Urgent.News

What's breaking now, across thousands of outlets.

AI

Anthropic Challenges Nvidia’s AI Chip Dominance

The artificial intelligence (AI) semiconductor market, where Nvidia has been running unchallenged, is showing signs of cracking. Anthropic, the developer of ‘Claude,’ announced that it can modify some of its models tailored to Nvidia chips to run on AMD semiconductors in just three days.Big tech com

The AI semiconductor market, once dominated by Nvidia, is experiencing a shift as Anthropic, the creator of 'Claude', has demonstrated the ability to adapt its models to run on AMD semiconductors in just three days. This is a significant reduction from the months it took previously, according to Co-founder Tom Brown. This development is part of Anthropic's broader strategy to avoid being reliant on a single semiconductor supplier.

Big tech companies, including Google and Amazon, are also expanding their AI semiconductor capabilities. Meanwhile, Nvidia's software 'moat' is under threat, with former CEO Jensen Huang, who was instrumental in the company's technological advancements, no longer leading the firm. At an investment briefing prior to its IPO, Anthropic announced the successful adaptation of its AI models to AMD semiconductors.

This trend aligns with analysts' expectations that Nvidia's dominance, which was largely based on developers' dependence on its GPUs, may be dismantled if AI facilitates easier conversions between different chipsets. Anthropic is now using AMD, Google TPU, and Amazon Trainium alongside Nvidia, while OpenAI is also working on securing AMD-based infrastructure.

Furthermore, a JPMorgan report indicated that Nvidia's share in future computing contracts for major AI companies is projected to decline, with AMD GPUs accounting for 18% of these contracts. Nvidia's market share has decreased significantly since 2023, when it held 80% to 90% of AI semiconductor revenues.

Big tech firms are also accelerating the development of their own AI semiconductors. Google is working on the TPU, Amazon on Trainium and Inferentia, Microsoft on Maia, and Meta on the MTIA. Amazon has even released software to make it easier to use Trainium.

The departure of former CEO Huang is seen as a significant blow to Nvidia, as he not only spearheaded technological development but also played a crucial role in establishing the broader AI industry ecosystem. Despite these challenges, there remains a belief that Nvidia's dominance is not likely to collapse in the near future, thanks to the robust CUDA ecosystem and its integration of GPUs, networks, and software.

Written by urgent.news from BusinessKorea's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at businesskorea.co.kr →

More in AI

More from Tuesday 17 November →