AI Insider’s Week in Review: News From ChatGPT, Amazon, Google, Nvidia, Plus the Latest Funding Rounds

Enterprise AI is tilting from experimentation to deployment. ChatGPT usage hit 700 million weekly users, with work use converging on decision support and writing, while Amazon rolled out an always-on agent to run seller ops and Google pushed AP2 to standardize agent-made purchases. Salesforce formed Missionforce to bring AI into U.S. defense workflows, and Beijing’s antitrust ruling on Nvidia’s Mellanox deal added fresh strain to U.S.–China chip ties.

On the frontier, labs and capital kept surging. Biostate’s multi-agent system compressed research timelines from years to weeks; OpenAI tightened youth safety rules as scrutiny grows. Funding flowed to scale: Nvidia earmarked £2B for U.K. startups; Figure topped $1B for humanoids; Groq raised $750M for inference chips; CuspAI secured $100M for materials discovery; Micro1 landed $35M to staff data work with experts; and OranAI picked up angel backing for agent-driven marketing.

Industry & Enterprise

What Are ChatGPT’s 700 Million Users Working on? What the Data Reveal About How People Actually Use AI

ChatGPT now sees 700 million weekly users exchanging 2.5 billion messages a day, with more than 70% of activity consumer-driven and clustered around practical guidance, information-seeking, and writing. On the job, writing dominates while programming is a small slice, and usage across industries is converging on decision support and information tasks. (AI Insider)

Salesforce Launches Missionforce to Bring AI to U.S. National Security

Salesforce launched Missionforce, a new unit led by Kendall Collins to apply AI across defense personnel, logistics, and decision-making. It pairs Salesforce’s cloud platform with AI to modernize workflows for U.S. military and federal agencies—building on existing Army, Navy, and Air Force work—and arrives amid a wider tech push into defense-focused AI from OpenAI, Anthropic, and Google. (AI Insider)

Google Unveils Agent Payments Protocol to Standardize AI-Driven Transactions

Google unveiled the Agent Payments Protocol (AP2), an open standard for AI agent-initiated purchases backed by 60+ merchants and financial firms to ensure interoperability and a verifiable audit trail. The spec adds layered approvals (“intent” and “cart” mandates) while allowing tightly governed automation, is published on GitHub, and is being extended to crypto via partnerships with Coinbase, MetaMask, and the Ethereum Foundation. (AI Insider)

China Rules NVIDIA in Violation of Antitrust Law, Heightening U.S.-China Semiconductor Tensions

China’s market regulator said Nvidia violated antitrust rules in its $7B Mellanox acquisition and will keep investigating, with no penalties announced yet. The move lands amid U.S.–China tariff talks in Madrid and ongoing export-control shifts over Nvidia’s AI chips; Nvidia says it will comply and cooperate with regulators. (AI Insider)

Amazon Expands Seller Assistant with Always-On AI Agent for Third-Party Sellers

Amazon rolled out an always-on AI agent for Seller Assistant that moves from monitoring to proactively running third-party operations—flagging slow movers, tuning prices, enforcing safety compliance, analyzing demand, generating shipment plans, and avoiding storage fees. It’s also pushing agentic AI into advertising with conversational campaign creation, part of a broader strategy to embed intelligent agents across commerce so sellers can scale and focus on growth. (AI Insider)

Research & Innovation

Biostate AI Launches an AI Agent That Compresses Research Cycles from Years to Days

Biostate AI launched K-Dense Beta, a multi-agent research system built on Google Cloud’s Gemini 2.5 Pro that aims to shrink scientific cycles from years to days and curb hallucinations; in tests with Harvard Medical School, it built a transcriptomic aging clock from 600,000 profiles and flagged stage-specific gene predictors in weeks. The company says K-Dense beat frontier models on a leading bioinformatics benchmark and is now being validated with academic, biotech, and pharma partners worldwide. (AI Insider)

Policy & Governance

OpenAI Introduces New Safety Policies for Underage ChatGPT Users Amid Rising Concerns

OpenAI CEO Sam Altman unveiled new ChatGPT safeguards for users under 18, banning sexual or flirtatious interactions with minors, tightening self-harm protections, and—in severe cases—enabling notifications to parents or authorities; parents will also be able to set “blackout hours” for child accounts. The move follows mounting scrutiny, including a wrongful-death lawsuit, and coincides with a Senate Judiciary hearing on AI chatbot risks around child safety and privacy. (AI Insider)

Startups & Capital

NVIDIA Announces £2B Investment in the United Kingdom AI Startup Ecosystem

Nvidia will invest £2 billion to expand capital and computing access for U.K. AI startups and researchers across London, Oxford, Cambridge, Manchester, and new government “AI growth zones.” Backed by Accel, Air Street, Balderton, Hoxton Ventures, and Phoenix Court, the plan targets supercomputing and funding gaps, aims to spur jobs and commercialization beyond London, and reinforces transatlantic ties as money is domiciled in the U.S. but deployed in Britain. (AI Insider)

Figure Exceeds $1B in Series C Funding at $39B Post-Money Valuation

Figure secured over $1B in committed Series C funding at a $39B valuation to speed deployment of its general-purpose humanoid robots. The round—led by Parkway Venture Capital with participation from Brookfield, NVIDIA, Macquarie, Intel Capital, and others—will scale BotQ manufacturing and real-world deployments, build next-gen NVIDIA GPU infrastructure for Helix, and fund expanded human-video and multimodal data collection to boost performance. (AI Insider)

Micro1 Secures $35M Series A at $500M Valuation to Scale Human Expertise for AI Training

Micro1 raised a $35M Series A led by 01 Advisors at a $500M valuation, as the three-year-old, Ali Ansari–founded startup surged to $50M ARR in 2025 by connecting AI labs with vetted domain experts—recruited via its AI agent “Zara”—for labeling and model training, with customers including Microsoft and Fortune 100 firms. Positioning itself as a trusted alternative amid shifting data-services dynamics, Micro1 will use the funds to expand into “AI environments,” virtual workspaces for training agents on simulated tasks. (AI Insider)

CuspAI Announces $100M Series A to Accelerate AI-Powered Materials Discovery

Cambridge-based CuspAI raised a $100M Series A co-led by NEA and Temasek, with backers including Samsung Ventures, Hyundai Motor Group, NVentures, Prosus, and angels such as Durk Kingma and Zoubin Ghahramani. The company—whose AI “search engine for materials” designs candidates up to 10x faster and already partners with Hyundai, Kemira, and Meta—will use the funds to scale in the U.S. and Asia, accelerate product development, and add board members Martin van den Brink and Lord Browne, a year after its $30M seed (advised by Geoffrey Hinton and Yann LeCun). (AI Insider)

Groq Announces $750M in Funding to Reach $6.9B Valuation as Demand for AI Inference Chips Accelerates

Groq raised $750M led by Disruptive, lifting its valuation to $6.9B, with investors including BlackRock, Neuberger Berman, Deutsche Telekom Capital Partners, Samsung, Cisco, D1, Altimeter, 1789 Capital, and Infinitum. The AI chip startup—focused on inference hardware for running pre-trained models—pitches itself against Nvidia and AMD as inference takes priority, following a $640M raise in August 2023 and a separate $1.5B Saudi commitment to expand regional deliveries. (AI Insider)

OranAI Raises Multi-Million-Dollar Angel Funding to Lead AI Content Marketing Through Its AI Agent PhotoG

OranAI raised a multi-million-dollar angel round to expand its AI marketing platform and strengthen PhotoG, its agent that automates insights, content creation, and publishing at scale. Founded in 2024, the company has surpassed $1.4M in revenue with 40 enterprise clients in six months and is building a compliance-first model library and multimodal “commercial brain” (PhotoG, DataG, VoyaAI) to deliver personalized, safe campaigns across beauty, FMCG, fashion, and consumer electronics. (AI Insider)

Check out AI Insider every day for the latest in AI and Robotics developments.

Greg Bock

Greg Bock is an award-winning investigative journalist with more than 25 years of experience in print, digital, and broadcast news. His reporting has spanned crime, politics, business and technology, earning multiple Keystone Awards and a Pennsylvania Association of Broadcasters honors. Through the Associated Press and Nexstar Media Group, his coverage has reached audiences across the United States.

Share this article:

AI Insider

Discover the future of AI technology with "AI Insider" - your go-to platform for industry data, market insights, and groundbreaking AI news

Subscribe today for the latest news about the AI landscape