bez wzajemnego zakłócania się.

Meta nie boi się bańki AI. Zuckerberg gotów jest „zmarnować” setki miliardów dolarów

Meta zintegruje przełączniki Spectrum-X ze swoim autorskim, otwartym systemem zarządzania siecią FBOSS (Facebook Open Switching System). Dla firmy, która musi obsługiwać miliardy użytkowników i trenować coraz większe modele AI, kluczowa jest otwartość i wydajność nowej infrastruktury sieciowej.

Z kolei Oracle planuje wykorzystać Spectrum-X w ramach Oracle Cloud Infrastructure do budowy wielkoskalowych fabryk AI. Firma zapowiedziała, że jej infrastruktura będzie również wspierać nadchodzącą architekturę NVIDII o nazwie kodowej Vera Rubin, co pozwoli klientom chmury Oracle na jeszcze szybsze wdrażanie nowych modeli AI. Wybór ten potwierdza rosnące znaczenie otwartych standardów sieciowych.

Cała inwestycja NVIDII w Spectrum-X jest już przygotowaniem pod wspomnianą kolejną generację architektury, czyli Vera Rubin, która ma być dostępna komercyjnie w drugiej połowie 2026 roku. Platforma sieciowa musi być gotowa, zanim na rynek trafią jeszcze potężniejsze procesory obliczeniowe.

OpenAI i Oracle budują superkomputer Stargate. Będzie miał moc wielkiego miasta

#AI #centraDanych #ethernet #fabrykiAI #Meta #news #nvidia #oracle #sieć #SpectrumX #sztucznaInteligencja #VeraRubin

#Intel Announces "#CrescentIsland" #Inference-Optimized #Xe3P #GPU With 160GB vRAM, but it's not shipping until at least H2'2026
Given the timing this will be going up against the AMD Instinct #MI450 series and NVIDIA #VeraRubin.
This new graphics card is built on next-gen Xe3P Celestial micro-architecture. Xe3P will be optimized around performance-per-Watt and Crescent Island will feature 160GB of LPDDR5x memory to allow plenty of space for #largelanguagemodels (#LLM).
https://www.phoronix.com/review/intel-crescent-island

First Ever Images from the Vera Rubin Telescope [Chile]
--
https://youtu.be/tu4cpaDZ68g?si=FaeVBYNRYIbEfqiU <-- shared technical video
--
https://rubinobservatory.org/ <-- Rubin Observatory home page
--
“The Vera C. Rubin telescope has just [mid-2025] released its first images, and they’re spectacular…
On a mountain in Chile, the Vera C. Rubin Observatory, with its massive 3.2-billion-pixel camera, is capturing images that rival those from the James Webb Space Telescope. Get ready to see some of the most stunning astronomical images ever taken as [they] explore how this groundbreaking project will help us unlock the secrets of the universe. This is just the beginning...”
#verarubin #fedscience #fedservice #Coquimbo #Chile #LegacySurveyofSpaceandTime #LSST #Andes #telescope #remotesensing #image #imagery #astronomy #astronomical #observation #spatialanalysis #spatiotemporal #extraterrestrial #research
@NSF @DOE @SLAC

Vera Rubin at Mayall Telescope

Deidre Hunter (left), Katy Garmany (center), and Vera Rubin (right) at the Nicholas U. Mayall 4-meter Telescope around 2000.

Credit: NOIRLab/NSF/AURA
https://noirlab.edu/public/images/VeraRubin_DeidreHunter_KatyGarmany-CC/

#VeraRubin #Astronomer #Astronomy #Rubin #DeidreHunter #KatyGarmany #science #STEM #WomeninSTEM #WomeninScience #Astrodon #KittPeak #Arizona #Xterminal #Xterminals #xterm

Vera Rubin at Mayall Telescope

Deidre Hunter (left), Katy Garmany (center), and Vera Rubin (right) at the Nicholas U. Mayall 4-meter Telescope around 2000.

Credit: NOIRLab/NSF/AURA
https://noirlab.edu/public/images/VeraRubin_DeidreHunter_KatyGarmany-CC/

#VeraRubin #Astronomer #Astronomy #Rubin #DeidreHunter #KatyGarmany #science #STEM #WomeninSTEM #WomeninScience #Astrodon #KittPeak #Arizona #Xterminal #Xterminals #xterm

Vera Rubin at Mayall Telescope

Deidre Hunter (left), Katy Garmany (center), and Vera Rubin (right) at the Nicholas U. Mayall 4-meter Telescope around 2000.

Credit: NOIRLab/NSF/AURA
https://noirlab.edu/public/images/VeraRubin_DeidreHunter_KatyGarmany-CC/

#VeraRubin #Astronomer #Astronomy #Rubin #DeidreHunter #KatyGarmany #science #STEM #WomeninSTEM #WomeninScience #Astrodon #KittPeak #Arizona #Xterminal #Xterminals #xterm

Vera Rubin at Mayall Telescope

Deidre Hunter (left), Katy Garmany (center), and Vera Rubin (right) at the Nicholas U. Mayall 4-meter Telescope around 2000.

Credit: NOIRLab/NSF/AURA
https://noirlab.edu/public/images/VeraRubin_DeidreHunter_KatyGarmany-CC/

#VeraRubin #Astronomer #Astronomy #Rubin #DeidreHunter #KatyGarmany #science #STEM #WomeninSTEM #WomeninScience #Astrodon #KittPeak #Arizona #Xterminal #Xterminals #xterm

Vera Rubin at Mayall Telescope

Deidre Hunter (left), Katy Garmany (center), and Vera Rubin (right) at the Nicholas U. Mayall 4-meter Telescope around 2000.

Credit: NOIRLab/NSF/AURA
https://noirlab.edu/public/images/VeraRubin_DeidreHunter_KatyGarmany-CC/

#VeraRubin #Astronomer #Astronomy #Rubin #DeidreHunter #KatyGarmany #science #STEM #WomeninSTEM #WomeninScience #Astrodon #KittPeak #Arizona #Xterminal #Xterminals #xterm

Vera Rubin: New Image of Andromeda (M31) Released in Tribute 
Study of Andromeda by Vera Rubin enriched our knowledge of galaxies, led to discovery of dark matter and transformed understanding of the universe............
#Andromeda #darkmatter #galaxy #homegalaxy #M31 #milkyway #NASA #VeraRubin
Umesh Prasad

https://www.scientificeuropean.co.uk/sciences/personality/vera-rubin-new-image-of-andromeda-m31-released-in-tribute/