Search

Intel CEO Pat Gelsinger unveiled the company's AI roadmap at the Vision 2024 conference, introducing the Gaudi 3 processor designed for enterprise genAI workloads. Intel also announced strategic collaborations and next-gen products, including AI-enabled PCs. The Gaudi 3 processor is positioned as a competitor to Nvidia's H100 GPU, offering better inference and power efficiency at a lower cost. Intel's roadmap includes the release of Xeon 6 processors and advancements in AI fabrics and networking. However, Intel faces challenges in competing with Nvidia and TSMC in the AI hardware market and chip fabrication. The company is set to receive funding from the CHIPS Act to support its future plans. [Extracted from the article]

Chinese researchers have developed a DVD-sized disc with a storage capacity of 200 terabytes by storing data in three dimensions. The data is recorded in dots that are 54 nanometers wide, allowing for a higher data density than Blu-ray discs or modern hard drives. The researchers believe that this new optical disc could be used in large data centers, potentially reducing their size from a stadium to a room. The disc is made of a new light-sensitive material called AIE-DDPR, and the researchers are working on improving the writing speed and energy consumption of the method. [Extracted from the article]

Intel is shifting its focus towards artificial intelligence (AI) as it faces financial difficulties, including an 85% drop in second-quarter profit. The company plans to cut over 15,000 jobs and invest in the development of next-generation AI chips and chip fabrication capabilities. Intel's revenue from AI and data center business has declined, while competitors like AMD have seen growth in this area. Intel aims to turn its financial situation around by investing in AI-driven product lines and reducing operating expenses. However, analysts are skeptical about Intel's ability to catch up in the AI market. [Extracted from the article]

An analyst from Info-Tech Research Group predicts that the US Department of Justice's antitrust probe into Nvidia's selling practices is unlikely to result in any significant action. The probe was initiated following allegations that Nvidia pressured cloud providers to purchase multiple products. While progressive groups and Democratic Senator Elizabeth Warren have called for an investigation into Nvidia's dominance in the market for AI chips, the analyst believes that the investigation is more about the battle between mega corporations and that there is little competition for Nvidia's product. The analyst also suggests that any action taken by the DOJ would likely come too late to have a significant impact on the market. [Extracted from the article]

The role of Chief Artificial Intelligence Officer (CAIO) is gaining prominence in organizations that are deploying generative AI technology. According to research firm IDC, 11% of midsize and large organizations have already filled a CAIO role, and another 21% are actively seeking one. CAIOs are expected to be strategic members of the organization and are responsible for extracting ROI from AI technology. The Biden administration has also mandated that all federal agencies appoint CAIOs to promote AI innovation and manage risks associated with the technology. However, the responsibilities and authority of CAIOs vary between organizations. Ensuring data quality and ethical considerations are important aspects of the role, as well as collaboration and continuous learning across all functions within an organization. While most organizations have not designated CAIOs as an official C-suite role, the position is often filled by CTOs or CISOs. However, PricewaterhouseCoopers (PwC) believes that the CAIO role should be dedicated and accountable for an organization's AI initiatives. The successful deployment of AI is crucial as it is expected to become a part of everyday life by 2030. [Extracted from the article]

Nvidia, a leading chipmaker in the AI industry, recently surpassed a $3 trillion market capitalization, making it the third company ever to achieve this milestone. While some may view this success as a temporary bubble, the author argues that Nvidia is actually undervalued. Unlike Apple, which faces challenges in sustaining its growth, Nvidia has significant potential for future expansion in areas such as AI processing, self-driving cars, robotics, and industrial automation. The company's CEO, Jensen Huang, introduced the concept of "Physical AI," which involves using virtual environments to simulate real-world physics and train robots. Nvidia's dominance in AI hardware and software, along with its lack of competition in the "Physical AI" space, positions it for significant future growth. [Extracted from the article]

Chinese companies are reportedly exploiting a loophole in US export control rules to gain access to high-performance computer chips. The loophole involves the sale of AI-related cloud computing services, which are not currently covered by the export controls. Chinese firm ByteDance, the company behind TikTok, is said to be renting US-based servers containing Nvidia's H100 chips from Oracle's cloud service to train AI models. The Biden administration is considering closing this loophole by restricting Chinese companies' access to US cloud computing services. However, some experts caution against a blanket ban, suggesting more tailored approaches to regulation. [Extracted from the article]

Intel has released its latest line of x86 processors, which it hopes will help it compete with Nvidia in the AI market. The new Xeon 6 processors are aimed at AI tasks and offer improved performance and energy efficiency compared to their predecessors. Intel is also targeting the AI PC industry with its Lunar Lake line of client processors. While Nvidia currently dominates the AI processor marketplace, Intel is focusing on affordability and edge devices to carve out its own share of the market. [Extracted from the article]

Apple CEO Tim Cook has promised that this year's WWDC event will be the company's best ever. The event, which will mostly take place online with some on-site guests, will feature a keynote speech, developer sessions, access to Apple engineers, and the annual Apple Design Awards. The focus of this year's event appears to be on spatial computing and augmented reality, with speculation that Apple may announce international availability dates for its Vision Pro headset. However, some industry watchers are skeptical that Apple's focus on mixed reality will be enough to satisfy the tech world's current obsession with generative AI. [Extracted from the article]

Apple's upcoming AI announcements at WWDC may not meet the high expectations that have been building. While there will be interesting features such as voice memo transcription and generative AI-powered editing tools, industry insiders believe that Apple still lags behind its competitors in genAI development. Apple has been in talks with OpenAI, Google, and Baidu, and there may be a potential deal between Apple and OpenAI, possibly involving integration of ChatGPT on the iPhone. Privacy and security are important considerations for Apple, and they are investing in secure data center processors for their cloud-based genAI services. The future evolution of server-based solutions like ChatGPT and Google Gemini may be constrained by AI regulation and data sovereignty, which could give Apple an advantage in the long run. [Extracted from the article]

The adoption of generative artificial intelligence (genAI) is increasing, leading to a supply and demand bottleneck for the high-performance chips needed to support this growth. Sixty-six percent of enterprises worldwide plan to invest in genAI in the next 18 months, with infrastructure accounting for 46% of the total spend. The shortage of high-bandwidth memory (HBM) chips, which are necessary for AI processing, is a key issue. SK Hynix, a major supplier of HBM chips, has stated that its products are almost fully booked through 2025 due to high demand. However, other suppliers like Samsung and Micron are increasing their HBM production to meet the demand. The shortage is expected to ease by the end of the year. Nvidia, Intel, and AMD are all ramping up production to address the GPU supply shortage. Despite the challenges, chipmakers are working to meet the increasing demand for AI-focused chips, with global spending expected to reach $53 billion this year. Intel and AMD have announced new processors specifically designed for AI workloads. On the software side, creators are developing smaller models tailored for specific tasks that require fewer processing resources. Overall, the industry is adapting to the high demand for AI chips, and supply is expected to improve in the near future. [Extracted from the article]

Intel's CEO, Pat Gelsinger, outlined a strategy at the Intel Vision conference that focuses on generative artificial intelligence (genAI) and industry standard Ethernet. Intel aims to regain dominance in the computing industry with its Gaudi 3 GPU and Xeon 6 processors, which are designed for training and running large language models (LLMs) that power genAI. Intel claims that its chips are more cost-effective and offer a friendlier ecosystem compared to Nvidia's offerings. However, industry analysts believe that Intel faces a challenging task in catching up to Nvidia, which currently dominates the data center AI GPU market. [Extracted from the article]

Data centers are playing a crucial role in our increasingly connected world, but their rapid growth is also contributing to environmental pressures. As data usage is expected to quadruple by 2025, it is essential to find ways to increase energy efficiency and reduce power consumption while meeting the growing demand for data. NTT, one of the world's largest data center companies, is taking steps to support sustainability. They have implemented innovative solutions in their data centers in Japan, California, and Germany, such as efficient cooling systems, reusing heat, and conserving water. NTT is also exploring other technological advancements, like transitioning to photonics and even putting data centers in space, to minimize climate impact. [Extracted from the article]

Nvidia's strong earnings report, driven by increased chip sales for AI and cloud use, has had a significant impact on the stock market. The Dow Jones Industrial Average and the S&P 500 both reached new records as a result. Nvidia's revenue grew by 265% in the fourth quarter, and its net income exceeded forecasts by nearly $2 billion. The company's success is attributed to its full-stack AI offerings and its position as the only chip maker with a solution for generative AI platforms. The demand for AI and cloud services is expected to continue growing, with on-device AI chipsets projected to reach over 1.8 billion units by 2030. However, there is currently a shortage of AI-specific GPUs, and efforts are being made to increase chip manufacturing capacity in the US. The CHIPS Act, which promises funding to spur chip manufacturing in the US, is expected to have an impact in the next few years. [Extracted from the article]

Data centers in space are no longer just science fiction. Trials and demonstrations are currently underway for a satellite networking system that would collect data from Earth and process and store it in space. This system, called the Space Integrated Computing Network, would use photonics and optical technology to reduce power consumption and increase data transmission speeds. By placing data centers in space, communication interruptions caused by Earth's weather or natural disasters would be eliminated. The launch of this space-based network is planned for 2025 and is expected to provide faster data communications, reduce electricity consumption, and create a sustainable infrastructure for future data needs. [Extracted from the article]

Regulators, investors, customers, and employees are pressuring companies to minimize the environmental impact of their AI initiatives. The energy consumption of AI models, particularly GPUs, has raised concerns about their contribution to global warming. To address these concerns, regulators are considering requiring large AI users to report on their environmental impact. Companies are finding success in streamlining their AI efforts by using more efficient processes and architectures, reducing the size of datasets, using renewable energy for data centers, and considering alternative approaches to AI when simpler solutions may suffice. However, challenges remain in implementing sustainable AI, such as identifying power-consuming workloads and obtaining information on the carbon footprint of AI models. [Extracted from the article]