🚀 Medial Secures Investment on Shark Tank India - Fueling the Future of Professional Social Networking. 🔥

Startup Showcase

Premium Content

Try our Valuation Calculator →

News on Medial

Meta trials in-house AI chip to reduce reliance on external suppliers: Report

YourStory

YourStory · 5m ago

Meta trials in-house AI chip to reduce reliance on external suppliers: Report

Meta is testing its first in-house AI chip to reduce reliance on suppliers like NVIDIA. Collaborating with a Taiwanese manufacturer, Meta aims to cut AI infrastructure costs. The chip is specialized for AI tasks, initiated after completing a crucial design milestone. Previously, Meta abandoned a custom chip, favoring NVIDIA GPUs. Despite this, Meta remains a significant NVIDIA customer. Meanwhile, Microsoft seeks alternatives for AI models, evaluating options to potentially rival OpenAI.

Related News

Meta tests in-house AI chip with TSMC to cut Nvidia reliance

Inshorts · 5m ago

Meta tests in-house AI chip with TSMC to cut Nvidia reliance

Meta is testing its first in-house AI training chip with TSMC to reduce reliance on Nvidia. The chip is a dedicated AI accelerator designed for better efficiency. If successful, Meta plans to scale it for recommendation systems and later for generative AI tools like Meta AI. This comes after a previous failed attempt at custom chips.

Google expands in-house chip efforts in costly AI battle

Livemint

Livemint · 1y ago

Google expands in-house chip efforts in costly AI battle

Google is expanding its in-house chip development with the introduction of a new chip called Axion. The move is part of Google's strategy to reduce reliance on external vendors and compete with partners such as Intel and Nvidia. The Axion chips, based on circuitry from chip-design firm Arm, can be used for a range of tasks, including powering Google's search engine and AI-related work. Google has chosen to rent its custom chips to cloud customers instead of selling them directly. The company's investment in Axion could pay off with significant performance improvements, but it faces tough competition from other cloud providers.

Meta debuts new generation of AI chip

Economic Times

Economic Times · 1y ago

Meta debuts new generation of AI chip

Meta Platforms has announced the details of its next-generation AI accelerator chip called Artemis. The chip aims to reduce Meta's reliance on Nvidia chips and lower energy costs. The Meta Training and Inference Accelerator (MTIA) chip focuses on providing the right balance of computing power, memory bandwidth, and capacity for serving ranking and recommendation models. Other than chip development, Meta has also invested in software to optimize its infrastructure efficiency. The company plans to produce the chip on Taiwan Semiconductor Manufacturing Co's 5nm process, and it has already been deployed in data centers for AI applications.

Tesla’s Dojo supercomputer is DOA — now what?

The Verge

The Verge · 15d ago

Tesla’s Dojo supercomputer is DOA — now what?

Tesla's Dojo supercomputer project has been disbanded after key talent, including project leader Peter Bannon, left to join another AI venture. The departure follows similar exits from other top executives, highlighting Elon Musk's struggle to retain talent. The closure marks a setback for Tesla's in-house AI development strategy for self-driving cars and indicates a shift to reliance on external suppliers like Nvidia and Samsung. This pivot comes amid broader challenges confronting Tesla's business.

OpenAI is reportedly getting closer to launching its in-house chip

The Verge

The Verge · 6m ago

OpenAI is reportedly getting closer to launching its in-house chip

OpenAI is advancing toward launching its own in-house AI chip design, targeting completion within this year before production by Taiwan Semiconductor Manufacturing Co (TSMC) using 3-nanometer technology. This development aims to reduce reliance on Nvidia chips for AI model training and operation, deploying initially on a "limited scale." As led by former Google TPU engineer Richard Ho, OpenAI's chip design team has grown, with future plans for more advanced iterations.

First crack in Nvidia AI wall? Meta set to use own AI chips in its servers in 2024 — but it will still use Nvidia H100 GPUs as well in its datacenters for now

Techradar

Techradar · 1y ago

First crack in Nvidia AI wall? Meta set to use own AI chips in its servers in 2024 — but it will still use Nvidia H100 GPUs as well in its datacenters for now

Meta Platforms, the parent company of Facebook, is planning to deploy its own custom-designed AI chips, called Artemis, in its data centers this year. This move aims to reduce Meta's reliance on Nvidia's H100 chips and control the rising costs of running AI workloads. The successful deployment of its own chip could save Meta hundreds of millions of dollars in energy costs and billions in chip purchasing costs. However, Meta will continue to use Nvidia's GPUs in its data centers, with plans to have 350,000 H100 processors in service by the end of the year.

Microsoft is building its own server tech — networking hardware will lessen reliance on Nvidia and possibly supercharge AI speeds

Techradar

Techradar · 1y ago

Microsoft is building its own server tech — networking hardware will lessen reliance on Nvidia and possibly supercharge AI speeds

Microsoft is reportedly developing its own network cards to boost the performance of its Maia AI server chip, according to a report by The Information. The project is believed to be led by Pradeep Sindhu, CVP Silicon at Microsoft and CEO of Fungible. By creating its own network cards, Microsoft could reduce its reliance on Nvidia, a crucial components provider for the industry. The new network card aims to optimize the performance of Microsoft's Maia AI server chip, potentially speeding up training time for OpenAI's models on Microsoft servers and improving cost efficiency.

Tesla shuts down Dojo: Here’s all you need to know - The Economic Times

Economic Times

Economic Times · 15d ago

Tesla shuts down Dojo: Here’s all you need to know - The Economic Times

Tesla has shut down its Dojo supercomputer project, ending its efforts to develop in-house chips for self-driving technology. This decision follows the departure of about 20 team members to form a new AI company, DensityAI. CEO Elon Musk emphasized the importance of consolidating resources rather than pursuing multiple AI chip designs. Tesla will now rely on external tech companies like Samsung, Nvidia, and AMD for chip production and computing needs.

China offers AI computing ‘vouchers’ to its underpowered start-ups

Techradar

Techradar · 1y ago

China offers AI computing ‘vouchers’ to its underpowered start-ups

China is introducing a new initiative to support AI startups by providing cloud computing vouchers to help them overcome chip shortages and reduce the financial burden of building their own data centers. The vouchers, valued between $140,000 and $280,000, will enable AI businesses to access essential infrastructure such as data centers for training and running AI models. This move comes in response to chip shortages caused by geopolitical tensions and US restrictions. In addition, there are reports of potential subsidies for AI groups using domestic chips to reduce reliance on foreign companies. However, questions remain about the sustainability of China's AI growth given the ongoing chip shortage.

Tech giants form AI group to counter Nvidia with new interconnect standard

Arstechnica

Arstechnica · 1y ago

Tech giants form AI group to counter Nvidia with new interconnect standard

Several major tech companies, including Google, Intel, Microsoft, Meta, AMD, Hewlett Packard Enterprise, Cisco, and Broadcom, have formed the Ultra Accelerator Link (UALink) Promoter Group. The group aims to develop an open standard interconnect technology for AI accelerator chips in data centers as an alternative to Nvidia's proprietary NVLink. The goal is to enable multiple companies to contribute and develop AI hardware advancements, reducing reliance on Nvidia's ecosystem and providing more options in the AI chip market. UALink 1.0, the first version of the proposed standard, is expected to improve speed and reduce latency compared to existing interconnect specifications.

Trackers

Active Indian VC’s

OG Capital Email

With a hands-on approach, OG Capital aims to invest in over 20 promising...

Accel Partners Email

Early and growth-stage investments in disruptive technology companies with...

Early-stage venture capital firm investing in technology startups in India. Focus on...

Access All Trackers

Startup Showcase Winners

June 2025

Helping your parents when you are miles away

The Pit Stop Your Cravings Deserve

The next generation E-commerce platform

Enter Ongoing Startup Showcase

Top Users

Trending News on Medial

Download the medial app to read full posts, comements and news.