AMD has just unveiled a super AI chip to compete with NVIDIA.

On June 13th, AMD announced the detailed specifications of its AI superchip with up to 192 gigabytes of internal memory, which will be launched in the third quarter to compete for Nvidia's market leadership position.

According to AMD CEO Lisa Su, the new product could help technology companies balance costs when offering services similar to ChatGPT.

AMD has just unveiled a super AI chip to compete with NVIDIA.

"The larger the memory, the larger the processing capacity of the chip," Lisa Su said. "We've seen workloads processed faster, and that's what makes the difference."

However, AMD has yet to mention which major customers are ready to spend money on the MI300X – the company's latest AI graphics superchip. Furthermore, they haven't disclosed details about pricing or how they plan to boost sales.

Since the beginning of the year, AMD shares have doubled, reaching a 16-month high on June 13, but closed down 3.6% after the company's AI strategy presentation.

Nvidia is currently the first chip manufacturer to have a market capitalization exceeding one trillion USD, dominating its competitors on a large scale. Although Intel and several other startups such as Cerebras Systems and SambaNova Systems have launched some competing products, the biggest threat to Nvidia's sales comes from Google Alphabet and Amazon's efforts to manufacture chips internally.

Besides the AI ​​technology market, AMD also announced that it has begun shipping large batches of its “Bergamo” central processing units to companies such as Meta Platforms.

On the same day, June 13th, chip manufacturer AMD also announced a software update for RoCM to compete with Nvidia's CUDA software platform.

Source: https://baoquocte.vn/amd-vua-trinh-lang-sieu-chip-ai-canh-tranh-voi-nvidia-230898.html