Intel Orders ASML Machine Still on Drawing Board as Chipmakers Look For an Edge

The Intel Corporation logo is seen on a display in a store in Manhattan, New York City, November 24, 2021. (Reuters)
The Intel Corporation logo is seen on a display in a store in Manhattan, New York City, November 24, 2021. (Reuters)
TT

Intel Orders ASML Machine Still on Drawing Board as Chipmakers Look For an Edge

The Intel Corporation logo is seen on a display in a store in Manhattan, New York City, November 24, 2021. (Reuters)
The Intel Corporation logo is seen on a display in a store in Manhattan, New York City, November 24, 2021. (Reuters)

Intel has placed the first order with ASML for an advanced chipmaking machine that is still on the drawing board and won't be delivered for years, as semiconductor manufacturers look to get ahead in a booming industry.

Alongside better-than-expected fourth quarter earnings, ASML said on Wednesday it had now received orders for five of its next generation of lithography machines, plus an order for the even newer model that is still being designed.

In a separate joint press release, the companies said Intel was the buyer.

ASML's most advanced commercial machines, known as EUV lithography systems because of the "Extreme Ultraviolet" light waves they use to map out the circuitry of computer chips, are the size of a bus and cost around $150 million each.

A cutting-edge chip plant needs 9-18 of them, making them one of chipmakers' biggest capital expenditure costs.

ASML is the only maker of such machines and its EUV customers include the world's biggest chipmakers, TSMC of Taiwan, Samsung of South Korea and Intel, among others.

ASML said it had now received orders for five of the next iteration of the system, known as "High NA" EUV machines, which will have a different lens system with a higher numerical aperture, will be even larger and cost around $300 million each.

The first prototypes are to be shipped in 2023. They are not expected to be used in volume manufacturing until 2025.

Intel, which is trying to win back its position as maker of the smallest and fastest chips from current leader TSMC, had previously identified itself as the first buyer of a High NA machine, ASML's EXE:5000, which is expected to help manufacturers shrink chip sizes later this decade.

Intel said on Wednesday it would buy the first EXE:5200 as well, which is expected to make several improvements including higher productivity. Neither company mentioned the price tag.



SDAIA Lists ALLaM 7B Arabic Language Model on Hugging Face

SDAIA Lists ALLaM 7B Arabic Language Model on Hugging Face
TT

SDAIA Lists ALLaM 7B Arabic Language Model on Hugging Face

SDAIA Lists ALLaM 7B Arabic Language Model on Hugging Face

The Saudi Data and Artificial Intelligence Authority (SDAIA) has disclosed the listing of the generative Arabic Large Language Model (ALLaM) —7 billion parameters category— on the Hugging Face platform. This platform, a prominent global open-source repository for natural language processing (NLP) resources, now hosts the nationally developed model, making it accessible to a global audience of developers and researchers for utilization in artificial intelligence applications.
SDAIA asserts that ALLaM is positioned as one of the preeminent generative language models for the Arabic language worldwide, SPA reported.
This initiative aligns with the strategic objectives of the Kingdom's Vision 2030, under the leadership of Prince Mohammed bin Salman bin Abdulaziz Al Saud, Crown Prince, Prime Minister and SDAIA Board of Directors Chairman, who supports SDAIA's mandate to strengthen the Kingdom's position as a global technology hub for artificial intelligence and digital innovation.
ALLaM is a large language model distinguished by its precision and efficacy in the comprehension and processing of Arabic. Its capabilities encompass text summarization, content generation, and other functionalities, positioning it as a tool for the development of smart applications. The model's training, conducted on an extensive and diverse Arabic language dataset, is cited as contributing to its performance across varied operational environments.
According to SDAIA, the availability of ALLaM on the Hugging Face platform is intended to facilitate access for the scientific and technical community, enabling the development of artificial intelligence-driven applications.
The listing on Hugging Face represents a phase in the model's ongoing development. ALLaM was initially integrated into IBM's Watsonx platform, which facilitated advanced Arabic language analysis and processing through artificial intelligence solutions. Subsequently, it was launched on Microsoft's Azure cloud platform, providing developers and enterprises access to advanced cloud computing environments.
During the LEAP25 conference, Qualcomm, in collaboration with the ALLaM development team, announced the model's availability on the Qualcomm AI Cloud. Additionally, the "ALLAM AI PC," a computer incorporating advanced artificial intelligence technologies designed to enhance user interaction with the Arabic language, was unveiled.