Yazılar

Nvidia Shifts Focus to New Advanced Packaging Technology

Nvidia’s CEO Jensen Huang confirmed that while the company’s demand for advanced packaging from TSMC remains robust, the specific type of technology required is evolving. At an event in Taichung, Taiwan, Huang explained that Nvidia is transitioning its focus from CoWoS-S to CoWoS-L for its upcoming Blackwell AI chips. This shift, however, does not signal a reduction in capacity, but rather an increase in the use of CoWoS-L, a newer, more advanced version of TSMC’s chip packaging technology.

Nvidia had previously relied heavily on CoWoS-S for its AI chips, including the Hopper platform. As the company moves into Blackwell, which was unveiled in March 2024, it plans to transition existing CoWoS-S capacity to CoWoS-L. This change will impact TSMC’s supply chain but is seen as a step forward in Nvidia’s push to meet the growing demand for its AI chips.

Huang also noted that while packaging capacity for these advanced chips had previously been a bottleneck, it had expanded significantly in recent years, with available capacity now approximately four times greater than it was two years ago. Despite the increased demand, Nvidia has not been cutting orders but is instead increasing its reliance on CoWoS-L, which is expected to better meet the needs of Blackwell’s design.

The move to CoWoS-L technology and changes in Nvidia’s order patterns have sparked speculation about the potential impact on TSMC’s revenue, particularly with analysts like Ming-Chi Kuo noting the shift in Nvidia’s focus. Huang declined to comment on recent U.S. export restrictions that limit AI chip sales to countries outside a select group of U.S. allies, but the company’s strategies continue to evolve in response to market demands and geopolitical factors.

 

Apple Partners with Broadcom to Create Its First AI Server Chip, Report Says

Apple Collaborates with Broadcom on AI Server Chip Development
Apple is reportedly working with semiconductor manufacturer Broadcom to develop a specialized server chip aimed at powering artificial intelligence (AI) features. The move aligns with Apple’s growing emphasis on AI-driven capabilities across its product ecosystem. While the company has previously announced plans to offload some processing for Apple Intelligence features to the cloud, this marks the first instance of Apple creating a dedicated server chip for AI applications. Recent updates to iOS, iPadOS, and macOS have already introduced advanced Apple Intelligence features, including integration of tools like ChatGPT with Siri.

Codenamed ‘Baltra’: Apple’s AI Server Chip in the Works
According to a report from The Information, which cites sources familiar with Apple’s plans, the new server chip, internally referred to as “Baltra,” is designed specifically for AI processing tasks. Unlike Apple’s existing processors, which primarily power on-device AI functionalities in iPhones, iPads, and Macs, the Baltra chip will handle AI tasks in the cloud. This could pave the way for more powerful and complex AI-driven features that require robust server-side computation.

Collaborative Effort Focused on Networking Technology
The partnership with Broadcom extends to the development of networking technology essential for the Baltra chip’s functionality. By optimizing the chip for cloud-based AI workloads, Apple aims to deliver faster and more efficient responses to user requests. This could include tasks like natural language processing, enhanced Siri capabilities, and improved integration with AI models, ensuring seamless performance across Apple devices.

A Strategic Step Toward AI Leadership
Developing an AI server chip represents a significant step in Apple’s strategy to remain competitive in the AI space. As rivals like Google and Microsoft advance their AI infrastructures, Apple’s investment in custom server chips highlights its commitment to innovation. The introduction of Baltra not only strengthens Apple’s control over its technology stack but also opens the door for new AI-powered experiences tailored to its ecosystem, reinforcing its position as a leader in consumer technology.

US Asks Nvidia to Investigate Unauthorized Exports of AI Chips to China

The U.S. Department of Commerce has reportedly asked Nvidia to investigate how the company’s artificial intelligence (AI) chips have ended up in China over the past year, according to a Thursday report by The Information. The investigation follows concerns over the potential unauthorized diversion of Nvidia products to Chinese entities, despite U.S. export restrictions.

Investigation into Distribution Channels

Nvidia has enlisted major distributors, including Super Micro Computer and Dell Technologies, to conduct spot checks of their customers in Southeast Asia. Nvidia’s AI chips are embedded in servers made by Super Micro and Dell, and the company is reportedly looking into how these chips may have been redirected to China without the required licenses.

The report suggests that multiple individuals involved in smuggling Nvidia chips have managed to evade detection during recent inspections conducted by Super Micro. The smuggling tactics reportedly involved duplicating serial numbers from Nvidia servers or altering them within the server’s operating system to disguise their origin.

Nvidia’s Response and Policies

In response, an Nvidia spokesperson reiterated the company’s commitment to adhering to U.S. export control regulations, stressing that any unauthorized diversion of its products would be detrimental to its business. Nvidia also insisted that its customers and partners must strictly follow these regulations, including prohibitions on grey market resales.

Both Dell and Super Micro have also emphasized their compliance with U.S. export laws. Dell stated that it requires all distributors and resellers to follow export controls and would sever relationships with any partner found to be non-compliant. Super Micro similarly asserted that it investigates and acts against any unauthorized exports, affirming its commitment to U.S. regulations.

Broader Context: U.S. Crackdown on Chinese Exports

This investigation comes as the Biden administration intensifies its crackdown on Chinese access to high-end AI chips. The U.S. government has already broadened its ban on the sale of advanced AI chips to China, including a move to limit semiconductor exports to 140 companies, including chip equipment makers, earlier this month. Despite these restrictions, there have been reports that Chinese universities and research institutions have still been able to procure Nvidia chips through resellers.