Today's AI Pulse
Odyssey lands $310M Series B, valued at $1.45B
Odyssey announced a $310 million Series B round that values the AI lab at $1.45 billion. The round was led by Natural Capital and included backers such as Amazon, AMD Ventures, GV, EQT and IQT. In parallel, Odyssey struck a preferred‑cloud deal with AWS to use Trainium silicon for large‑scale model training.
Also developing:
Merck and Mastercard Are Seeing Real Agentic AI Results. Both Say the Plumbing Came First.
Merck is leveraging AI agents to accelerate drug discovery and marketing, cutting research cycles by a third and delivering compliant marketing drafts up to 80% faster. The gains stem from a "plumbing‑first" strategy that now supports 2,500 AWS accounts, multiple Azure and GCP subscriptions, and dozens of edge locations. AI‑generated drafts are 99% compliance‑ready, shrinking review times from months to days. Mastercard is applying similar agentic AI to streamline transaction dispute workflows, seeking cost reductions while maintaining consumer trust.
Seeds | FIVEAGES Closes Hundreds of Millions of Yuan in A-Round Financing
FIVEAGES, a Chinese robotics startup founded in September 2024, closed an A‑round financing of roughly ¥200 million (about $28 million), its third funding round in 2026. The round was led by Futeng Capital and included several domestic venture firms and an increased...
Volkswagen's Full-Domain AI Agents Expected to Launch in Second Half of Year
Volkswagen announced that its Agentic AI for All roadmap will see full deployment in new models built on the CEA architecture starting in the second half of 2026. The system, powered by a locally trained large language model, moves beyond...
Tensormesh Raises $20M, Launches AI Inference Platform Built on KV Caching
Tensormesh announced a $20 million Series A extension, bringing its total capital to $24.5 million, and launched Tensormesh Inference, a SaaS platform that leverages key‑value (KV) caching to eliminate redundant GPU computation. The technology promises up to a tenfold reduction in latency and...
Supermicro and Verda Deliver Sustainable, Full-Stack AI Cloud Infrastructure for Next-Gen AI Workloads
Supermicro announced that European AI cloud provider Verda has chosen its NVIDIA Blackwell‑based, rack‑scale systems to build a full‑stack AI cloud across Europe, the U.S., and Asia. The deployment includes NVIDIA GB300 NVL72, HGX B300, HGX B200, and RTX PRO 6000 servers,...
VC: The $575B AI Bet (Biggest Since the World Wars) and the New Two-Buyer Reality in Software | Tomasz Tunguz,...
In this GTM Now episode, Tomas Tunguz of Theory Ventures explains that AI infrastructure spending is on track to become the fifth‑largest U.S. infrastructure project ever—about $575 billion by 2030—as hyperscalers pour roughly $12 into data‑center build‑out for every $1 earned from...
Exploits Arrive Before Patches: AI Defense Becomes Essential
The most important number in cybersecurity right now is negative seven. That's what Mandiant (part of Google Cloud)'s M-Trends 2026 report found: mean time to exploit is now an estimated -7 days. Exploitation is routinely happening before a patch is even...
Spielberg Tells Michelle Obama, “Do Not Use AI as the Final Word on Anything Creative”
Steven Spielberg appeared on Michelle Obama’s IMO podcast to discuss his upcoming film *Disclosure Day* and his perspective on artificial intelligence in filmmaking. He emphasized that AI can serve as a practical tool for tasks like location scouting, but should...
Robinhood Allows Users to Use AI Agents to Trade Stocks
Robinhood announced AI‑driven agents that can execute stock trades on behalf of retail customers. The service currently supports equities, with plans to add options, futures and cryptocurrency later, and also includes agents for credit‑card purchases. Users set spending limits and...
SecureFlag Launches AI-Assisted Development Labs to Tackle Security Risks in AI-Generated Code
SecureFlag, a developer security enablement platform, has introduced AI‑Assisted Development Labs to train developers on safely using AI coding assistants such as GitHub Copilot, Claude and ChatGPT. The hands‑on program teaches prompting techniques, code review, and integration with security tools...
“There Is No Accountability”: AI Coding Agents Are Installing Packages No One Owns
AI coding assistants such as GitHub Copilot, Claude Code and Cursor are increasingly installing packages and dependencies without clear ownership, creating a security accountability gap across enterprises. Aikido Security’s new Endpoint product monitors and blocks unknown installations for a configurable...
“Not a Blip”: AI Was Canada’s Venture Market Mover in 2025
Canada’s venture capital market saw AI become the dominant sector in 2025, accounting for 54 cents of every dollar invested and 23 percent of all financing rounds. Osler’s Deal Points Report, based on 140 deals worth $6.3 billion CAD (≈$4.66 billion USD), shows AI deal sizes grew...
AI Tool Helps Remote Monitoring Clinics Explore Patient Data
Octagos introduced Ask Atlas, an AI-powered natural‑language interface for its electrophysiology remote‑monitoring platform. The tool lets clinicians pose simple questions and retrieve answers from device transmissions, electronic health records, workflow status, and billing data within seconds. By replacing manual chart...
SAP Sapphire Madrid Customer Keynote: Enterprise AI Value Depends on Business Ownership
SAP’s Sapphire Madrid keynote shifted the AI conversation from pilot announcements to operational adoption, emphasizing that enterprise AI value hinges on governance, observability, and clear business ownership of agents. Executives highlighted "verified agents" built under ISO‑certified controls, clean‑core SAP S/4HANA...
The Impact of AI on Traditional Development Processes
Rob Zuber explains how AI is reshaping the software development life cycle by introducing autonomous reliability—continuous quality feedback that guides AI‑generated code. While AI speeds prototyping, moving to production reveals gaps in security, scaling and trust that require new platform‑engineer...
Paychex Launches WISE AI Platform
Paychex announced the Workforce Intelligence Strengthened by Expertise (WISE), an AI‑powered platform that deploys autonomous agents to execute HR tasks within customer‑defined workflows. The solution moves beyond static chatbots, offering agents that can schedule shifts, approve timesheets, and provide real‑time...
AI Robot Can Spot ‘Invisible’ Signs of Plant Disease
RoboCrops, a robotic phenotyping platform developed by the University of Lincoln and the Lincoln Institute for Agri‑food Technology, won a Silver Gilt medal at the RHS Chelsea Flower Show. Its core system, PhenAIx, fuses AI, high‑resolution imaging and robotics to...
Brightcove Adds New Features to Its AI Suite for Video Advertising
Streaming‑tech provider Brightcove unveiled two AI‑driven features—AI Contextual Ads and Smart Ad Breaks—within its AI Suite for video advertising. AI Contextual Ads scans video scenes against the IAB taxonomy, delivering scene‑level contextual signals to ad servers, while Smart Ad Breaks...
Google Moves AI Agents Into the Mainstream
At its I/O conference Google unveiled Gemini‑powered AI agents that move beyond chat, embedding themselves in Search, Gmail, Docs, Chrome, YouTube, shopping and even smart glasses. The new Gemini models are designed to plan tasks, operate across apps, interpret images...
GCHQ Teases “Blueprint” For Agentic AI National Cyber Defense
British signals agency GCHQ announced it has drafted a "blueprint" to embed cutting‑edge agentic AI into a new national cyber‑defence capability. Director Anne Butler‑Keast highlighted the urgency for businesses to bolster security as AI threats accelerate. The initiative ties into...
FAA Turns to AI to Spot Aviation Risks Faster, Aiming to Cut Incident Lag
The Federal Aviation Administration unveiled a program to deploy artificial‑intelligence tools that can sift massive flight‑data streams and flag safety threats in near‑real time. Officials say the move addresses long‑standing criticism that the agency lags behind in predictive analysis, while...
Hyundai Motor Group Fast‑tracks Atlas Humanoid Robot Production in Georgia
Hyundai Motor Group announced an accelerated rollout of mass‑produced Atlas humanoid robots at its Metaplant America facility in Georgia, backed by a new Software Defined Factory (SDF) unit. The move pairs robot manufacturing with a dedicated robotics parts procurement office...
ING Uses AI ‘Vibe Coding’ to Cut Trading Tool Development to Hours
ING Groep has rolled out an in‑house generative‑AI system called “vibe coding” on its electronic trading desk, compressing code‑prototype timelines from days or weeks to a matter of hours. The bank says the approach, which turns natural‑language prompts into production‑ready...
Dataiku and Snowflake Unveil Cobuild for Governed AI Workflows
Dataiku announced Cobuild on Snowflake, a joint solution that turns natural‑language intent into visual, governed AI workflows using Snowflake Cortex AI. The offering targets Global 2000 enterprises seeking transparent, production‑ready AI pipelines.
Scaling Obsession Delayed AGI; Neurosymbolic Focus Needed
If we had done everything I suggested in my 2020 arXiv article “The Next Decade in AI”, we might actually have reached AGI by now. In the last three years, after a detour driven by the false promise of pure scaling,...
Salesforce’s AI‑First Bet Faces Crucial Test in Q1 FY2027 Earnings
Salesforce (CRM) will release its fiscal first‑quarter 2027 earnings on May 27, putting its Agentforce AI platform under the microscope. Analysts expect $11.05 B in revenue and $3.12 EPS, while the company’s guidance targets $45.8‑$46.2 B full‑year revenue and a 34.3% non‑GAAP...
AI‑Driven Autonomous Robot Labs Unveiled to Speed Scientific Experiments
Researchers have introduced AI‑driven autonomous robot laboratories that can conduct routine experiments without human hands, promising to free scientists for higher‑level inquiry. The development signals a convergence of robotics and artificial intelligence that could reshape how research is performed.
Continual Learning Agents Improve Like Students Over Time
New trend: continual learning. I first showed you this with Levangie Labs from @blevlabs. You "grow" an agent by teaching it. It gets better over time. It works because it has a better memory, so can try things, and learn, and remembers...
Dean Gonsowski: Your AI Doesn’t Have a Hallucination Problem — It Has a Data Problem
Dean Gonsowski argues that AI hallucinations stem from poor data rather than model flaws. He cites early 2023 incidents like the Mata v. Avianca brief where ChatGPT fabricated citations, and notes that modern LLMs have reduced hallucination rates. Yet more...
Introducing Planetary Intelligence: AI Gains Real‑World Vision
I’m introducing a new type of machine intelligence that understands our physical world in real-time… a powerful expansion of AI’s capabilities. The AI models today that have captivated the world are very powerful – wrapping most of human knowledge as committed...
Inside Optum Health's Push to Make AI Practical for Clinicians
Optum Health piloted an AI‑driven chart summarization tool to cut clinicians' after‑hours record‑review time. The feature automatically condenses patient histories into a concise view, letting providers prepare faster before appointments. Early results showed smoother visit flow, reduced documentation workload, and...
CelerData Rebrands as PhoenixAI, Introduces Analytical Engine Designed for AI Agents
PhoenixAI, the rebranded CelerData, unveiled its Agentic AI Database, an analytical engine built for AI agents to query live enterprise data with sub‑second latency and massive concurrency. The platform unifies real‑time and historical data, eliminating brittle pipelines while embedding SOC 2‑grade...
AI-Driven Science Is Turning Data Storage Into a Competitive Advantage
AI is reshaping scientific storage from archival capacity to real‑time accessibility. Researchers now need continuous, high‑throughput access to massive datasets for training foundation models, turning old genomics, climate and imaging archives into active assets. Storage vendors such as Western Digital,...
RemoFirst Goes All-In on AI, Becoming the First Major AI-Native Employer of Record and Launches RemoAI
RemoFirst has completed a year‑long AI‑centric rebuild and unveiled RemoAI, a conversational AI platform that embeds intelligence across its global Employer of Record services. The new suite includes a Global Recruiter Agent and a Task Manager Agent, offering instant compliance...
OpenAI Says No Jobs Apocalypse
OpenAI’s CEO Sam Altman told Reuters that AI is unlikely to spark a global jobs apocalypse, emphasizing productivity gains and the enduring value of human interaction. In contrast, Anthropic’s leadership has warned that AI could cause significant workforce disruption, especially...
Why Publishers Are Betting on Video to Counter AI Search
Publishers are losing an estimated $2 billion in ad revenue as AI‑driven zero‑click search climbs from 56% to 69% in a single year. To counter this, they are shifting focus to video, a format that resists AI summarization and drives higher...
AI Underwriter FutureProof Enters E&S Lines Market
FutureProof Technologies, a venture‑backed InsurTech, has launched an AI‑driven personal lines program for condo and renters insurance in catastrophe‑exposed southeastern states, focusing on excess and surplus (E&S) coverage in Florida and Texas. The initiative partners with wholesale broker Bridge Specialty...
AI‑crafted Spam Emails Masquerade as Human Outreach
My email inbox is filling up with AI-generated emails: an example. I get an email from seemingly a human, asking for a link insertion to promote a company that is their client. Except this is not from a human. It's with AI/automated: the...
What the Streaming Wars Can Teach Utilities About the AI Data Center Boom
Utilities are confronting a surge in AI‑driven data‑center load that could triple electricity demand by 2030, echoing the disruption of the streaming wars in media. The article warns that treating each interconnection as a standalone project risks cost shifts to...
NATO Chooses France for AI
NATO is set to establish a Centre of Excellence for artificial intelligence in Rennes, France, with formal member approval expected in early June. The new hub will employ roughly 50 specialists and sit alongside the city’s existing cyber‑defence ecosystem. France...
Getting Value Out of Enterprise AI Consistently Hits a Data Road Block. Box CEO Aaron Levie Wants to Have a...
Box CEO Aaron Levie highlighted that enterprises are eager to harness AI agents for faster product execution, better customer service, and operational efficiency, but they hit a roadblock when trying to connect agents to fragmented, unstructured data stored in legacy...
Potential DLSS 5 References Have Been Found in the Latest Nvidia Driver, but It's All Gone Quiet on the AI...
Nvidia’s latest GeForce Game Ready driver (610.47) introduces three new profile options—Enable DLSS NR override, Enable DLSS NR SL override, and Override DLSS NR presets—suggesting ongoing work on its next‑gen DLSS 5 neural‑rendering technology. The earlier DLSS 5 reveal sparked fierce debate...
RMX Industries Sees Accelerating Engagement for AI-Driven Visual Intelligence
RMX Industries (OTCQB: RMXI) reports accelerating engagement for its AI‑driven visual intelligence solutions, noting a customer shift from bandwidth concerns to edge AI judgment and autonomy. The company is preparing a new class of edge‑deployed visual systems that interpret video...
AI May Increase Nursing Costs, Workflow Burden: Penn Study
A University of Pennsylvania study warns that artificial‑intelligence tools, while promising reduced paperwork and better patient monitoring, can impose hidden costs on nursing staff. Researchers highlight that AI systems often require extensive training, maintenance, workflow redesign, and continuous oversight, which...
EltegraAI Launches Enterprise AI Platform That Cuts Legacy Modernization Time From 18 Months to 3.5
Eltegra Inc. unveiled the EltegraAI Enterprise AI Platform, promising to shrink legacy system modernization timelines from an average of 18 months to just 3.5 months. In a pilot involving a 2.5‑million‑line PowerBuilder application, the platform cut delivery time by 15...
Employee-Built AI Tool Saves $120K Annually
An accountant in Vancouver just saved us $120K a year. Prescott Thompson got tired of a close-management vendor. So he built his own using our internal AI platform Mesh. Mesh is our internal AI productivity platform / AI coworker that gives...
TorqueAGI Announces Collaborations with NVIDIA, John Deere, and Dexterity to Advance Physical AI for Enterprise-Grade Robots
TorqueAGI announced strategic collaborations with NVIDIA, John Deere, and logistics firm Dexterity to accelerate deployment of Physical AI in enterprise‑grade robots. As an NVIDIA Perception Partner, TorqueAGI will leverage GPU‑powered foundation models for real‑time perception and edge reasoning. John Deere selected the...
AI Scaffolding Boils Down to Memory, Skills, Scripts
Thinking about simplifying what makes a functional AI scaffolding and how to think about building it up. Seems like you can reduce most approaches into three components: 1. A memory system: Contains relevant context about the domain the AI is operating on....
AWS CEO Calls AI Junior‑dev Replacement Idea Foolish
"We think that with AI we can replace all of our Jr developers in our company" AWS CEO Matt Garman: "That's the dumbest thing I've ever heard" https://t.co/lotYn0OLQr
Cost per Token Driven by Hardware, Not Model
Most enterprises are optimizing the wrong layer of AI. Cost per token doesn't start with the model. It starts underneath it. Huawei's Tau Scaling Law (Her's Law) changes the conversation entirely. Here's the breakdown… #HuaweiPartner @huawei https://t.co/P5fnE5CtGu