Weekly News

Nvidia acquires AI chip challenger Groq’s tech for $20B.

A turbulent week saw hardware bankruptcies from iRobot, Luminar, and Rad Power, while AI advanced with Luma’s new video model, GPT-5.2-Codex, app submissions to ChatGPT, and new ChatGPT Images. Meta also updated its AI glasses and released the SAM audio model.

AI continues to rapidly reshape industries, with BBVA, Instacart, and Disney partnering with OpenAI to transform banking, shopping, and entertainment, alongside the launch of GPT-5.2. At the same time, Google unveiled its most advanced research agent, AI leaders were named TIME’s 2025 Person of the Year, OpenAI expanded certifications, and copyright tensions around AI intensified.

Meta drives major AI updates with the Limitless acquisition, real-time news in Meta AI, and a new 24/7 support hub. OpenAI expands with an Australia launch and the Neptune acquisition, while Snowflake’s $200M Anthropic deal, EY–NVIDIA’s physical AI platform, Accenture’s OpenAI partnership, and UK–Germany quantum plans round out a big week in tech.

Instagram added global creator tools, ChatGPT launched shopping research, and GPT-5 helped solve a decades-old math problem. Major funding poured into US and Indian AI ecosystems, while Australia boosted AI safety and global businesses gained broader data residency options.

AI developments are speeding up as OpenAI teams with Target and Intuit and advances its tools with GPT-5.1-Codex-Max, while Meta unveils SAM 3 and SAM 3D for interactive 3D worlds. Meanwhile, the Trump administration may step back from opposing state-level AI regulations.

A wave of major AI updates arrived this week, from OpenAI’s Ireland launch, GPT-5.1 upgrades, and new ChatGPT features to Google’s Private AI Compute and Apple’s crackdown on third-party AI data sharing. Industry momentum is accelerating as Cursor raises $2.3B, ElevenLabs partners with celebrities, and Chinese models like Baidu’s ERNIE and Moonshot claim to surpass GPT and Gemini.

Tinder uses its users' Camera Roll photos for personalisation, Apple is close to a $1B deal to use Google AI in Siri, SoftBank and OpenAI formed a Japan venture, Perplexity will power Snapchat search for $400M, WhatsApp arrives on Apple Watch, and OpenAI is spreading a $600B cloud investment across major providers.

Nvidia plans to invest up to $1B in Poolside, while Google will offer free AI Pro access to Jio users in India. OpenAI introduced Aardvark, a new agentic security researcher, and launched gpt-oss-safeguard for open-source safety. Cursor released version 2.0 with a multi-agent Composer model, and ChatGPT Go will be free for a year in India.

OpenAI is expanding rapidly with a new generative music tool, the acquisition of Sky’s maker, and the launch of ChatGPT Atlas, a browser with ChatGPT built in. Meanwhile, Intel’s foundry recovery, Instagram’s new AI editing tools, and a cloud partnership between Anthropic and Google highlight broader AI momentum across the tech industry.

OpenAI’s Sora tops the App Store as AI agent competition grows with Google’s Jules. OpenAI partners with Japan, MIT links AI to reduced brain activity, and phishing detection emerges as a 2026 cybersecurity focus. ChatGPT adds parental controls and instant checkout.

Penske Media is suing Google over AI-generated summaries, while Google's practices face further criticism from the People CEO. Meanwhile, OpenAI announced new initiatives including OpenAI Grove and a partnership in APAC, as Perplexity raised $200M at a $20B valuation and xAI laid off 500 workers.

OpenAI reorganised research team behind ChatGPT’s personality, launched ‘OpenAI for Greece’ with Greek Government, announced AI-powered hiring platform, and shared its research outome on why language models hallucinate. Warner Bros sues Midjourney for stealing Superman and Switzerland releases 100% open AI model.

India boosts AI with Llama solutions, OpenAI adds real-time updates and safety findings, Nvidia hits $46.7B revenue, Nous releases Hermes 4, and Intel faces U.S. foundry sale limits.

Elon Musk’s xAI open-sourced Grok 2.5 and made moves around OpenAI, while rivals like Anthropic, ByteDance, and DeepSeek launched upgraded enterprise and open-source models. Nvidia also released a compact open model, Nemotron-Nano-9B-v2, marking rapid competition in AI releases.

ChatGPT app earns $2B; Meta criticized for AI chat rules with minors. Google launches Gemma 3 270M and adds Gemini personalization. GPT-4o returns; GPT-5 updated on the fly. Claude Sonnet 4 handles full software projects. Sam Altman challenges Neuralink.

OpenAI launched GPT-5 and an open model, gpt-oss, amid a bumpy rollout that led to bringing GPT-4o back. Meta bought AI audio startup WaveForms, Tesla ended its Dojo program, and ChatGPT became available to the entire U.S. federal workforce.

Anthropic blocked OpenAI from Claude and took the lead in enterprise LLMs; OpenAI raised $8.3B at a $300B valuation. New AI models launched from Gemini, Deep Cogito, Meta, and Z.ai’s open-source GLM-4.5. ChatGPT queries are now indexed by Google, and ads are being embedded in AI chats.

No ChatGPT therapy confidentiality, Anthropic safety agents, a 100x faster architecture, America’s AI plan, Alibaba’s Qwen milestone, OpenAI–Oracle data deal, Gemini 2.5 efficiency, OpenAI–UK partnership, and longer reasoning reducing accuracy.

Netflix used generative AI, Anysphere acquired Koala, and OpenAI launched an agentic model, Meta rejected EU AI rules, Claude Code added limits, and LLMs were found to change answers under pressure.

OpenAI delays its open model and cancels the Windsurf deal, with Google absorbing key staff. AI adoption accelerates across industries, from Tesla and AWS to Goldman Sachs, while new research tackles AI memory and data integration. Meanwhile, concerns grow over Nvidia’s China ties, and OpenAI joins the EU’s responsible AI efforts.

The European Union's landmark rules on artificial intelligence will be rolled out.... Google has been hit by an EU antitrust complaint over its AI.... Other developments include openai co-founder Ilya Sutskever says he is now formally ceo of ssi,... and business insider has learned meta is training customizable chatbots to be more....