9.5 C
Ташкент
Пятница, 13 февраля, 2026
ДомойArtificial IntelligenceHow Facebook’s new AI system may be ‘good news’ for Nvidia

How Facebook’s new AI system may be ‘good news’ for Nvidia

Date:

Related stories

AQSh va Xitoyning kiberxavfsizlik strategiyasi: kim nimani ko‘zlamoqda?

Xalqaro tizimda ko‘plab davlatlar o‘zining kibermudofaa va hujum qobiliyatini...

Huawei yangi smartfonlarini namoyish etdi

P50 smartfoni Kirin 9000, P50 Pro smartfoni esa Snapdragon...

Mercedes-Benz kompaniyasi S-Class sedanining zirhlangan versiyasini ishlab chiqdi

Mercedes-Benz kompaniyasi yangi avloddagi S-Class sedanining zirhlangan versiyasini ishlab...

Telegram va TikTok ishida uzilishlar kuzatildi

Downdetector ma'lumotlariga ko‘ra, Telegram foydalanuvchilarida ilovani ishga tushirishda, serverga...

O‘zbekcha “Alisa” qanday ishlab chiqildi? — loyiha asoschilari bilan suhbat

O‘zbekchani tushunib, unda gapira oladigan sun’iy intellekt yaratildi. Endilikda...
spot_imgspot_img

Facebook parent company Meta is reportedly working on a new artificial intelligence (AI) system that could be as powerful as the most advanced large language model (LLM) offered by OpenAI, a report has claimed.

The Microsoft-backed company currently offers GPT-4 LLM that powers the ChatGPT and is available as a part of ChatGPT Plus subscription. The same technology also powers Microsoft’s Copilot AI that the company has incorporated in its products and services, including Windows 11.

Citing people familiar with the matter, a report by the Wall Street Journal claimed that Meta is aiming for its new AI model to be ready next year and plans to begin its training early in 2024. The report also claims that the new model will be several times more powerful than its commercial version dubbed Llama 2.

Meta launched Llama 2 earlier this year and the company said that this model was trained on 40% more data than the Llama model and has double the context length. This translates to, in layman’s terms, more contextual answers and better results.

‘Good news’ for Nvidia

In order to run this model, Meta is reportedly buying more Nvidia H100 AI-training chips and is making the infrastructure more powerful so that it won’t need to rely on Microsoft’s Azure cloud platform to train the new chatbot.

The chip is in high demand and is used by companies, including Microsoft, Google and OpenAI. Google recently introduced its A3 supercomputers with Nvidia H100 GPUs. Meta buying more chips could help the company’s profits to swell.

According to a report by the Financial Times, TSMC is on track to deliver 550,000 H100 GPUs to Nvidia this year, with potential H100 revenues of $13.75 billion to $22 billion this year.

“With estimated margins of 60-90% per chip, Nvidia could make over $10 billion in profits from H100 sales this year alone,” the FT report claimed.

Other players in the AI business also reportedly include Apple. It has been reported that the company is building its own framework, known as ‘Ajax’, to create LLM and is also testing a chatbot which is said to be called ‘Apple GPT’.

Subscribe

- Never miss a story with notifications

- Gain full access to our premium content

- Browse free from up to 5 devices at once

Latest stories

spot_img

ОСТАВЬТЕ ОТВЕТ

Пожалуйста, введите ваш комментарий!
пожалуйста, введите ваше имя здесь