Introducing the first AI-native CRM

Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.

With AI at the core, Attio lets you:

• Prospect and route leads with research agents

• Get real-time insights during customer calls

• Build powerful automations for your complex workflows

Join industry leaders like Granola, Taskrabbit, Flatfile and more.

👉Try Attio Pro for free

Today:

• Anthropic Eyes $70B Revenue, $17B Cash Flow

• OpenAI’s Sora App Lands on Android

• Amazon Slaps Perplexity with Legal Threat Over AI Shopping Bot

• NVIDIA Brings AI-Powered Traffic and Safety to Global Cities

• Microsoft Inks AI Licensing Deal with People Inc. as Web Shifts

Anthropic Eyes $70B Revenue, $17B Cash Flow

Anthropic is forecasting $70 billion in revenue and $17 billion in cash flow by 2028, driven by rapid growth in API sales and enterprise AI deals. Its Claude models are gaining traction with major firms like Microsoft, Deloitte, and Salesforce, signaling strong business demand.

KEY POINTS

• Massive Growth Forecast: Anthropic aims for $70B revenue and 77% profit margins by 2028, with $17B projected cash flow.

• Enterprise Expansion: Claude is rolling out across Microsoft 365, Deloitte, Cognizant, and Salesforce platforms, showing strong B2B momentum.

• Model Development: Smaller, cost-efficient Claude 4.5 models (Sonnet, Haiku) and new Enterprise Search tools position Anthropic for scale.

Why it matters

Anthropic’s success signals rising demand for business-friendly AI tools. Its rapid revenue growth challenges OpenAI, showing that leaner, targeted models like Claude 4.5 can win large enterprise clients. This AI race affects how fast AI spreads in workplaces — and who controls the tools.

OpenAI’s Sora App Lands on Android

OpenAI’s AI video app Sora is now available for Android in the U.S., Canada, and Asia. Already popular on iOS, Sora lets users create shareable AI-generated videos, including using their own likeness. OpenAI faces challenges over deepfakes, copyright, and a legal fight over the “Cameo” feature.

KEY POINTS

• Sora launches on Android in multiple countries, expanding reach beyond iOS and boosting OpenAI’s presence in short-form video.

• Deepfake and copyright concerns prompt policy shifts, including halting content of Martin Luther King Jr. and switching to opt-in for copyrighted characters.

• New features coming soon, like pet videos, video stitching, and more personalized social feeds, aim to rival TikTok and Meta’s Vibes.

Why it matters
This launch puts OpenAI in direct competition with TikTok, Instagram, and Meta’s new AI video tools. Sora’s growth shows how fast AI is reshaping social media. But it also raises big questions about how to handle deepfakes, privacy, and creative ownership.

Amazon Slaps Perplexity with Legal Threat Over AI Shopping Bot

Amazon sent a legal warning to Perplexity AI, demanding it block its Comet browser from making purchases on Amazon. Perplexity called the move anti-competitive and accused Amazon of trying to limit innovation and user choice in the growing market for AI-powered shopping assistants.

KEY POINTS

• Legal Threat: Amazon told Perplexity to stop its Comet assistant from shopping on Amazon’s site, citing poor experience and lack of transparency.

• Perplexity Pushes Back: The startup rejected the claims and said Amazon is using legal pressure to crush competition and protect its ad-driven business model.

• Broader Battle: This highlights a larger clash over AI agents reshaping web browsing and who controls online commerce, with Amazon also building its own shopping AI tools.

Why it matters
This fight shows how tech giants are trying to control how AI assistants interact with their platforms. If companies like Amazon block tools like Comet, it could limit consumer freedom and slow AI innovation in everyday tasks like online shopping.

🧠RESEARCH

π_RL: Online RL Fine-tuning for Flow-based Vision-Language-Action Models

The π_RL framework improves robot learning by using reinforcement learning (RL) instead of traditional methods. It trains flow-based vision-language-action models more efficiently, handling tough problems in action prediction. Tests show major performance boosts in tasks like object manipulation, proving π_RL’s success in scaling robotic skill learning with better generalization.

INT v.s. FP: A Comprehensive Study of Fine-Grained Low-bit Quantization Formats

This study compares low-bit floating-point (FP) and integer (INT) formats for large language models. It finds that fine-grained INT formats like MXINT8 outperform FP in both accuracy and efficiency. While FP works better for coarser settings, INT shines in finer setups, challenging current hardware trends and guiding future AI accelerator designs.

Every Activation Boosted: Scaling General Reasoner to 1 Trillion Open Language Foundation

Ling 2.0 is a new family of reasoning-focused AI models that scale up to one trillion parameters using sparse Mixture-of-Experts (MoE). With smart training techniques and efficient architecture, it boosts reasoning power while cutting compute costs. Ling-1T sets a new benchmark for performance and efficiency in large-scale thinking models.

🛠️TOP TOOLS

Each listing includes a tutorial so you can get started right away, whether you’re a beginner or a pro.

2Short AI: Turn Long Videos into Viral Shorts - finds the best moments in your long‑form videos and turns them into short

Affogato AI: Agentic Video Ad Generator for TikTok, Reels & Shorts - agent-driven video tool built to turn a simple idea into ready‑to‑post short‑form ads

Adrenaline AI: AI Codebase Q&A & Visualizer - AI assistant for developers that lets you ask technical questions, chat with your codebase

📲SOCIAL MEDIA

🗞️MORE NEWS

• NVIDIA and partners are using AI to help cities like Dublin and Ho Chi Minh manage traffic, safety, and planning. Their new tools simulate real-world conditions, automate inspections, and improve urban mobility with real-time data.

• People Inc. signed a new AI licensing deal with Microsoft, joining its content marketplace as Google traffic drops. The publisher uses Cloudflare to block AI crawlers, pressuring companies to pay for content use.

• Google’s Project Suncatcher explores powering AI in space using solar satellites equipped with TPUs and optical links. It aims to create scalable, eco-friendly ML compute by overcoming challenges in bandwidth, radiation, and orbital control.

• AWS announced Fastnet, a new high-capacity subsea cable linking Maryland and Ireland by 2028. It boosts global cloud and AI service reliability, offering faster, more secure data routes while supporting local communities.

• Amazon is integrating Alexa Plus into its Music app, enabling users to ask for vibe-based playlists, obscure song info, and tailored recommendations. The AI assistant understands moods, lyrics, instruments, and exclusions for personalized music discovery.

• Microsoft has launched its first in-house image generator, MAI-Image-1, in Bing Image Creator and Copilot Audio Expressions. Known for photorealistic and nature imagery, it's fast and detailed but not yet available in the EU.