● Covered by 4 sources · 10 reportsMedium impact1 negative4 neutral
NVIDIA Unveils Vera CPU for AI-centric Workloads, Challenging AMD and Intel
Key points
- NVIDIA introduced the Vera CPU for AI-focused data centers.
- Vera CPU is designed for high single-threaded performance.
- Targets AI inference workloads, challenging AMD and Intel.
- Vera CPUs delivered to clients like OpenAI and SpaceX.
- Rivals x86 CPUs in performance benchmarks.
- Vera CPU is a new category of CPUs built for the agentic AI era.
- Nvidia's next-gen Rigel Arm v9.2 CPU core will ship as part of its Rosa CPU.
- Rigel core will deliver higher per-core performance than Vera's Olympus core.
- Vera is Nvidia's first CPU with a custom core design.
- Vera is on track for general availability in the back half of this year.
- Nvidia released a Vera white paper with SPEC CPU 2026 results.
- Vera chips were delivered to clients, including Anthropic, in June.
- NVIDIA Vera Rubin platform is ramping up production with partners.
- Vera Rubin NVL72 production is spanning 350-plus factory sites in 30 countries.
- Vera Rubin platform delivers 10x more throughput per megawatt than Grace Blackwell NVL72.
- Vera Rubin platform is built from seven chips and five rack trays.
- Vera CPU's Olympus core delivers 2x single-threaded performance.
- AMD's Zen 6 'Venice' CPUs show a 20% per-core performance advantage over Vera.
- AMD's Venice CPUs show 2.2x higher throughput compared to Vera.
- NVIDIA is deploying Vera CPU internally to optimize electronic design automation (EDA) workflows.
- NVIDIA is collaborating with Cadence and Synopsys to optimize EDA applications for Vera CPU.
- Vera CPU shows up to 1.5x performance improvement on selected EDA tasks.
NVIDIA Launches Vera CPU
NVIDIA has introduced the Vera CPU aimed at AI server environments, focusing on AI inference workloads where single-threaded speed is critical. NVIDIA revealed this new CPU as a direct challenge to the traditional CPU giants, AMD and Intel, by targeting AI-centric data centers.
Design and Performance
The Vera CPU is built on NVIDIA's custom Olympus core architecture, optimized for single-threaded performance. NVIDIA frames it as a 'max single-threaded CPU at scale', rather than focusing on parallel processing, providing strong performance per core. This design is a shift from the high core-count CPUs typically used in data centers.
Market Impact
Delivered to prominent clients such as OpenAI and SpaceX, the Vera CPU comes as part of NVIDIA’s strategy to vertically integrate and capture a larger share of the expanding AI server market. As companies expand AI infrastructures, NVIDIA aims to position itself against established players Intel and AMD.
Benchmark Comparisons
NVIDIA disclosed benchmark results comparing Vera to AMD's Epyc CPUs, showing significant performance advantages in certain workloads. These comparisons highlight Vera's potential to advance AI data center efficiency, particularly in power-constrained scenarios.
Industry Implications
NVIDIA's Vera CPU marks a significant advancement in AI infrastructure architecture and signifies a competitive challenge to established CPU makers. The chip's optimization for AI workloads represents an important shift in how data centers might approach processor selection in the AI era.
Updates
- NVIDIA is deploying Vera CPU internally to optimize electronic design automation (EDA) workflows.
- NVIDIA is collaborating with Cadence and Synopsys to optimize EDA applications for Vera CPU.
- Vera CPU shows up to 1.5x performance improvement on selected EDA tasks.
- Vera CPU is a new category of CPUs built for the agentic AI era.
- Nvidia's next-gen Rigel Arm v9.2 CPU core will ship as part of its Rosa CPU.
- Rigel core will deliver higher per-core performance than Vera's Olympus core.
- Vera is Nvidia's first CPU with a custom core design.
- Vera is on track for general availability in the back half of this year.
- Nvidia released a Vera white paper with SPEC CPU 2026 results.
- Vera chips were delivered to clients, including Anthropic, in June.
- NVIDIA Vera Rubin platform is ramping up production with partners.
- Vera Rubin NVL72 production is spanning 350-plus factory sites in 30 countries.
- Vera Rubin platform delivers 10x more throughput per megawatt than Grace Blackwell NVL72.
- Vera Rubin platform is built from seven chips and five rack trays.
- Vera CPU's Olympus core delivers 2x single-threaded performance.
- AMD's Zen 6 'Venice' CPUs show a 20% per-core performance advantage over Vera.
- AMD's Venice CPUs show 2.2x higher throughput compared to Vera.
✨ This summary was generated by AI from the outlets' reporting listed below. It is not independently verified and may contain errors — check the original sources. How BrevFeed works →
The daily brief
One email each morning: the day's tech stories, clustered across outlets and summarized. No account needed.
One email a day. Unsubscribe in one click, any time.
Today's brief
Spend a few minutes, get the whole day. Every topic's top stories in one hands-free rundown — listen, watch, or read the transcript.
▶ Play today's brief
New every morning, and the back catalogue is archived by date.
Primary sources
GitHub dsheffie/mem-latHow outlets covered it
NVIDIA released a 45-page whitepaper detailing Vera, its first server CPU with an 88-core Olympus chip, but the paper's marketing claims are criticized for misrepresenting technical aspects. The Olympus core itself is noted as genuinely formidable, making the marketing narrative unnecessary.
Elon Musk announced that SpaceX and xAI will exclusively use Nvidia GPUs, specifically the Vera Rubin NVL72 architecture, for their AI computing needs. This decision is based on the perceived superior design of the NVL72 system, and SpaceX plans to deploy an optimized version of this system in space starting next year.
NVIDIA announced new storage advancements and open-sourced its cuFile APIs at the Future of Memory and Storage (FMS) conference. These developments aim to address the increasing demands of AI workloads on data storage infrastructure by enabling GPUs to directly access storage and improving data processing efficiency.
NVIDIA is deploying its Vera CPU to optimize electronic design automation (EDA) workflows for developing its next-generation CPUs and GPUs. This internal adoption, in collaboration with Cadence and Synopsys, demonstrates how high-performance CPU architecture can accelerate demanding engineering workloads in chip design, with initial tests showing up to 1.5x performance improvement on selected tasks.
AMD conducted its own SPEC CPU 2026 benchmarks comparing its Zen 6 'Venice' CPUs against Nvidia's Vera CPU, using Nvidia's published configuration. AMD claims its Venice CPUs show a 20% per-core performance advantage and 2.2x higher throughput compared to Vera, indicating a competitive stance in the high-performance computing market.
NVIDIA has launched the Vera Rubin platform, designed for high-performance AI factories, with a significant 10x throughput improvement per megawatt compared to previous models. This system features integrated extreme co-design with optimized components for better power efficiency and networking capabilities, marking a substantial advancement in AI infrastructure architecture.
Nvidia has launched its Vera CPU targeting the AI server market, delivering chips to clients such as OpenAI and SpaceX. This move positions Nvidia as a contender against traditional CPU giants AMD and Intel, emphasizing a shift in server architecture driven by the needs of AI applications.
Nvidia unveiled benchmarks for its Vera CPU targeting the growing AI data center market, showing comparisons with AMD's Epyc 9755. The release signifies Nvidia's strategy to capture market share from hyperscalers expanding AI infrastructure.
Nvidia's Vera CPU has shown strong single-threaded performance, challenging x86 competitors. It targets AI inference workloads in which single-thread speed is crucial. The upcoming Rigel CPU core promises improved performance per core and efficiency.
NVIDIA has introduced the Vera CPU, designed for optimal single-threaded performance to enhance AI operations. This new CPU category addresses the inadequacies of current data center CPUs that prioritize core count over speed, which hinders efficiency in AI factory environments.