Yazılar

Nvidia Unveils “Rubin CPX” AI Chips for Video and Software Generation

/in /tarafından

Nvidia (NVDA.O) announced plans to launch a new AI chip, dubbed Rubin CPX, by the end of next year, targeting highly complex workloads such as video generation and AI-assisted software coding. The chip will be built on Nvidia’s upcoming Rubin architecture, the successor to its current Blackwell technology.

Why It Matters

  • AI systems are rapidly evolving, with tasks like video generation and “vibe coding” (AI-assisted software creation) pushing hardware to new limits.

  • Processing one hour of video can require up to 1 million tokens, a massive challenge for current GPUs.

  • Rubin CPX will integrate video decoding, encoding, and inference into a single system, making processing faster and more efficient.

Economic Angle

  • Nvidia estimates a $100 million investment in Rubin CPX systems could generate $5 billion in token revenue.

  • Wall Street is closely watching the ability of AI hardware firms to turn capital spending into measurable returns.

Market Impact

  • Nvidia already dominates the AI chip market, with its high-end processors fueling the latest wave of generative AI.

  • The company’s move reflects both its defensive strategy against rivals and its offensive push to expand AI capabilities beyond text and images into full-scale video and software generation.

The Bigger Picture

  • Nvidia’s rise has made it the world’s most valuable company, but competition in AI infrastructure is intensifying.

  • With Rubin CPX, Nvidia is betting that integrated, video-ready AI chips will anchor the next phase of AI growth — and cement its lead in the sector.

US Servers in Singapore Fraud Case May Contain Nvidia Chips, Minister Says

/in /tarafından

Servers involved in a fraud case recently announced by Singapore were supplied by U.S. firms and may contain advanced Nvidia chips, Singapore’s Home Affairs and Law Minister K Shanmugam said on Monday.

Three men, including a Chinese national, were charged last week in Singapore with fraud linked to these servers. The servers, supplied by Dell Technologies and Super Micro Computer to companies based in Singapore, were then sent to Malaysia, though it is unclear if Malaysia was the final destination.

Shanmugam noted the authorities are investigating the case independently following an anonymous tip-off. Singapore has asked U.S. officials to confirm if the servers contained U.S. export-controlled items and expressed willingness to cooperate in any joint investigation.

This case is part of a broader investigation involving 22 individuals and companies suspected of false representation, amid concerns over organized smuggling of AI chips, particularly to China. The U.S. is probing whether Chinese company DeepSeek used U.S. chips prohibited from export to China. Earlier reports revealed that Chinese research institutions obtained Nvidia’s advanced AI chips embedded in servers made by Dell, Super Micro, and Gigabyte Technology.

Singapore is Nvidia’s second-largest market after the U.S., accounting for 18% of Nvidia’s latest fiscal year revenue, though actual shipments to Singapore represent less than 2% since the country functions largely as an invoicing hub.

Some AI industry figures, like Scale AI CEO Alexandr Wang, allege DeepSeek has as many as 50,000 higher-end Nvidia chips banned for export to China, though these claims remain unverified. DeepSeek stated it legally purchased Nvidia H800 chips in 2023 and disclosed use of Nvidia A100 chips in its supercomputing AI cluster.

Dell stated it enforces strict trade compliance and takes action if customers violate obligations but declined further comment due to the ongoing investigation. Super Micro affirmed compliance with U.S. export controls and said it investigates any unauthorized re-exports. Nvidia declined to comment, and DeepSeek has not responded to requests for comment.

AMD Collaborates with AI Startups to Enhance Chip and Software Development

/in /tarafından

Advanced Micro Devices (AMD) is strengthening its ties with a range of artificial intelligence startups to improve both its chip hardware and software ecosystems, aiming to better compete against dominant players like Nvidia.

In its pursuit to build a competitive line of AI chips, AMD has acquired companies such as server maker ZT Systems and several small software firms to boost its talent pool and software capabilities. The company is focused on enhancing ROCm, its AI software platform, to better support the complex needs of AI developers.

Vamsi Boppana, AMD’s senior vice president of AI, described the effort as a “thoughtful, deliberate, multi-generational journey” aimed at creating hardware and software that meet evolving AI demands.

One of AMD’s key customers benefiting from these software improvements is AI enterprise startup Cohere, which focuses on building AI models tailored for large businesses. Cohere CEO Aidan Gomez said AMD’s enhancements have drastically shortened the time it takes to adapt their software to AMD chips—from weeks down to days—improving agility in deployment.

On the hardware front, OpenAI has played a notable role in shaping AMD’s upcoming MI450 series AI chips. Forrest Norrod, AMD’s executive vice president, revealed that feedback from OpenAI influenced the design of the MI450’s memory architecture and chip scalability to accommodate thousands of chips working together for AI applications. OpenAI also guided AMD on optimizing the chips for specific mathematical operations vital for AI workloads.

The MI400 series will underpin AMD’s new “Helios” server expected next year, designed to rival Nvidia’s integrated AI servers that link hundreds or thousands of GPUs.

OpenAI CEO Sam Altman appeared alongside AMD at their recent San Jose event, highlighting the ongoing collaboration between the two organizations.