Models & Labs
Meta to Begin Production of AI Chips in September
TechCrunch AIhigh confidence
Why it matters
- →Meta's move to produce its own AI chips could significantly reduce its dependency on external GPU suppliers like Nvidia and AMD.
- →The modular design of the chips allows for adaptability to rapidly changing AI workloads, potentially enhancing Meta's AI capabilities.
- →This development is part of a larger industry trend where major tech companies are investing in proprietary chip development to meet growing AI demands.
Meta plans to begin production of its latest AI-specific chips in September, as part of its strategy to reduce GPU costs amid a component shortage. The chips, developed under the Meta Training and Inference Accelerator program, are designed with a modular approach to adapt to changing AI needs. Meta is working with Broadcom on the design and TSMC on manufacturing. This initiative is part of Meta's broader effort to secure computing capacity for its AI models, reflecting a trend among tech giants to develop proprietary chips.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
More from TechCrunch AI
© TechCrunch AIAgentsagents
Meta Muse AI Agent Now Places US Business Calls
Meta’s consumer agent Muse is rolling out voice calling to US businesses for paid subscribers, targeting the tedious task of customer service navigation. With over 10 million users, Meta is betting that AI’s infinite patience will replace human frustration with phone trees and hold times. The feature leverages improved voice mimicry to interact with receptionists, though many still hang up upon detecting non-human voices. This moves Muse from text-based assistance into active, autonomous execution of real-world tasks, marking a significant step toward mainstream consumer AI utility.
TechCrunch AI
© TechCrunch AIMore in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
