Shulou(Shulou.com)12/24 Report--

CTOnews.com, December 15, at the "make AI everywhere" event yesterday, Intel officially released the new Core Ultra mobile processor and the fifth-generation Xeon scalable processor.

In addition, Intel CEO Pat Kissinger (Pat Gelsinger) also publicly attacked Nvidia, saying that its CUDA technology moat is shallow and small, and that reasoning technology will be more important than training for artificial intelligence.

▲ Touyuan: in an interview, Intel hinted that Nvidia CUDA's dominance would not last forever. "you know, the whole industry is thinking about how to eliminate the CUDA market." He cited examples such as MLIR, Google and OpenAI that they are turning to "Pythonic" to make AI training more open.

"We think the CUDA moat is shallow and narrow because the industry is committed to bringing a wider range of technologies to a wide range of training, innovation, data science and other areas," he said.

Kissinger believes that AI can not only rely on training, reasoning is the really feasible way. "when you embrace reasoning, once you've trained the model, you don't have to rely on CUDA," he said. "the question is, can you run the model well?"

This is not to say that Intel will not compete in training, but that "fundamentally, the reasoning market is the focus of competition," he said.

Kissinger also took the opportunity to promote OpenVINO, a standard Intel developed specifically for its AI work, and predicted a world of "hybrid computing" in which some computing exists in the cloud while others exist on local PC.

He also announced that "We will compete for 100% data center AI TAM in three ways," that is, "our leader CEO, leader accelerator and foundry". "We will pursue every internal opportunity: TPU, reasoning engine, training chip, etc. We will also pursue every business opportunity, working with NVIDIA, AMD and so on. We will become a contract factory."

At yesterday's press conference, Intel brought a new generation of core Ultra mobile processors. The Core Ultra is the first processor based on the Intel 4 process and represents Intel's most significant architectural change in 40 years, officials said.

Intel said the Intel Core Ultra processor used the Neural Network processing Unit (NPU), Intel's first on-chip AI accelerator for the client, to take energy-efficient AI acceleration to new heights, delivering 2.5 times the energy efficiency performance of the previous generation.

Intel says the core Ultra processor will bring AI features to more than 230 models from global notebook and PC manufacturers next year. By 2028, AI PC will account for 80% of the PC market and bring new tools to our work, study, and creation.

Intel's first Core Ultra series processors include U and H series, with basic power consumption of 15W and 28W, respectively. U series is 2'8'2 core specification, with the highest 4Xe display. H series has the highest core specification of 6'8'2 and the highest 8Xe display. In the first quarter of next year, Intel will also release a flagship Ultra 9 model with a basic power consumption of 45W, a core specification of 6x8002, and a 8Xe display. The detailed parameters of CTOnews.com are as follows:

Related readings:

"Intel's new generation AI accelerator Gaudi 3 makes its debut and is expected to go public as scheduled next year."

"Intel Core Ultra Mobile processor release: double Core display performance using Intel 4 Technology"

Intel Fifth Generation Xeon Extensible processor release: average performance improvement of 21%, built-in AI Accelerator

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.