Alphabet, Meta in Hollywood

Today:

• Alphabet, Meta in Hollywood

• Cohere Unveils Aya 23

• Amazon, Hugging Face Partnership

• Truecaller, Microsoft Collaboration

• 20,000 Startups Use Nvidia

• Nvidia’s Annual AI Chips

• Snowflake Acquires TruEra

Alphabet, Meta Offer Millions to Partner With Hollywood on AI

Alphabet and Meta are in talks with Hollywood studios to license content for AI-driven video generation software. They offer millions to create realistic scenes from text prompts. OpenAI, backed by Microsoft, is having similar discussions. The aim is to leverage AI's potential without losing control over content. 

Warner is considering licensing, while Disney and Netflix have declined. These discussions highlight the growing interest and investment in AI technology within the entertainment industry. Alphabet, Meta, and OpenAI have not commented on the ongoing negotiations.

BLOOMBERG

Cohere launches open weights AI model Aya 23 with support for nearly two dozen languages

Cohere for AI, the research arm of Canadian AI startup Cohere, has released the open weights for Aya 23, a new multilingual AI language model. Available in 8B and 35B parameter variants, Aya 23 supports 23 languages, including Arabic, Chinese, English, and Spanish. This release aims to improve performance over previous models by focusing on fewer languages with greater depth. 

Aya 23 surpasses other open models like Google’s Gemma in quality. The open weights allow researchers to fine-tune the model for their needs, enhancing multilingual AI applications. The models are available on Hugging Face under a permissive license.

VENTUREBEAT

Truecaller and Microsoft will let users make an AI voice to answer calls

Truecaller, a caller ID app, has partnered with Microsoft’s Azure AI Speech to let users create an AI version of their voice to answer calls. This feature, available through Truecaller’s AI Assistant, allows users to record their voice, which Azure AI Speech then uses to generate an AI replica. This innovation aims to add a personal touch to digital interactions. 

The AI Assistant, introduced in 2022, screens incoming calls and informs users of the call's purpose, allowing them to choose whether to answer or let the assistant respond. This personalization marks a significant enhancement in user experience.

THE VERGE

Nvidia says 20,000 GenAI startups are now building on its platform

Nvidia's CEO, Jensen Huang, reported a surge in generative AI (GenAI) startups, with 20,000 now using Nvidia’s accelerated computing platform. The demand for Nvidia’s GPUs is immense as companies across various sectors seek to leverage AI. Nvidia's new H100 "Hopper" and upcoming "Blackwell" GPUs promise significant speed improvements. 

Despite record revenue of $26 billion in Q1, Nvidia struggles to meet demand. Huang highlighted the strong financial returns for companies hosting AI models on Nvidia’s platform. Nvidia also made advances in AI networking with its Spectrum-X platform, further expanding its market reach.

VENTUREBEAT

Amazon, AI startup Hugging Face pair to use Amazon chips

Amazon Web Services (AWS) has partnered with AI startup Hugging Face to run thousands of AI models on Amazon's custom Inferentia2 chips. Hugging Face, valued at $4.5 billion, is a key platform for AI developers. This collaboration aims to enhance efficiency and cost-effectiveness in running AI models. 

While Nvidia leads in training models, AWS focuses on cost-effective inference operations using their chips. Jeff Boudier of Hugging Face emphasizes the importance of efficiency, while Matt Wood of AWS highlights the frequent need for inference, making Inferentia2 an ideal solution for AI developers.

REUTERS

Nvidia will now make new AI chips every year

Nvidia will now release new AI chips annually, shifting from its previous two-year cycle. CEO Jensen Huang announced this change during the Q1 2025 earnings call, highlighting the success of the H100 "Hopper" chip and the upcoming "Blackwell" chip. Nvidia's new strategy includes accelerating the production of CPUs, GPUs, and networking hardware. 

The company's recent profits, driven by AI chip sales, reached $14 billion in a single quarter. Nvidia's AI GPUs, compatible with existing data centers, will continue to meet high demand from various sectors, including automotive and consumer internet companies like Meta.

THE VERGE

Snowflake acquires TruEra to deliver LLM observability inside data cloud 

Snowflake has acquired AI startup TruEra to enhance its data cloud with tools for testing, debugging, and monitoring machine learning (ML) models and large language model (LLM) applications. TruEra, founded in 2019, provides observability solutions that address the "black box" problem in ML, ensuring accurate and relevant AI outputs. 

This acquisition marks Snowflake's continued investment in AI capabilities, aiming to offer an end-to-end platform for building trustworthy generative and predictive AI applications. TruEra’s technology will help Snowflake users identify and mitigate performance gaps in AI models, complementing Snowflake’s existing data governance functionalities.

VENTUREBEAT