Today's AI Pulse
-1.png)
Anthropic shuts down Fable 5 and Mythos 5 under U.S. export controls
Anthropic abruptly disabled its flagship Fable 5 and Mythos 5 models to comply with a U.S. export‑control directive, exposing the vulnerability of firms that depend on closed‑source AI services. The move sparked a sharp rally in Chinese open‑source offerings such as MiniMax and Zhipu as investors shift to downloadable, on‑premise models.
Also developing:
By the numbers: CNTXT AI raises $60M Series A
Sovereign‑Wealth Fund Coalition Pushes Alphabet for Government AI Oversight
More than 40 investors, led by Zevin Asset Management and controlling $1.15 trillion in assets, have filed Proposal 11 to force Alphabet to disclose how its AI and cloud tools are used by governments. The group holds roughly $2.2 billion in Alphabet shares, less than 1% of the company, and cites the $1.2 billion Project Nimbus contract with Israel as a case study. The move adds pressure on Google amid growing scrutiny of AI in public‑sector applications.
Walmart Deploys ‘Code Puppy’ AI Coding Assistant to Dodge Vendor Lock‑In
Walmart has rolled out Code Puppy, an AI‑driven coding assistant that can switch among dozens of large‑language models, giving its developers freedom from single‑vendor dependence. The tool, built by distinguished engineer Mike Pfaffenberger, aims to curb token‑price volatility and keep...
Agentic AI Elevates Ops Platforms to Enterprise Core, Says DocuSign and AVEVA
DocuSign, ServiceTitan, Rubrik and AVEVA announced AI‑driven enhancements that turn operational platforms into agentic engines, while Microsoft’s internal “addicted” strategy underscores the shift. COOs must now view AI‑enabled ops layers as the primary lever for growth and efficiency.
GE HealthCare Secures FDA Clearance for AI Radiation Therapy Planning Software
GE HealthCare announced FDA 510(k) clearance for its AI‑enabled auto‑contouring software MIM Contour ProtégéAI+ 2.0. The clearance paves the way for faster radiation‑therapy planning for the roughly 2 million U.S. patients treated each year, and introduces a Predetermined Change Control Plan to...
OpenAI's "Dreaming" Gives ChatGPT Long-Term Memory
OpenAI has introduced "Dreaming," a new memory architecture for ChatGPT that provides true long‑term, cross‑session recall. The upgrade replaces the earlier Memory 1.0 system, which relied on manual context injection and static snapshots. Dreaming automatically captures user preferences, project details,...
AGIBOT WORLD CHALLENGE 2026 Advances Embodied AI Competition From Simulation to Real-Robot Testing at ICRA 2026
AGIBOT hosted the WORLD CHALLENGE 2026 at ICRA 2026 in Vienna, drawing 526 teams from 27 countries to compete in Reasoning to Action and World Model tracks. The event shifted evaluation from pure simulation to closed‑loop testing on the AGIBOT G2 humanoid, using...
Anthropic Files Confidential S‑1 as AI IPO Wave Gains Speed
Anthropic submitted a confidential S‑1 to the SEC on June 1, positioning the Claude‑maker for one of the largest AI IPOs ever. Valued at roughly $965 billion and reporting $47 billion in annualized revenue, the filing underscores the flood of capital chasing generative‑AI...
Bitfarms Rebrands as Keel Infrastructure, Shifts Focus to AI Power Assets
Bitfarms (KEEL) announced it has rebranded as Keel Infrastructure and is scaling back Bitcoin mining to chase AI‑related power and data‑center opportunities. The move leverages its North‑American electricity footprint, a strategic asset as AI workloads drive soaring energy demand.
Collate Raises $95M to Transform Life Sciences AI
Collate announced a $95 million Series B round that pushes its valuation close to $1 billion, marking it as one of the fastest‑growing regulated‑SaaS firms. The San Francisco startup targets life‑science companies burdened by massive regulatory paperwork, using specialized large‑language models...
Your AI Bill Is Out of Control. Cloudflare Can Fix It Now.
Cloudflare announced new spend‑control features for its AI Gateway, letting enterprises set dollar‑based budgets and enforce limits across models, providers, teams, or individual users. A closed beta adds identity‑driven budgets that tie usage to employees or service tokens via Cloudflare...
Article Intro - MICCAI Endoscopic Imaging Challenge
The MICCAI conference has launched the Open Suturing Skills Vision‑Based Assessment Challenge for 2024‑2025, the first large‑scale competition to evaluate open‑surgery proficiency using computer‑vision techniques. Organizers released a curated dataset of high‑resolution video recordings annotated with surgical outcomes, OSATS scores,...
Anthropic Suggests Slowing AI Research Until We Can Align It with Human Goals
Anthropic researchers Marina Favaro and Jack Clark warned that fully recursive self‑improving AI could outpace current alignment safeguards, urging society to consider slowing development. They outlined three futures, with the worst‑case scenario involving AI systems that build and improve themselves...
Ramp Debuts Accounting Product, Raises Additional $750M
Ramp announced the launch of "Ramp Stack," an AI‑driven operating system designed to automate core accounting functions such as month‑end close, cash reconciliation, and journal entry posting. The same day the company disclosed a $750 million equity raise that lifts its...
Microsoft Trained Its MAI Models on Unlicensed Web Data Despite Promising "Enterprise Grade, Clean and Commercially Licensed Data"
Microsoft’s latest MAI models were partially trained on data harvested from the open web, including the Common Crawl corpus, despite earlier statements that training would rely solely on "enterprise‑grade, clean and commercially licensed" data. The technical paper released by Microsoft...
Ardian, Artefact, Bull, EDF, Capgemini, the Iliad Group, Orange and Scaleway Combine Their Expertise to Run as a Candidate for...
A coalition of eight French leaders—Ardian, Artefact, Bull, Capgemini, EDF, the iliad Group, Orange and Scaleway—has formed the AION consortium to submit a bid for the European Union’s AI Gigafactory program in France. The partnership aims to deliver sovereign, high‑performance...
Design GenAI From Scratch: Beginner-Friendly Masterclass
We wrote a full breakdown of how GenAI systems are actually designed from scratch. It went out in Neo Kim's "The System Design Newsletter": Generative AI Masterclass: Everything You Need to Know to Design GenAI Systems From Scratch. If you're a beginner in...
Huawei's Ascend 910C Powers DeepSeek-V4-Pro Breakthrough
Huawei chips refine DeepSeek model in leap for China’s AI self-reliance A research team that includes Huawei Technologies says it has successfully used the firm’s Ascend 910C chips to complete post-training for the DeepSeek-V4-Pro model. https://t.co/tnlLSzzs7V
Bulgaria’s Information Services and Google Cloud Launch AI-Powered National Cybershield
Bulgaria’s national integrator, Information Services, has teamed with Google Cloud to deploy the AI‑powered Cybershield platform across its government. The project creates a federated Security Operations Center that will connect 54 ministries and agencies, leveraging Google’s cloud security suite and...
Canada Launches Multi‑hundred‑million AI Growth Fund
Canada’s government will set up a technology growth fund to help fund promising artificial-intelligence companies, part of the hundreds of millions of dollars it plans to pump into building a homegrown AI industry and encourage adoption of the technology https://t.co/KibBdvA3DW...
DELEGATE-52 Shows That LLMs Corrupt Your Documents Over Time: Artificial Intelligence Trends
Microsoft Research’s DELEGATE‑52 benchmark evaluates how large language models (LLMs) perform in long‑running, delegated document‑editing tasks across 52 professional domains. Testing 19 models—including Gemini 3.1 Pro, Claude 4.6 Opus and the newly released GPT 5.4—researchers measured reconstruction scores after multiple forward‑and‑inverse edits. After 20 interactions,...
Veeam Says Enterprises Need AI Agents to Monitor AI Agents
Veeam announced three AI‑driven compliance agents—Consent, Data Subject Request and Assessment—designed to monitor and enforce data‑privacy policies for autonomous AI agents. The company’s Data and AI Trust Gap report shows 88% of senior executives are already using AI agents, yet...
1X Launches World Model Lab to Advance Humanoid Robot Autonomy
1X has launched the 1X World Model Lab to build large‑scale AI world models that will power its Neo humanoid robot. The lab will pretrain foundation models using a mix of internet data, first‑person human video, simulation, teleoperated robot streams...
AI Server Earnings: Wall Street Sees One Clear Standout
Wall Street highlighted Dell Technologies as the clear AI‑server leader, reporting an 88% revenue surge and a 700% jump in AI‑server sales. Hewlett Packard Enterprise (HPE) beat estimates, raised its EPS guidance two years ahead of schedule, and now carries a near‑$65...
Infosecurity Europe: Practical Lessons From Lloyds' Agentic AI Security Playbook
Lloyds Banking Group is approaching agentic AI as an engineering problem, embedding security controls across the AI lifecycle. The bank created an internal agent marketplace, multi‑cloud identity management, and automated red‑team testing to govern more than 50 GenAI use cases....
Only 10% of SOCs Say They’re Getting Excellent Value From AI. Here’s What the Second Wave Has to Deliver
The SOC‑CMM 2026 report reveals a surge in AI adoption across security operations, with off‑the‑shelf LLMs up 55 % YoY, AI co‑pilots up 145 %, and AI agents up 118 %. Despite this, only 10 % of SOCs report excellent AI value, while 71 %...
Anthropic's Mythos Model Is Reportedly Powering NSA Offensive Cyber Ops Against China and Iran
Anthropic's Mythos AI model is being used by the U.S. National Security Agency for offensive cyber operations targeting networks in China and Iran. The company has embedded roughly six engineers within the NSA to tailor the model for these missions....
What Past Technology Waves Teach Us About AI Adoption
The article argues that today’s surge in agentic AI mirrors earlier technology waves, where hype, early adopters, and gradual value realization followed a predictable arc. It uses Salesforce’s Einstein Copilot rollout as a case study, highlighting early successes, data‑quality gaps,...
The 20 AI Agent Platform & Framework CEOs You Need to Know in 2026
The AI Insider cataloged the most influential CEOs steering the rapid expansion of AI agent platforms in 2026, from enterprise‑focused sovereign solutions at Cohere to open‑source frameworks like LangChain and large‑scale cloud bets by Microsoft. These leaders are delivering massive...
Skylytics Launches VEXΛ and VΛST: AI Sovereignty for the Business
Skylytics Data unveiled two new AI solutions: VEXΛ, a private generative‑AI accelerator that runs inside a customer’s Azure environment, and VΛST, an automated platform that validates, scores and tests AI outputs for accuracy and security. Both products are designed to...
Investors Have Poured Billions Into Plaintiff-Side Legal AI, But Defense Could Be The Next Big Opportunity
Investors have funneled roughly $682 million into plaintiff‑side legal AI firms such as EvenUp, Eve, Supio and Darrow, accounting for about 71% of disclosed capital in the sector. These startups benefit from standardized intake, case evaluation and demand‑generation workflows that lend...
Britain Set to Award AI Military Decision-Support Contract
The UK Ministry of Defence will award a direct contract to Defence Holdings PLC to prototype an artificial‑intelligence decision‑support system called Project Strong. The £227,000 contract (about $288,000) covers three months of work and is exempt from open competition under...
Music AI’s $5.4B Valuation Driven by Data, Not Tricks
A music AI model is now valued at $5.4 billion. More than most century-old record labels. Investors aren't paying for the audio trick. They're paying for the licensed training data. The model is the commodity now. The corpus is the company. https://t.co/yCzTci7k2i
The Friday File: Anthropic; EU; FCC
Anthropic announced it will extend access to its Claude Mythos AI model to 150 additional companies across power, healthcare and communications, still under the guarded Project Glasswing framework. The European Commission rolled out a new digital‑sovereignty package that refreshes the...
Google’s NotebookLM Serves Up Jay Shetty‑Style Coaching, Highlighting ChatGPT’s Gaps
Google’s NotebookLM was tasked with delivering a Jay Shetty‑style personal‑coach experience. The AI notebook produced daily affirmations and reflective prompts that the author found more nuanced than ChatGPT’s standard advice, underscoring gaps in mainstream conversational AI’s motivation tools.
How Breast Cancer Screening Can Predict Heart Disease Risk
Researchers have created an artificial‑intelligence model that automatically scans routine mammograms for breast arterial calcifications (BAC) and quantifies their severity. The study, covering over 120,000 women at Emory and Mayo Clinic sites, found that even modest BAC levels raise cardiovascular...
Former Salesforce SVP Launches AI Governance Startup, Credits Generative AI
Gabrielle Tao, a former senior vice president at Salesforce, left the cloud‑CRM giant in September to found Tovix AI, an AI‑governance startup. She says generative AI tools like Claude accelerated product development and lowered barriers for small teams, making a...
SuperX Unveils 1.6 Tbps Optical Modules at Interop Tokyo, Targeting AI Data Center Supply Chain
SuperX AI Technology Limited announced the launch of its 1.6 Tbps optical module solution at Interop Tokyo 2026, alongside a full‑stack AI compute and power portfolio. The rollout aims to accelerate AI data‑center deployments in Japan and globally, tightening the telecom‑compute...
UK MP and New Claimants Sue Elon Musk's xAI Over Grok-Generated Sexualised Images
Labour MP Jess Asato and a growing group of claimants have filed a high‑court lawsuit against Elon Musk’s xAI, accusing the company’s Grok AI tool of producing non‑consensual sexualised images and video. The case, described as the first test of...
Anthropic Unveils Open‑Source AI Framework for Autonomous Vulnerability Discovery
Anthropic has published an open‑source reference implementation that uses its Claude model to automate the full vulnerability discovery and remediation loop. The framework, hosted on GitHub, is positioned as a building block for DevSecOps pipelines, while a managed Claude Security...
Texas DIR Announces New Agency CIO to Oversee IT and AI Programs
The Texas Department of Information Resources (DIR) is hiring a dedicated agency chief information officer to direct its IT and artificial‑intelligence initiatives. The role, reporting to state CIO and DIR executive director Tony Sauerhoff, offers a monthly salary of $17,417‑$17,917...
SharonAI Holdings Inc. (NASDAQ: SHAZ)
SharonAI Holdings (NASDAQ: SHAZ), valued at roughly $1.3 billion, provides purpose‑built cloud infrastructure for AI workloads, leveraging a unified Vast Data AI Operating System. The company recently closed a $350 million senior convertible‑note offering and has upgraded its data‑center capacity target to...
Asana Deploys AI “Chief of Staff” To Convert Slack Chaos Into Structured Work
Asana introduced Dash, an AI “chief of staff” that pulls follow‑ups from Slack, email and calendars into its Work Graph, promising to turn unstructured communication into trackable work. The launch, announced at the Work Innovation Summit in London, is part...
Before You Buy the Model: A Healthcare AI Readiness Framework for IT Leaders
Healthcare AI demos often look impressive, but many pilots stall when deployed at scale because the underlying infrastructure is weak. The article argues that the real bottleneck is not the model but fragmented data, poor integration, and lack of governance....
GrowthIQ and Winnesota Show Operations-First AI Payback with NAVI Platform
GrowthIQ’s NAVI platform is delivering an operations‑first AI solution for Winnesota, a regional cold‑chain logistics provider. By connecting live trucking data, custom rules, and workflow memory, NAVI automates daily customer updates, proof‑of‑delivery reports, and document generation. The pilot produced 2,729...
AI Boom Triggers $1.1 Trillion Power Push and $5 B High‑Density Rack Market Surge
Exelon and other utilities are committing roughly $1.1 trillion over the next five years to power the AI‑driven data‑center boom, while the high‑density AI rack market is projected to grow from $1.6 billion in 2026 to $5 billion by 2036. The twin pressures...
Ginkgo Bioworks Uses AI Robots for 30,000 Experiments, Cutting Costs 40%
Ginkgo Bioworks, the MIT‑spun biotech startup, deployed AI‑controlled laboratory robots that ran more than 30,000 experiments in six months, delivering a 40% reduction in protein‑synthesis costs. The breakthrough showcases how generative AI can move from drafting protocols to directing autonomous...
How ‘Talking to Your Data’ Is Reshaping Financial Decision Making
AI-driven conversational analytics are moving from a novelty to a core capability in private equity. Investors can now ask natural-language questions of their data, receiving instant analysis, visualizations, and commentary that adapt as hypotheses evolve. This real-time access shortens due-diligence,...
Three‑quarters of K‑12 Teachers Say AI Will Outpace Internet’s Impact on Education
An NPR/Ipsos poll of 545 K‑12 teachers reveals that nearly three‑quarters think artificial intelligence will have a bigger effect on schooling than the internet or computers. The same survey shows most educators are already using AI for tasks, but they...
AI Coding Agents Are Creating A New Management Challenge For Business Leaders
Anthropic warns that AI coding agents are shifting from assistive tools to collaborative partners that can write tests, debug, and generate documentation, but humans still retain most responsibility. Internal data shows developers use AI for about 60% of their tasks...
Fintechs Redesigning with AI Achieve 5× Developer Productivity
AI is delivering 5x developer productivity at fintechs that fully redesign their engineering model around it. Not just copilots — end-to-end workflow transformation. The firms still running scattered pilots are being outbuilt. BCG + FT Partners Global Fintech Report 2026 https://t.co/zPh99Mx5T0