Last Week’s AI News #10

Artificial Intelligence (AI) continues to transform the world of business, technology, and creative industries - from advanced language models to 3D visualizations and strategic investments. Last week was particularly eventful: new models were released, startups secured massive funding, and tech giants reinforced their positions in the AI space.

In this roundup, you’ll find the key developments and trends shaping the AI ecosystem, offering businesses new opportunities, insights, and competitive advantages. Whether you’re interested in integrating AI into business processes, automating creative workflows, or exploring long-term investments, this overview has all the essential information.

Quick Highlights:

• OpenAI launched ChatGPT group chats

• The 2wai model sparks ethical debate: digital resurrection of the deceased

• Google unveiled Gemini 3: the next generation AI and leadership positioning

• Microsoft and Nvidia invest $15B in Anthropic

• Codex-Max outperforms Gemini 3 Pro in coding tests

• HUMAIN startups partner with xAI and Nvidia: 600k GPUs and 500+ MW data center

• Meta introduces SAM 3 and SAM 3D: new 3D computer vision models

• Nano Banana Pro: next-level image and text generation

• And more AI news from the past week

OpenAI Launches ChatGPT Group Chats

OpenAI has introduced a group chat feature for all subscription plans, allowing up to 20 users to collaborate simultaneously in a single ChatGPT chat.

Chats are shared via invitation links, with ChatGPT monitoring the conversation and stepping in when needed or directly mentioned. Response limits apply, and privacy features isolate group sessions from personal memory - ChatGPT does not retain information from collaborative chats.

The feature was tested in four Asia-Pacific markets and is now available to Free, Go, Plus, and Pro users.

Why it matters for businesses:

• Enhanced collaboration: Group discussions and projects can happen in real time with AI assistance.

• Natural AI integration: Models can participate as full-fledged team members, contributing to idea generation and problem-solving.

• Privacy control: Group chats are isolated from personal memory, ensuring confidentiality.

• Accessible to all plans: Available to both free and paid subscribers.

Google Introduces Gemini 3: Next-Gen AI and Leadership

Google unveiled Gemini 3, its most advanced AI model yet, describing it as the start of a “new era of intelligence.” The model achieved record results across multiple benchmarks and was released alongside Antigravity, a new agent-first development platform.

Gemini 3 and Deep Think reached top scores on complex reasoning tests like Humanity’s Last Exam and ARC-AGI-2, surpassing GPT-5. The model also excels in scientific knowledge, mathematics, multimodal reasoning, and tool usage, lagging only slightly behind Claude Sonnet 4.5 in coding tests.

One of the standout innovations is real-time generative UI and visual layout creation, integrated directly into AI mode search. Antigravity is a free agent-based coding platform with browser control, asynchronous tasks, and multi-agent orchestration.

Why it matters for businesses:

• Stronger AI models = more automation: Businesses can delegate increasingly complex analytics, development, and decision-making tasks.

• Generative UI transforms product creation: Prototypes and designs are generated instantly, reducing costs and speeding up cycles.

• Antigravity agents redefine workflows: Multi-agent systems plan and execute processes across the organization.

• AI competition intensifies: Google’s challenge to OpenAI accelerates innovation and AI strategy reviews.

Microsoft and Nvidia Invest $15B in Anthropic

Microsoft and Nvidia announced a major strategic partnership with Anthropic, totaling $15B in investments, alongside a commitment of $30B in Azure cloud resources. Nvidia will invest up to $10B, Microsoft up to $5B, bringing Claude’s valuation close to $350B. Claude models will be available through Azure AI Foundry, making Anthropic the only “frontier” model provider operating across all three major cloud platforms.

Together, Nvidia and Anthropic will develop new AI chips optimized for Claude, with Anthropic committing $30B in Azure compute resources and 1GW capacity. Satya Nadella emphasized that the AI industry must move beyond “zero-sum” rhetoric toward collective progress.

Why it matters for businesses:

• AI power becomes more accessible: Claude on Azure expands choice and competition among advanced models.

• Multi-cloud strategies become practical: Anthropic models operate across all three major clouds, easing migration and flexibility.

• Specialized AI chips reduce cost and increase speed: Critical for businesses managing expensive compute workloads.

• New industry alliances form: Microsoft, Nvidia, and Anthropic strengthen AI ecosystem consolidation.

Codex-Max Surpasses Gemini 3 Pro in Coding Tests

OpenAI released GPT-5.1 Codex-Max, an upgraded agent-based coding model using new “compaction” technology, enabling longer tasks and over 24-hour development sessions.

Codex-Max significantly improved results, outperforming even Gemini 3 Pro in coding tests. It uses 30% fewer tokens and runs much faster in real-world tasks due to more efficient reasoning. The compaction technology trims session history while retaining context, allowing the model to handle millions of tokens continuously.

It is available via Codex CLI and IDE extensions for Plus, Pro, and Enterprise users, with API access coming soon.

Why it matters for businesses:

• Longer development sessions: AI can be applied to long-term projects without losing context.

• Efficient token usage: Reduces compute costs and speeds up workflows.

• Competitive edge: Coding performance often lags other AI tasks—Codex-Max closes this gap.

• Flexible integration: Ready for IDE and CLI, with upcoming API support.

HUMAIN Startups Partner with xAI and Nvidia: 600k GPUs & 500+ MW Data Center

Saudi Arabia’s HUMAIN announced AI partnerships with xAI, Nvidia, and other major companies, planning 600k GPUs, a 500+ MW data center, and nationwide Grok deployment. xAI will build the first international data center in Saudi Arabia, while HUMAIN expands into the U.S. with Nvidia data centers. AWS will supply 150k chips for the “AI Zone” data center, with AMD, Cisco, Adobe, and Qualcomm also collaborating.

Luma AI secured $900M funding for a 2GW supercluster for multimodal model training, starting in 2026. The U.S. approved AI chip exports to HUMAIN and UAE’s G42, opening regional export opportunities.

Why it matters for businesses:

• Regional AI leadership grows: Saudi Arabia emerges as a key AI player, enabling international partnerships.

• Massive infrastructure = more capability: Large-scale GPUs and data centers support complex AI projects.

• Global collaboration: Partnerships expand the innovation ecosystem.

• Export potential: U.S. approval boosts accessibility and investment.

Meta Introduces SAM 3 and SAM 3D: New 3D Computer Vision Models

Meta launched SAM 3 and SAM 3D, models that segment and reconstruct objects and people in 3D from text descriptions. The new Segment Anything Playground allows free testing, and SAM 3 code and weights are open-source.

SAM 3D reconstructs scenes from a single image, outperforming competitors by 5:1. Technology is already integrated into Facebook Marketplace “View in Room” and will soon expand to Edits and Vibes apps.

Why it matters for businesses:

• Enhanced computer vision: Precise segmentation and 3D reconstruction enable e-commerce, creative work, robotics, and visualizations.

• Open-source advantage: Companies can integrate models into products, saving time and cost.

• Product integration: Features like “View in Room” improve user experience and engagement.

• Innovation acceleration: Public access encourages developer community growth.

Nano Banana Pro: Next-Level Image and Text Generation

Google released Nano Banana Pro, an image model based on Gemini 3 with professional editing, 4K resolution, advanced text, and world knowledge.

The model supports up to 14 visual references, maintains up to five people’s identities, and allows precise control over camera angles, focus, and lighting. Text can be long, multilingual, with various fonts and layouts. Integration with Google Search ensures real-time accuracy.

Why it matters for businesses:

• Professional visuals and text: Ideal for high-quality infographics, ads, and presentations.

• Internet integration: Real-time world knowledge ensures accuracy.

• Multi-reference and identity support: Simplifies new compositions and visual storytelling.

• Multi-language and font support: Suitable for international projects and layered designs.

2wai Model Launch Sparks Ethical Debate: Digital Resurrection of the Deceased

Former Disney Channel actor Calum Worthy introduced 2wai, an AI platform creating interactive avatars of deceased loved ones. These “HoloAvatars” can speak, react, and “participate” in life events, reminiscent of a techno-dystopian Black Mirror scenario.

The app generates avatars from minutes of video and lets users interact with digital representations of loved ones. A viral clip showing an AI grandmother interacting with a grandson sparked thousands of critical reactions on X.

Beta is now free on the Apple App Store, with a planned paid subscription and Android rollout. Critics call it “morally distorted,” warning that it exploits grief and simulates deceased individuals without consent.

Why it matters for businesses:

• New market: Emotional experience, memory, and “digital immortality” services could become multibillion-dollar segments.

• Ethical risk: Companies handling human likenesses, voices, or genetic data must establish consent, data management, and posthumous rights policies.

• Regulatory risk: Technologies like this accelerate AI legislation; businesses must adapt.

• Consumer trust: Overly invasive or emotionally manipulative applications can damage brand reputation.

• Competitive signal: Investment in sensitive areas indicates AI personalization is reaching new heights.

Other AI Developments Last Week

• Google: $40B Texas investment in data centers and AI infrastructure by 2027

• Microsoft: Agent 365 platform for AI agent management, security, and compliance

• Lambda: $1.5B funding for AI data center expansion

• OpenAI: GPT-5.1 Pro rollout for improved writing, analytics, and business tasks

• Adobe acquires Semrush for $1.9B

• Google AI holiday shopping improvements

• Replit Design tool launches using Gemini 3

• ByteDance Seed: Depth Anything 3 for 3D depth mapping

• Sakana AI valued at $2.6B, Japan’s most valuable private AI company

This week’s roundup is just the beginning - subscribe to our newsletter to get regular AI news delivered straight to your inbox and stay ahead of the latest developments in the AI industry.

Don't miss news from us

Or join AI Mastery Program for business leaders >