Meta Platforms Inc META is testing its first in-house chip for training artificial intelligence systems. This marks its efforts to reduce reliance on suppliers like Nvidia Corp NVDA and bring down its infrastructure costs to offset the impact of its bets on AI tools to drive growth.

Meta tapped Taiwan Semiconductor Manufacturing Co TSM to produce the chip, Reuters reported, citing unnamed sources familiar with the matter.

Also Read: Microchip Launches New PIC32A Chips For AI, Auto, Medical, and Industrial Applications

Meta told Reuters they want to use their chips by 2026 for training.

In 2024, Meta started using a Meta Training and Inference Accelerator (MTIA) chip to perform inference.

As with the inference chip, the training chip aims to start with recommendation systems and later use them for generative AI products like chatbot Meta AI.

The Facebook and Instagram parent company has begun deploying a small chip and plans to ramp up production for wide-scale use if the test goes well.

Meta expects total 2025 expenses of $114 billion—$119 billion, including up to $65 billion in capital expenditures driven by AI infrastructure spending.

Meta previously scrapped an in-house custom inference chip.

Since then, Meta has remained one of Nvidia’s leading customers.

The AI chip market was worth $61.45 billion in 2023 and is expected to reach $621.15 billion by 2032, growing at a CAGR of 29.4% over the forecast period 2024-2032.

Big Tech giants remain engaged in their endeavors to develop in-house chips to reduce dependence on suppliers.

Reportedly, OpenAI is nearly ready to finalize the design for its first proprietary AI chip, which could reduce the company’s reliance on Nvidia.

OpenAI is on track to achieve its goal of commercializing at Taiwan Semiconductor by 2026. 

Apple Inc AAPL reportedly collaborated with Taiwan Semiconductor to launch in-house Wi-Fi and Bluetooth chips by 2025.

Price Action: META stock is up 1% at $604.00 at last check Tuesday.

Also Read:

Photo: Shutterstock