Listen to Full Audio at

🚀 Welcome to the AI Unraveled Weekly Rundown.

This episode is made possible by AIRIA. 🛑 The “AI Sprawl” is here. Can you see it? From Gemini 3.1 to OpenClaw agents and Alibaba’s Qwen, your employees are now using a fragmented, high-speed AI stack. AIRIA is the industry-leading Control Plane for enterprise AI. It allows you to orchestrate multiple models, discover “Shadow AI,” and provide unified security and cost control in one single environment. 👉 Orchestrate Your Stack: Get the Airia Demo here.

Weekly Highlights:

This was the week the AI industry moved from “Chatbot” to “Agentic Operating System.” We cover Google’s massive reasoning breakthrough with Gemini 3.1 Pro, the $30 billion Nvidia-OpenAI equity deal, and OpenAI’s move into hardware with a Jony Ive-inspired smart speaker.

Key Segments:

• The Intelligence Crown: Gemini 3.1 Pro hits a 77.1% ARC-AGI score, doubling its reasoning power.

• The Wealth Explosion: Nvidia takes a $30B stake in OpenAI as average employee compensation hits $1.5 million.

• Hardware & Privacy: OpenAI’s upcoming smart speaker, glasses, and “smart lamp” project.

• Cybersecurity Shakeup: Anthropic’s Claude Code Security sends legacy cybersecurity stocks (CrowdStrike, Okta) tumbling.

• AI Geopolitics: The U.S. launches a “Tech Corps” to counter Chinese AI exports in the global south.

• The Ad-Free Pivot: Perplexity drops ads entirely to save user trust, while Google and OpenAI lean in.

• Workplace Mandates: Accenture begins tying promotions to weekly AI tool usage.

Credits: This podcast is created and produced by Etienne Noumen, Senior Software Engineer and passionate Soccer dad from Canada.

Keywords: Gemini 3.1 Pro, Nvidia OpenAI Stake, Claude Code Security, OpenAI Smart Speaker, US Tech Corps, Project Silica, OpenClaw, Perplexity Ads, Accenture AI Promotions, AIRIA, Shadow AI, Reasoning Breakthrough, ARC-AGI-2

🚀 Reach the Architects of the AI Revolution

Want to reach 60,000+ Enterprise Architects and C-Suite leaders? Download our 2026 Media Kit and see how we simulate your product for the technical buyer: https://djamgamind.com/ai

Connect with the host Etienne Noumen: https://www.linkedin.com/in/enoumen/

⚗️ PRODUCTION NOTE: We Practice What We Preach: AI Unraveled is produced using a hybrid “Human-in-the-Loop” workflow. While all research, interviews, and strategic insights are curated by Etienne Noumen, we leverage advanced AI voice synthesis for our daily narration to ensure speed, consistency, and scale.

OpenAI plans an AI smart speaker with camera

• OpenAI is working on a $200 to $300 smart speaker with a built-in camera, part of a broader push into AI-powered hardware that also includes smart glasses and a smart lamp.

• The speaker could identify objects on a nearby table, listen to conversations, and support facial recognition for authenticating purchases, with a team of over 200 employees building the project.

• The smart speaker could ship in early 2027 at the earliest, while smart glasses may not arrive until 2028, and the project has already faced delays over technical, privacy, and logistical issues.

Supreme Court blocks tariffs costing Apple billions

• The Supreme Court ruled 6 to 3 that Trump’s sweeping “Liberation Day” tariffs were illegal because he imposed them without Congress, costing Apple around $2 billion in fees paid to the government.

• Chief Justice John Roberts wrote that IEEPA does not authorize the president to impose tariffs, and Justice Kavanaugh warned that the refund process for billions collected from importers will be a “mess.”

• Beyond direct tariff fees, companies like Apple spent hundreds of millions reorganizing manufacturing and distribution, while smaller US firms were put out of business with no clear path to recovery.

OpenAI targets $600 billion in compute spending by 2030

• OpenAI now projects it will spend a total of $665 billion on training and running AI models through 2030, raising its cumulative cash burn estimate by roughly $111 billion compared to previous forecasts.

• Revenue more than tripled to $13.1 billion in 2025, but adjusted gross margin fell to 33 percent from 40 percent the prior year, as inference costs alone quadrupled during that period.

• OpenAI does not expect to become cash-flow positive until 2030 and is negotiating a funding round exceeding $100 billion at a $750 billion valuation, with SoftBank, Amazon, Nvidia, and Microsoft involved.

Anthropic launches Claude Code Security

• Anthropic has launched Claude Code Security, a tool built into Claude Code that scans codebases for security vulnerabilities and suggests patches, though humans must review and approve every fix.

• The company says Claude Opus 4.6 has found over 500 vulnerabilities in production open-source codebases, including bugs that went undetected for decades despite years of expert scrutiny.

• Cybersecurity stocks dropped sharply after the announcement, with CrowdStrike falling 8 percent, Cloudflare 8.1 percent, Okta 9.2 percent, and SailPoint 9.4 percent on the same day.

US plans Tech Corps to counter China AI exports

• The U.S. government is creating a “Tech Corps” volunteer program, run through the Peace Corps, to send people abroad and promote American AI products as it competes with China for influence.

• Chinese open models from Alibaba, Minimax, and Moonshot rank among the most downloaded on Hugging Face and OpenRouter because they are cheaper, customizable, and can run on local infrastructure.

• Volunteers with STEM degrees will spend one to two years helping integrate American AI into farms, hospitals, and schools, but a Brookings fellow said persuasion alone won’t overcome economic realities.

Nvidia nears $30 billion investment in OpenAI

• Nvidia is close to finalizing a $30 billion equity stake in OpenAI, replacing a previously announced $100 billion infrastructure deal, with the new agreement potentially closing as early as this weekend.

• The restructured deal lets Nvidia take direct equity in OpenAI at a $730 billion pre-money valuation while still selling chips to the company, giving it both ownership upside and hardware revenue.

• Wall Street analysts remain bullish, with 57 out of 61 ratings at Buy and a 12-month average NVDA price target of $253.88, representing a roughly 35 percent upside from current levels.

Google launches Gemini 3.1 Pro

• Google released Gemini 3.1 Pro today, a new model designed for complex tasks where a simple answer falls short, marking the first time Google has used a .1 increment between generations.

• The model scores 77.1% on ARC-AGI-2, which Google says is more than double the reasoning performance of 3 Pro, thanks to upgraded core intelligence that debuted in Gemini 3 Deep Think.

• Gemini 3.1 Pro is rolling out in the Gemini app, NotebookLM, and developer tools like Google AI Studio and Vertex AI, though it launches in preview before becoming generally available soon.

Microsoft stores data in glass for 10,000 years

• Microsoft’s Project Silica can store data inside a thin piece of borosilicate glass using laser technology, and the material could theoretically last 10,000 years without degrading.

• The system encodes information in phase voxels — three-dimensional equivalents of pixels — using femtosecond lasers, and a single glass chip can hold 4.8 terabytes of data.

• Project Silica was built for organizations like the National Archives and museums that need storage immune to malware, decay, and ongoing maintenance costs over long periods.

Accenture ties promotions to AI adoption

• Accenture is now tying employee promotions to regular use of its internal AI tools, tracking weekly logins and making that data a visible factor in summer promotion decisions.

• Senior staff appear more reluctant to adopt AI than junior employees, and some workers have called the tools ineffective “broken slop generators,” with one threatening to quit.

• Not all employees face the new rule equally — staff in 12 European countries and the US federal government contracts division are exempt from having AI adoption tracked for promotions.

Anthropic launches Claude Sonnet 4.6 as new default model

• Anthropic released Claude Sonnet 4.6 as the new default model for Free and Pro plan users, continuing the company’s pattern of updating its mid-size Sonnet model roughly every four months.

• The beta release includes a 1 million token context window, double the previous largest Sonnet window, which Anthropic says can hold entire codebases, lengthy contracts, or dozens of research papers.

• Sonnet 4.6 posted record benchmark scores in computer use and software engineering, plus a 60.4% on ARC-AGI-2, though it still trails Opus 4.6, Gemini 3 Deep Think, and a refined GPT 5.2.

Perplexity drops ads from its AI search engine

• Perplexity, the AI search startup that was among the first AI companies to introduce ads, is now dropping advertising from its product and has no plans to pursue it further.

• A Perplexity executive told the Financial Times that ads made users “start doubting everything,” undermining trust in the chatbot’s answers and making people less willing to pay for it.

• The move comes as competitors head the other direction: OpenAI recently started testing ads on ChatGPT, Google shows ads in AI search results, while Anthropic has committed to keeping Claude ad-free.

Google Gemini can now generate music with Lyria 3

• Google is rolling out a music-generation feature in the Gemini app, powered by DeepMind’s Lyria 3 model, which can create 30-second tracks with lyrics and cover art from text prompts.

• Users can upload a photo or video to generate a matching song, control elements like style, vocals, and tempo, and all output carries a SynthID watermark to identify AI-generated content.

• Google is also expanding Dream Track on YouTube globally for creators, though the tool won’t mimic artists directly — naming one in a prompt only inspires a track with similar style or mood.

SpaceX competes in Pentagon AI drone swarm contest

• SpaceX and its subsidiary xAI are competing in a secretive $100 million Pentagon contest to build voice-controlled, autonomous drone swarming technology, marking a new move into AI-enabled weapons development for Elon Musk.

• The six-month competition, launched by the Defense Innovation Unit and the new Defense Autonomous Warfare Group, will progress in five phases from software development to real-life testing of drones for offensive purposes.

• OpenAI is also involved through a partner company called Applied Intuition but limited to voice-command translation, while SpaceX and xAI are expected to work on the entire project together, raising concerns among defense officials.

OpenAI adds Lockdown mode to ChatGPT

• OpenAI has added a new Lockdown Mode and an “Elevated Risk” warning label to ChatGPT, both designed to protect users from prompt injection attacks that can trick AI into leaking sensitive data.

• Lockdown Mode is optional and deterministically limits or fully disables high-risk features like web browsing, while the “Elevated Risk” label warns users before actions like opening unverified external links.

• Business plan users on ChatGPT Enterprise, Edu, Healthcare, and Teachers already have the protection, but OpenAI said it will roll out to other users in the coming months across unknown payment tiers.

OpenClaw creator Peter Steinberger joins OpenAI

• Peter Steinberger, the creator of the autonomous AI tool OpenClaw, is joining OpenAI to work on what the company calls “the next generation of personal agents” for ChatGPT and other products.

• OpenClaw operates autonomously by accessing personal services like email and computer files to handle tasks such as clearing your inbox, and it sends updates through iMessage or WhatsApp.

• Altman confirmed that OpenClaw will continue as an open-source project with OpenAI “support,” though security experts have raised concerns about the tool’s broad access to users’ information and services.

Alibaba launches Qwen 3.5 AI model

• Alibaba released Qwen 3.5, a large language model designed for the “agentic AI era,” which the company says outperforms GPT-5.2, Claude Opus 4.5, and Gemini 3 Pro on several internal benchmarks.

• Qwen 3.5 is 60% cheaper than the previous version, up to eight times better at handling large workloads, and includes an open-weight model released under an Apache 2.0 license supporting 201 languages.

• The launch comes days after ByteDance updated its Doubao chatbot, which leads China with nearly 200 million users, while DeepSeek is expected to introduce a next-generation model soon.

What Else Happened in AI From February 15th to febraury 22 2026?

1. Pope Leo XIV has urged priests to not to use artificial intelligence to write their homilies or to seek “likes” on social media platforms like TikTok.

2. Google VP warns that two types of AI startups may not survive.

3. NVIDIA Releases DreamDojo: An Open-Source Robot World Model Trained on 44,711 Hours of Real-World Human Video Data.

4. Anthropic Launches Claude Code Security for AI-Powered Vulnerability Scanning.