
Meta in Talks with TSMC to Launch its First In-House Own Chip
The company aims to use in-house chips by 2026 for both training and inference tasks, Reuters reported.
Meta AI: Cutting-edge AI assistant powered by Llama 3. Available on Facebook, Instagram, WhatsApp, and Messenger. Offers text generation, image creation, and web search integration. Enhances productivity, sparks creativity, and facilitates connections. Currently expanding globally with improved features and real-time image generation capabilities.
The company aims to use in-house chips by 2026 for both training and inference tasks, Reuters reported.
They are also reportedly launching a standalone app for its AI assistant.
Meta’s new experimental framework aims to improve LLM agents that help in research.
Like Meta, tech giants like Apple, Nvidia, X, FigureAI, and Apptronik are investing heavily in robotics to shape the future of AI-driven automation.
Meanwhile, even OpenAI is building its custom chip.
The framework is structured around a three-stage process: anticipating risks, evaluating and mitigating risks, and deciding whether to release, restrict, or halt the model.
The company is maintaining a $60 billion to $65 billion capital expenditure plan for 2025 to strengthen its AI initiatives.
Meta will deploy over 1.3 million GPUs by year-end to power its AI initiatives.
NCLAT stayed the order directing WhatsApp not to share data with Meta’s products for five years. However, Meta has been ordered to deposit 50% of the ₹213.14 crore penalty within two weeks.
The Pragati: AI for Impact CSR initiative offers grants of ₹40 lakh to ₹1.5 crore to 10-12 startups each to use AI for challenges in healthcare, education, agriculture and inclusion.
Elon Musk and Jeff Bezos are new best friends in town.
“AI will not replace mid-level engineers entirely but will augment their capabilities.”
Apollo models excel at video tasks by addressing key challenges, including how videos are sampled, encoded, and trained.
The model excels in both text-based and image-based medical tasks, such as analysing X-rays, CT scans, and MRIs.
Samsung’s first XR headset, codenamed Project Moohan, is expected to debut next year.
Is it as good as OpenAI’s o1, though?
The researchers found that the accuracy rates could be 30-40% higher than big techs’ initial low rates (10-11%).
At redBus, OpenAI’s models power chatbot interactions in India, while Anthropic’s Claude is being tested in Malaysia.
In addition to Isaac Lab, six specialised workflows for their humanoid development platform, Project GR00T, were also released.
Meta has introduced an AI model that can evaluate the work of other AI models.
Meta’s three new research artefacts bring touch, dexterity and interaction to robots.
The rise of AI-driven search engines, driven by ChatGPT and Perplexity, poses a significant threat to Google’s search and ad revenue dominance.
Meta’s stance to help government agencies leverage their open-source AI models comes after China’s rumoured adoption of Llama for military use.
AI now recommends over 50% of the content on Instagram and 30% of posts on Facebook, according to Meta chief Mark Zuckerberg.
“We’re training the Llama 4 models on a cluster that is bigger than 100k H100s or bigger than anything that I’ve seen reported,” said Mark Zuckerberg on the earnings call.
This model can fill in missing frames in videos, turn a single image into a moving scene, and even extend a short clip by adding natural, continuous frames.
Meanwhile, OpenAI might have already cracked AGI internally as it delays its release, waiting for Google, Meta and others to catch up.
Like NotebookLM, it generates conversational, podcast-style summaries from uploaded text files.
As an early adopter of Llama 3.1 and 3.2 models, Infosys is integrating these models with Infosys Topaz, the in-house AI platform, to create tools that deliver business value.
The new model allows users to generate more natural-sounding speech and demonstrates the capability to learn tasks across different modalities, including ASR, TTS, and speech classification.
AIM’s Happy Llama 2025 brings AI startups, investors, and experts under one roof to drive meaningful growth and innovation.
© Analytics India Magazine Pvt Ltd & AIM Media House LLC 2025