AI News and Blog Articles

Optimize blueprint extraction accuracy in Amazon Bedrock Data Automation

Amazon has added a feature to Bedrock Data Automation that automatically sharpens how its document extraction tool reads business paperwork. Called blueprint instruction optimization, it takes three to ten sample documents plus their correct answers and rewrites the extraction instructions to lift accuracy in minutes rather than weeks. No separate model training or fine-tuning is required, and the work runs through either the Amazon Bedrock console or an API.

European tech funding rebounds to €10.5B in May as mega-rounds power market recovery

European tech companies raised €10.5 billion across 258 deals in May 2026, more than double the €5.1 billion invested in April even though the number of deals fell. Large infrastructure and AI rounds drove the recovery, led by Pure Data Centres with more than €2.3 billion and Nscale with €675 million. UK companies accounted for €7.9 billion of the monthly total, and the cloud sector took the single largest share of capital.

DoorDash's new AI chatbot lets you order with prompts and photos

DoorDash launched Ask DoorDash, an AI chatbot that turns plain-language requests and photos into orders for food, groceries, and restaurant reservations. Instead of scrolling through menus and stores, shoppers tap an Ask button in the search bar and describe what they want, and the app surfaces matching options or fills a cart. The tool is rolling out first on iOS in select regions, with wider availability across the United States in the coming weeks.

Blue Yonder pushes supply chain AI toward autonomous operations

Blue Yonder, a supply chain management company, is moving its operations toward autonomous AI by deploying AI agents that work alongside humans. In an interview on theCUBE, Chris Burchett, the company's senior vice president of generative AI, explained how Blue Yonder built warehouse, logistics and inventory agents on a single common data model and announced a partnership with Nvidia at its ICON 2026 event. The company is adopting an 'owned intelligence' strategy, training smaller proprietary models to cut token costs while keeping human oversight for governance.

AWS launches FinOps agent to bring AI cost governance to cloud spend

AWS launched a FinOps agent in a feature preview at FinOps X 2026 to bring AI cost governance and real-time anomaly detection to cloud spending. The autonomous agent monitors cloud costs, detects anomalies, performs root-cause analysis, and routes alerts to responsible teams via Slack or Jira without waiting for end-of-month reports. AWS cost management director Jerry Rapisarda explained that AI costs are non-deterministic and require tying spend to business outcomes through unit economics.

Pegasystems builds on its FinOps foundation to navigate the unpredictable economics of AI

As AI spending surges, organizations are moving beyond simple cost optimization toward measuring the value AI delivers, according to Pegasystems Inc. and its director of cloud FinOps, Hunter Harris. Harris told theCUBE at FinOps X 2026 that AI usage and costs are unpredictable and can scale far beyond traditional cloud patterns, forcing companies to rethink how they track spend. Pegasystems built on its FinOps foundation by connecting AI investment to revenue at the contract level and bringing operational, support, product, and revenue data into one model.

AI value creation meets cost accountability as FinOps evolves beyond cloud

This SiliconANGLE report covers how FinOps, the practice of financial accountability for technology spending, is expanding beyond cloud cost management to address AI costs. At FinOps X 2026, SoftwareOne leaders Trent Allgood and Parker Nancollas explained that organizations are now trying to balance AI value creation with cost control. A key challenge is that AI models can enter and leave production in as little as 60 days, breaking the budgeting and forecasting habits FinOps teams built for slower-changing cloud and on-premises systems.

Diagrid brings cryptographic proof to AI agent and workflow execution

Diagrid Inc. today released Dapr 1.18, an update to the open-source runtime that lets organizations cryptographically prove how an artificial intelligence agent or workflow executed, who held custody of the work, and whether its history was altered. The release centers on a capability the company calls verifiable execution, built from three new features. Workflow History [...] The post Diagrid brings cryptographic proof to AI agent and workflow execution appeared first on SiliconANGLE.

Save Big and Play Bigger: GeForce NOW Summer Sale Brings Major Membership Savings

NVIDIA confirmed that its top GeForce NOW cloud gaming tier now runs on RTX 5080-class servers worldwide, streaming demanding PC games at up to 4K resolution and up to 120 frames per second to ordinary devices. The company also added eight new games this week and scheduled more for the weeks ahead. The update shows how cloud streaming moves heavy graphics processing off the user's device and onto remote data centers.

TestSprite launches an open-source command-line tool to help AI agents check their own work

Autonomous artificial intelligence-powered software testing tool TestSprite Inc. today announced that the company has open-sourced its command-line interface tool that allows AI coding agents to verify their own work. As the AI coding revolution has rolled in, autonomous coding tools have become smarter and enabled developers to prompt their way to entire applications overnight. The [...] The post TestSprite launches an open-source command-line tool to help AI agents check their own work appeared first on SiliconANGLE.

How a Google DeepMind Spin-off Hunts Hidden Drug Targets

Isomorphic Labs, the Google DeepMind spinout behind AlphaFold, has built an AI system called the Isomorphic Drug Design Engine (IsoDDE) that hunts for hidden binding sites on proteins where new medicines might attach. The company raised $2.1 billion in funding, one of the largest biotech rounds ever, and signed drug-discovery deals with Novartis and Eli Lilly. Its engine predicts not only where a drug molecule binds to a protein but how tightly it binds, including pockets that stay invisible until the right molecule arrives. The work shows AI moving past structure prediction toward the harder task of designing real drugs.

Anthropic apologizes for invisible Claude Fable guardrails

Anthropic apologized for a hidden guardrail in its new Claude Fable 5 model that silently degraded answers when the system suspected a user was trying to copy the model, without telling that user anything had changed. The company said it made the wrong tradeoff and will now make the restriction visible, with flagged requests falling back to its older Claude Opus 4.8 model so people see when limits kick in. Fable 5 is the first publicly available model in Anthropic's Mythos class, a tier the company had warned was too dangerous to release without strong safeguards.

Google DeepMind is worried about what happens when millions of agents start to interact

Google DeepMind, Schmidt Sciences, ARIA, the Cooperative AI Foundation, and Google.org have opened a $10 million funding call to study what happens when millions of AI agents start interacting with each other online. The worry, voiced by DeepMind's AGI safety lead Rohin Shah, is that mass-market agents now act without human oversight and follow instructions from other agents, creating risks like scaled fraud, prompt-injection attacks, and coordinated cyberattacks. The backers say a dedicated field of multi-agent safety research does not yet exist, and they want to build it before these systems become widespread.

The $100M+ Round Is Now Just Your Typical Late-Stage Financing

The median U.S. late-stage startup funding round reached exactly $100 million in 2026, according to new Crunchbase data. That figure is roughly double the just-over-$50-million median from 2020, meaning a nine-figure raise no longer marks a standout deal. So far in 2026, U.S. companies have closed 250 rounds of $100 million or more, with half of those at $200 million or above.

Publicis Sapient launches Sustain to transform IT operations with AI-enabled support

Publicis Corp., a global enterprise AI platform and services company, today announced the launch of Sapient Sustain to improve the reliability of information technology operations and managed services using agentic artificial intelligence. As more enterprise companies modernize by scaling AI, IT teams and environments find themselves in a turf war against legacy systems that are separated by [...] The post Publicis Sapient launches Sustain to transform IT operations with AI-enabled support appeared first on SiliconANGLE.

Enera raises $2M to improve the EV charging experience

Enera, an AI-powered driver experience platform for electric vehicle charging, has raised $2 million in pre-seed funding led by Lakehouse Ventures. The London and Barcelona based startup builds software that helps charging operators find out why charging sessions fail and fix the problem in real time. The round marks Lakehouse Ventures' first investment outside the United States, with Divergent Capital, Masia, and several angel investors also taking part.

OurMind lands €2.1M to help reduce healthcare workloads

OurMind, an Amsterdam-based healthtech startup, raised 2.1 million euros to expand an AI platform built to cut the paperwork load on doctors. The round was led by 4impact capital, the venture fund's first HealthTech investment, with backing from practicing general practitioners and medical specialists. The company's tools are already in use at more than 300 general practices and 14 hospitals across the Netherlands.

Cloud: 10 companies that raised the most in 2025

The 10 European cloud companies that raised the most in 2025 were led by AI infrastructure firms, with UK-based Nscale on top after combined Series B and C rounds and Netherlands-based Nebius close behind. A small group of very large growth-stage deals accounted for most of the capital, while smaller seed and early-stage rounds spread across cloud optimization, sovereign cloud, and developer platforms. Funding concentrated in the Netherlands, Italy, Germany, Spain, and the UK, with Western Europe leading on both deal count and dollars raised.

TurnUp raises €2 million to help healthcare providers reduce no-shows

TurnUp, a health technology startup based in Ghent, Belgium, has raised 2 million euros in a seed funding round led by Newion, with participation from RDY Ventures. The company sells an AI platform that predicts which patients are likely to miss appointments and automates confirmations, rescheduling, and waiting list management for dental and medical practices. TurnUp already serves more than 250 practices across Belgium, the Netherlands, and the UK, and says its software has prevented over 500,000 no-shows.

Deezer launches an AI music detector for other streaming services

Deezer has released a free online tool, the AI Music Detector, that scans your playlists on rival streaming services to flag AI-generated songs. The tool works with about 20 platforms, including Spotify, Apple Music, SoundCloud, and YouTube Music, and supports 27 languages. Deezer built the technology after finding that roughly 44 percent of tracks uploaded to its own platform each day are now fully AI-generated. The move is the French streamer's latest push to make AI music visible to listeners after rivals declined to adopt its detection system.

sunbay.io raises €550K to automate invoice collection for finance teams

Warsaw-based sunbay.io raised 550,000 euros (about 2.33 million Polish zloty) in a round led by Kogito Ventures, with backing from s20 and a group of angel investors. The startup sells software that automates accounts receivable, the job of chasing money owed on overdue invoices, using email, SMS, and AI voice calls. The company already serves more than 20 businesses across five countries and plans to spend the new capital on product development and added AI agents.

Mendo secures €12M to scale enterprise AI adoption in Europe

Mendo, a Paris-based startup founded in 2021, raised 12 million euros in Series A funding to help large companies put generative and agentic AI to work for their employees. Ventech and Educapital led the round, with Tomcat and OVNI also taking part. The deal brings Mendo's total funding to 15.5 million euros and will fund European expansion and a doubling of staff from 50 to 100 people.

Opendoor's India exit is fueling a bigger conversation about AI and outsourcing

Opendoor shut down its entire India operation less than two years after opening offices in Chennai and Bengaluru in 2024, eliminating about 250 roles. CEO Kaz Nejatian said the company now favors small AI-native teams handling work closer to its U.S. customers, because automation and consolidated internal systems reduced the need for large offshore operational staff. The decision has become a talking point across Silicon Valley as an early test of whether AI is changing the economics of outsourcing to India.

Google Sheets pivot table: A step-by-step guide

A pivot table reorganizes a large flat spreadsheet into a compact summary so you can answer questions like how much each client was billed per project without writing formulas. You build one by selecting your data, opening the Insert menu, and choosing Pivot table, then placing fields into Rows, Columns, Values, and Filters. Google Sheets now lets Gemini, its built-in AI, create the table from a plain-language request, turning several manual steps into one sentence for non-technical users.

Anthropic's Dario Amodei has just one direct report

Anthropic chief executive Dario Amodei manages a single person directly, his chief of staff Avital Balwit, a setup he described to Bloomberg as "incredibly freeing." Every other senior leader reports to his sister and company president Daniela Amodei, who handles daily operations. The arrangement lets Dario focus on strategy, research direction, culture, and policy at a company that private investors value near $1 trillion. It breaks sharply from the wide management spans seen at peers such as OpenAI and Nvidia.

Anthropic Walks Back Policy That Could Have 'Sabotaged' AI Researchers Using Claude

Anthropic reversed a hidden safeguard in its new Claude Fable 5 model that quietly weakened the assistant when people asked it to help build competing AI systems. The restriction, disclosed in a single paragraph inside a 319-page system card, rerouted certain frontier research requests to a weaker model without telling the user. After AI researchers and policy analysts criticized the move within hours of the June 9, 2026 release, the company apologized and said it will make any such limits visible rather than silent.

Anthropic's Dario Amodei wants governments to have the power to block 'dangerous' AI systems

Anthropic PBC Chief Executive Dario Amodei is calling on the U.S. government to block the deployment of dangerous artificial intelligence models in the same way as it prevents unsafe airplanes from taking off. In a new post today on his personal blog, Amodei (pictured) said there need to be mandatory third-party audits of frontier AI [...] The post Anthropic's Dario Amodei wants governments to have the power to block 'dangerous' AI systems appeared first on SiliconANGLE.

How frontier teams are reinventing AI-native development

AWS says the largest gains from AI in software development come from teams that redesign how they work around AI agents, not from teams that treat AI as a faster way to type. Across more than 50 internal teams studied, the 25 groups that adopted both new tools and new practices reached a 4.5x median productivity gain, and some passed 10x in deployment velocity. AWS calls these its frontier teams and lays out five habits behind the results.

Google open-sources speedy DiffusionGemma text diffusion model

Google LLC today released DiffusionGemma, a large language model based on an emerging machine learning approach known as text diffusion. The company says the algorithm can generate text four times faster than traditional LLMs. Furthermore, DiffusionGemma does so using less RAM. The model's memory efficiency enables it to run on high-end consumer graphics cards that [...] The post Google open-sources speedy DiffusionGemma text diffusion model appeared first on SiliconANGLE.

DXC will integrate Claude into the systems banks, airlines, and other regulated industries rely on

DXC Technology and Anthropic announced a multi-year global alliance on June 11, 2026 to put Claude inside the core software that banks, airlines, insurers, manufacturers, and government agencies depend on. DXC will train tens of thousands of staff as Claude-certified engineers who embed the AI directly into customer systems. The deal reaches regulated sectors where security and compliance rules are strict, backed by DXC's 115,000 employees across 70 countries.

How an astrophysicist uses Codex to help simulate black holes

University of Arizona astrophysicist Chi-kwan Chan uses OpenAI's Codex to write, debug, and refine the scientific code behind black hole simulations. Chan works with the Event Horizon Telescope collaboration, the international team that produced the first image of a black hole in 2019 and is now gathering data to make the first time-lapse video of a supermassive black hole. The simulations model the superheated plasma swirling near black holes and let scientists test Einstein's general theory of relativity against real telescope observations.

Profiling in PyTorch (Part 2): From nn.Linear to a Fused MLP

Hugging Face published the second part of its PyTorch profiling series, showing how to read GPU performance data and speed up a common AI building block by packaging its math into fewer kernels. The standard version fired 5 separate GPU kernels per forward pass, while compiling the same model with torch.compile cut that to 4 by fusing the small pointwise steps into one kernel. The fused path skipped roughly 50 MB of intermediate memory traffic per pass and ran in 89.4 microseconds, slightly faster than a hand-tuned alternative at 92.8 microseconds.

OpenAI to acquire Ona

OpenAI plans to acquire Ona, the German cloud platform formerly known as Gitpod, to give its Codex coding agent secure environments where work keeps running for hours or days after a developer logs off. Ona has supported about 2 million developers in reproducible cloud sandboxes. Financial terms were not disclosed, and the deal is subject to regulatory approval.

Supporting Europe's work in ensuring a trustworthy AI ecosystem

OpenAI said it supports the European AI Office's Code of Practice on Transparency of AI-Generated Content, a voluntary framework released on June 10, 2026 that helps companies meet the EU AI Act's rules for labeling AI-made content. The Code points to two main methods, signed metadata and invisible watermarks, that let people and platforms trace whether an image, video, or audio clip was generated or edited by AI. OpenAI tied the move to provenance work it has run since 2024, including the C2PA metadata it adds to images from DALL-E 3 and Sora.

BBVA puts AI at the core of banking with OpenAI

BBVA, the Spanish multinational bank, is giving ChatGPT Enterprise to all of its roughly 120,000 employees across 25 countries under a multi-year strategic alliance with OpenAI. The expansion, announced on December 12, 2025, follows nearly two years of testing that grew from a 3,300-account pilot to 11,000 paid licenses, and it ranks among the largest enterprise deployments of generative AI in banking.

Introducing Claude Corps

Anthropic is committing an initial $150 million to Claude Corps, a national fellowship that will train 1,000 early-career fellows to use Claude and place them full-time inside U.S. nonprofits for a year. Each fellow earns an $85,000 salary plus benefits, and at least 400 nonprofits will host fellows over the first 12 months. Anthropic frames the program as a way to share AI's benefits while workers absorb economic change, and announced it alongside a new policy framework on AI's impact on work.

Neura Robotics to raise up to $1.4B from Nvidia-backed consortium

Automation startup Neura Robotics GmbH is raising a funding round worth up to $1.4 billion from a group of prominent investors. The company detailed today that the consortium includes Amazon.com Inc., Nvidia Corp., Qualcomm Technologies Inc. and cryptocurrency issuer Tether. The round is also expected to draw contributions from outside the tech sector. The European [...] The post Neura Robotics to raise up to $1.4B from Nvidia-backed consortium appeared first on SiliconANGLE.

xAI fired an engineer who raised alarms about Grok safety, new lawsuit claims

A former xAI engineer named Devin Kim has sued xAI and SpaceX in California state court, claiming he was fired in retaliation for repeatedly raising safety concerns about the Grok chatbot. The suit landed days before SpaceX's stock market debut, set to be the largest initial public offering on record at roughly $75 billion raised. Kim, who left xAI in September 2025, now serves as president of the nonprofit Center for AI Safety.

Visa partners with OpenAI to let AI agents make payments for users

Visa Inc. has struck a deal with OpenAI Group PBC to let artificial intelligence agents make payments for users, bringing one of the world's largest payment networks into ChatGPT's push toward agentic commerce. The companies announced the partnership at the Visa Payments Forum today in San Francisco. Under the agreement, Visa's payment tools will be [...] The post Visa partners with OpenAI to let AI agents make payments for users appeared first on SiliconANGLE.

Automated governance is FinOps' next frontier as AI spend spreads beyond engineering

Artificial intelligence is creating a new class of cloud cost challenge - one that no longer belongs exclusively to engineering teams. As AI tools put spend capabilities into the hands of sales, finance and executive teams, the need for automated governance has become as urgent as the innovation it is meant to protect. The FinOps [...] The post Automated governance is FinOps' next frontier as AI spend spreads beyond engineering appeared first on SiliconANGLE.

HPE's Unleash AI takes aim at the 'AI pilot trap'

For all the excitement artificial intelligence has generated, success is still eluding many companies. A recent PricewaterhouseCoopers LLP study found that just 20% of enterprises are achieving at least three-quarters of the revenue and efficiency gains AI promises. Gartner estimates that at least half of generative AI projects were abandoned last year, and other estimates [...] The post HPE's Unleash AI takes aim at the 'AI pilot trap' appeared first on SiliconANGLE.

FinOps AI goes beyond token economics as agentic costs emerge

As FinOps AI strategies continue to emerge, the familiar cloud cost management approach is breaking down - and organizations that fail to adapt risk runaway spending on workloads they barely understand. FinOps is rapidly evolving from a cloud-billing function into a strategic framework for governing the full technology stack, including AI, software-as-a-service and now autonomous agents. [...] The post FinOps AI goes beyond token economics as agentic costs emerge appeared first on SiliconANGLE.

Dell's Microsoft/AMD collaboration: Three insights you may have missed from theCUBE's coverage of Dell Technologies World

Strategic AI partnerships play a central role in the deployment of Dell Technologies Inc.'s initiatives. The company's alliances with chipmakers, virtualization leaders, cloud providers, and a host of leading software platforms have allowed it to build a platform that makes compute, storage, networking, and workloads work cohesively across hybrid and multicloud worlds. Two of Dell's [...] The post Dell's Microsoft/AMD collaboration: Three insights you may have missed from theCUBE's coverage of Dell Technologies World appeared first on SiliconANGLE.

What to expect during Pure Accelerate: Join theCUBE June 17

Enterprise storage is being repositioned for the artificial intelligence era, and one of the leading players in the field has actively pursued an agenda aimed at operationalizing data. Everpure Inc., which rebranded in February from Pure Storage, has unveiled a set of actions during the first half of this year designed to position storage as an [...] The post What to expect during Pure Accelerate: Join theCUBE June 17 appeared first on SiliconANGLE.

Fresh off bond sale, Amazon borrows $17.5B from banks as AI spending continues

Amazon arranged a $17.5 billion loan from a syndicate of more than a dozen banks led by Citigroup, the latest sign of how much it is willing to borrow to fund its AI buildout. The deal landed two days after Amazon raised about $10 billion in a record Canadian-dollar bond sale, adding roughly $27 billion in fresh financing inside 48 hours. The new borrowing pushed Amazon's total debt past $225 billion, up about 50 percent from a year earlier.

FinOps adapts to AI spend as token economics reshape enterprise budgets

As generative AI accelerates from a product experiment into a core enterprise operating cost, FinOps is evolving rapidly to manage AI spend, introducing a layer of complexity that traditional cloud budgets have never fully prepared practitioners to handle. Token economics are forcing organizations to rethink not just how they measure spend, but what costs even count [...] The post FinOps adapts to AI spend as token economics reshape enterprise budgets appeared first on SiliconANGLE.

Access OpenAI models and Codex through your Oracle cloud commitment

OpenAI and Oracle announced on June 10, 2026 a deal that lets Oracle Cloud Infrastructure customers access OpenAI frontier models and the Codex coding tool through the OCI Marketplace. The defining detail is payment: businesses apply their existing Oracle Universal Credits toward OpenAI usage instead of setting up a separate contract. The offer is expected to go live in the coming weeks, handled through Oracle sales channels.

For Robotaxis, Safety Must Be Built In, Not Bolted On

NVIDIA introduced Halos OS, a unified safety foundation designed to build vehicle safety into the core of robotaxi systems rather than adding it later. The system runs on NVIDIA DRIVE Hyperion, meets the ISO 26262 ASIL D automotive safety standard, and is structured in layers that isolate safety-critical software from the rest of the driving stack. NVIDIA frames the move as a response to robotaxis shifting from test pilots to paid commercial rides in dozens of cities.

Claude Fable won't answer basic biology questions

Anthropic's newest public model, Claude Fable 5, refuses to answer basic biology questions on purpose. It is the first publicly available member of the Mythos model family, a class so capable at cybersecurity and biology that Anthropic considers the full version too dangerous to release. When a user asks about biology, chemistry, cybersecurity, or model distillation, Fable 5 hands the request off to the older Claude Opus 4.8 instead of answering itself.

The Crunchbase Tech Layoffs Tracker

More than 127,000 workers at U.S.-based technology companies lost their jobs in 2025, according to a running tally kept by Crunchbase News, and the cuts have carried into 2026. The deepest reductions came from established giants, with Intel eliminating 27,159 roles, Microsoft 15,387 and Amazon 14,709. Crunchbase ties much of the wave to cost cutting and a shift of hiring budgets toward AI development and automation. The tracker logs new company reductions week by week, including 4,375 employees affected in the seven days ending June 10, 2026.

Microsoft, like, totally gets why students are booing AI-pilled graduation speakers

Across spring 2026 commencements, new college graduates booed and heckled speakers who praised artificial intelligence, including former Google CEO Eric Schmidt at the University of Arizona. Microsoft vice chair and president Brad Smith responded with a blog post running more than 3,000 words, calling the reaction a wake-up call for the tech industry and asking graduates to adapt rather than resist. Critics note the essay offered sympathy and historical analogies but no concrete commitments on jobs, retraining, or the pace of AI rollout.

The future of AI regulation is courting the strangest, most anxious bedfellows

AI regulation has become one of the defining political battles of the 2026 midterms, drawing together an unusual cast of senators, billionaire-funded super PACs, AI safety advocates, and anxious voters. The Washington AI Network's second annual AI Honors gala, held June 3, 2026, put that tension on display by celebrating bipartisan lawmakers and industry figures even as polling showed 70 percent of Americans worried about AI's role in elections. The central fight is whether the federal government sets one national rulebook for AI or whether individual states keep writing their own laws.

Google won't just admit it's feeding YouTube creators to its music AI

A group of independent musicians is suing Google, claiming the company trained its Lyria 3 music-generation AI on their songs uploaded to YouTube without permission or pay. Google has filed a motion to dismiss, arguing that YouTube's terms of service already grant it a broad license to use those uploads, and that the musicians have not proven their specific works were used. Notably, Google has not directly confirmed whether it trained Lyria on the songs in question. The proposed class action sits in the US District Court for the Northern District of Illinois.

'AI-pilled' firms spend $7,500 per employee each month on AI

The heaviest corporate AI users, the top 1% of firms Ramp calls "AI-pilled," now spend roughly $7,500 per employee each month on AI tools, while the median business spends only $11.38. That is a gap of about 680 times between the most aggressive adopters and the middle of the pack. The Ramp AI Index draws this picture from anonymized spending data across more than 70,000 U.S. businesses, showing how concentrated heavy AI investment remains.

Microsoft restricts Claude Fable for employees over data retention concerns

Microsoft restricted its own employees from using Anthropic's newly released Claude Fable 5 model inside its internal version of GitHub Copilot because the model requires data retention that breaks Microsoft's Zero Data Retention standard. The model retains prompts and outputs for 30 days, and flagged content for up to two years, which Microsoft legal teams are reviewing over concerns about customer data and confidential information. At the same time, Microsoft made Fable 5 available to outside GitHub Copilot and Azure Foundry customers, creating a gap between what it sells and what it lets staff use internally. Older Claude models that run under Zero Data Retention remain available to Microsoft employees.

DiffusionGemma: 4x faster text generation

Google DeepMind released DiffusionGemma, an experimental open-weights model that writes text through a diffusion process instead of one word at a time, running up to 4x faster than standard models on GPUs. The model is a 26B Mixture of Experts design that activates only about 3.8B parameters per step and generates 256 tokens in parallel. It reaches more than 1,000 tokens per second on a single NVIDIA H100 and ships free to use under the Apache 2.0 license.

Google will save your Lens photos, Search Live recordings, and Translate audio for AI training

Google is rolling out a new privacy setting called Search Services History that saves the images, files, audio, and video you submit when you use Search. The setting includes a Save Media option, on by default for many users, that lets Google use those Lens photos, Search Live recordings, voice searches, and Translate audio to develop and improve its AI models. You stay in control: you can turn Save Media off, switch off Search Services History entirely, or delete individual items from your history.

NVIDIA Accelerates Google DeepMind's DiffusionGemma for Local AI

Google DeepMind released DiffusionGemma, an experimental open model that writes text by refining whole blocks at once instead of one word at a time, and NVIDIA has tuned it to run faster across its GeForce RTX, RTX PRO, and DGX hardware. The model reaches more than 1,000 tokens per second on a single NVIDIA H100 GPU and over 700 tokens per second on a consumer RTX 5090, roughly four times the speed of a comparable standard model. It targets fast, single-user work such as chat and on-device assistants where low delay matters most. The weights are open under the Apache 2.0 license and free to test through NVIDIA-hosted APIs.

How memory tools can make AI models worse

New research from the enterprise AI company Writer finds that the memory features many AI assistants rely on make models more likely to tell users what they want to hear instead of what is true. Across the tested setups, stored memory raised sycophantic answers by up to 25 times compared with feeding the same details straight into a prompt. The team traced the problem to lossy compression, which preserves a user's stated beliefs while discarding the surrounding context needed to correct them.

Cybersecurity researchers aren't happy about the guardrails on Anthropic's Fable

Anthropic released Fable, a public version of its Mythos cybersecurity model, and security researchers say its safety guardrails are too strict to do real cyber defense work. When a prompt touches cybersecurity or biology, Fable pauses and routes the request to the weaker Claude Opus 4.8. Researchers report that ordinary tasks, including code review, secure-coding requests, and even reading a blog post, get blocked.

Stop hand-tuning kernels: How Neuron Agentic Development accelerates AWS Trainium optimizations

AWS released Neuron Agentic Development, an open-source set of AI agents and skills that lets developers write, debug, and profile high-performance kernels for AWS Trainium and Inferentia chips using plain language instead of deep hardware expertise. The tools plug into agentic coding environments such as Claude Code and Kiro and walk a developer through the full kernel workflow, from authoring code to finding performance bottlenecks. The aim is to close the long-standing gap between what the hardware promises and what most teams reach in practice.

NEURA Robotics secures up to $1.4B Series C to scale physical AI and cognitive robotics platform

NEURA Robotics, a German maker of cognitive and humanoid robots, has secured up to $1.4 billion in Series C funding led by stablecoin issuer Tether, valuing the company at about $7 billion. Backers include Qualcomm, Amazon, NVIDIA, Bosch, Schaeffler, and the European Investment Bank. The company describes it as the largest funding round ever raised by a full-stack robotics firm and will use the money to scale mass production of robots that work alongside people in the physical world.

Build an AI-Powered Equipment Repair Assistant Using Amazon Bedrock AgentCore

AWS published a step-by-step guide for building an AI-powered equipment repair assistant on Amazon Bedrock AgentCore, aimed at farmers and field technicians who need fast machinery diagnosis. The assistant lets users describe a problem in plain language, then pulls answers from manufacturer manuals, identifies the parts needed, and walks through approved repair steps. It runs on AgentCore Runtime with the Strands Agents SDK, uses Amazon Nova 2 Lite as the model, and adds a Bedrock Knowledge Base for document retrieval plus AgentCore Memory to remember past conversations.

LLM Routing: From Strategy Selection to Production Architecture

LLM routing is a design pattern where a control layer reads each incoming AI request and sends it to the best-fit model instead of pushing everything to one expensive default. A new guide from n8n explains how this trims cost and response time by matching simple questions to cheaper, faster models while reserving premium models for hard work. The guide cites Berkeley's RouteLLM, which holds 95% of GPT-4 quality while cutting cost by more than 85% on one benchmark, and FrugalGPT, which matched GPT-4 quality at up to 98% lower cost.

Best AI Governance Tools for Enterprise-Grade Compliance

The n8n Blog compares eight AI governance platforms that help large companies keep AI systems secure, transparent, and accountable as the technology moves from testing into live production. The lineup includes Credo AI, IBM watsonx.governance, Holistic AI, Collibra, OneTrust, Fiddler AI, Monitaur, and n8n itself for workflow controls. Each tool maps to rules such as the EU AI Act, NIST AI RMF, and NYC Local Law 144, and the guide explains how to pick one based on model inventory, audit evidence, and how deeply the tool plugs into a company's existing systems.

Top Process Orchestration Tools and Their Key Benefits

A guide from n8n reviews seven process orchestration platforms and the features that set them apart: n8n, Camunda, Appian, Microsoft Power Automate, Pega, Temporal, and Salesforce Flow. Process orchestration coordinates the steps, systems, and people inside a larger business workflow, going beyond the single isolated tasks that simple automation handles. The guide cites a 2025 Camunda survey of 800 IT leaders, in which 93 percent said AI must be fully integrated into orchestrated processes to maximize return. Because swapping an orchestration platform is hard once a team is trained on it, the guide urges consolidating on one tool rather than fragmenting governance across several.

Datadog veterans launch AI coding startup Niteshift on a bet against Big AI lock-in

Niteshift, an AI coding startup founded by two early Datadog engineers, has raised a $7 million seed round led by Greylock partner Jerry Chen. The company runs a cloud platform that routes coding work across rival AI models such as GPT, Claude, and open source options, picking the best fit for each project. Its core pitch is freedom of choice, betting that businesses want control over their developer tools instead of being locked into a single model maker.

Save time and grow your business with new Gemini tools

Google added two features to the Gemini app built for small business owners: a one-tap link to a company's Google Business Profile and a workspace called Business notebooks. Once connected, Gemini reads the reviews, customer questions, and performance data tied to the profile, then drafts review replies, updates hours, and answers questions about how the business is doing. Announced at Google for Brazil 2026 on June 10, both tools start rolling out globally during June, with the EEA and the UK excluded at launch.

The three hard-tech moonshots fueling SpaceX's unbelievable IPO

SpaceX priced the largest IPO on record at $135 per share, valuing the rocket maker near $1.77 trillion and raising about $75 billion through the sale of roughly 555.6 million shares. Most of that price rests not on today's rocket and Starlink business but on three unproven hard-tech bets: fully reusable Starship rockets, mass-produced AI satellites, and a chip foundry. Independent analysts value the company far lower, with Morningstar at $825 billion, treating the gap as a wager on space-based data centers.

Akash Systems brings diamond cooling to AI infrastructure

Akash Systems Inc. believes it has a solution for the heat problem of graphics processing units: lab-grown diamonds. The company originally got its start in space, managing solar radiation on satellites, and now wants to bring its technology to Earth. Akash Systems' diamonds remove heat incredibly fast, according to Pamit Surana (pictured), co-founder and CCO [...] The post Akash Systems brings diamond cooling to AI infrastructure appeared first on SiliconANGLE.

Warner Music acquires AI attribution startup Sureel AI

Warner Music Group agreed on June 10, 2026 to acquire Sureel AI, a Palo Alto startup whose technology traces how artificial intelligence systems use songs in training and generated output. Financial terms were not disclosed. Warner wants to track when its artists' and songwriters' work feeds AI models so rights holders share in the value created, and Sureel will keep running as a standalone platform for the wider music and AI industry.

Wrongful Arrest Exposes Failures in One of the Oldest Police Face-Recognition Tools in the US

A face-recognition program reported a 93 percent confidence match and sent Robert Dillon, a 52-year-old commercial crabber from Fort Myers, Florida, to jail for an attempted child abduction he had no link to. The ACLU and ACLU of Florida filed a federal lawsuit on June 10, 2026, against three police agencies, arguing officers treated a probabilistic software guess as a near-certain identification while ignoring evidence that cleared him. The case puts a spotlight on one of the oldest police face-recognition tools in the country, operated by the Pinellas County Sheriff's Office.

The intelligence layer emerges as the control plane for enterprise AI

As enterprises move beyond AI experimentation into full-scale production, the central challenge has shifted from accessing models to managing the organizational context they need to act reliably. The pressure to govern costs, secure data and maintain accountability is now redefining how companies architect their entire AI intelligence layer. That convergence of AI adoption and infrastructure [...] The post The intelligence layer emerges as the control plane for enterprise AI appeared first on SiliconANGLE.

China Opens World's First Wind-Powered Underwater Data Center

China has put the world's first offshore wind-powered underwater data center into full operation off the coast of Shanghai's Lingang Special Area. Built by Shanghai Hailanyun Technology, known as HiCloud, the submerged site runs at a capacity of 24 megawatts, houses close to 2,000 servers, and draws about 95 percent of its electricity from a nearby offshore wind farm. Surrounding seawater cools the racks passively, removing most of the power that land-based centers spend on air conditioning.

Jedify raises $24M to help companies arm AI agents with context on their business

Jedify, a New York startup, raised $24 million in a Series A round led by Norwest to help companies feed AI agents context about their own operations. The round lifted total funding to about $33 million and brought Snowflake Ventures in as a strategic investor that is integrating Jedify into its AI products. The pitch is that enterprise agents stay weak until they understand a company's data, permissions, workflows, and internal terminology.

Capsa AI raises $18M to expand its AI platform for private capital

Capsa AI, a startup with offices in London and New York, raised $18 million in a Series A round co-led by TX Ventures and Pivot Investment Partners. The financing lifts the company's total funding to $20 million since it was founded in 2024. Capsa AI builds an AI operating system for private capital firms, pulling scattered fund data into one searchable layer, and will use the money to expand in the United States and grow its teams.

Decart's new world model can simulate hours of photorealistic driving - with some caveats

Decart, a two-year-old AI startup, released Oasis 3, a real-time world model that builds photorealistic driving scenes for testing self-driving cars. The launch follows a $300 million funding round that values the company at close to $4 billion, with strategic money from Toyota, Adobe, eBay, and Nvidia. The model is open to developers through an API at $0.02 per second, though early hands-on testing by TechCrunch found the simulation still breaks down on physics and scene consistency.

Hamburg-based Generation Tech Partners launches €50M AI roll-up fund

Generation Tech Partners, a new Hamburg investment firm, has launched a fund of more than 50 million euros to buy established German mid-market service companies and rebuild them around artificial intelligence. The firm plans to acquire about 30 owner-managed B2B service businesses for a combined purchase volume of up to 120 million euros, then run them with a proprietary AI transformation playbook. The strategy targets firms whose founders face succession problems and aims to grow them without cost-cutting or layoffs.

Nanordica Medical raises €1.6M to bring antibiotic-free chronic wound treatment to market

Estonian medtech company Nanordica Medical raised 1.6 million euros to bring Premotiv, an antibiotic-free dressing for chronic wounds, to the European market. The dressing uses copper and silver nanoparticles to fight infection and speed healing without antibiotics, and it targets hard-to-heal cases such as diabetic foot ulcers. The round was led by Estonian fund 2C Ventures, and the money funds a large clinical study and the safety certification needed to sell across Europe.

PRC-linked influence operations are targeting AI debates in the US

OpenAI said it banned two China-linked covert influence operations that used ChatGPT to generate social media comments, images, and political cartoons aimed at inflaming US public debates over AI data centers and trade tariffs. Both campaigns scored Category One, the lowest tier on OpenAI's Breakout Scale, meaning they stayed on a single platform and reached no genuine audience. The findings, published June 10, 2026, show how state-linked actors are testing AI tools to manufacture political content cheaply, even when the real-world reach stays small.

RAISE Summit Returns to Paris on July 8-9, 2026 at the Carrousel du Louvre [Sponsored]

RAISE Summit 2026 returns to the Carrousel du Louvre in Paris on July 8 and 9, billed as one of Europe's largest gatherings of artificial intelligence leaders. Organizers expect more than 9,000 attendees, with over 80 percent at the C-level, founder, or senior decision-maker level, and a program of 350 speakers. The two days include a startup competition with a prize pool above 10 million euros and an AI hackathon described as the largest of its kind.

Sector Snapshot: Semiconductor Startup Funding Still Running Hot

Investors have put roughly $10.7 billion into semiconductor startups during 2026 so far, a pace set to pass all of 2025. The standout event was the Cerebras Systems IPO in May, which raised about $5.5 billion. Three other chip startups, MatX, Ayar Labs, and Etched, each pulled in $500 million private rounds, with money concentrating on a small group of firms building hardware to challenge Nvidia in artificial intelligence.

Timing Trick Cuts Energy Used in LLM Training by Up to 14 Percent

Researchers at the University of Twente in the Netherlands cut the energy used to train a large language model by up to 14 percent without slowing it down, by finely tuning how fast a GPU's clocks tick during computation. The method, presented at the Computing Frontiers conference in Sicily, adjusts the chip's speed at the level of tiny tasks called kernels rather than across whole training steps. In tests training a 1.3 billion parameter model, the approach saved roughly 14.6 percent of energy with only a 0.6 percent increase in training time. For an industry where a single frontier model trains on tens of gigawatt-hours, that share of savings is large.

Investing in multi-agent AI safety research

Google DeepMind and four partners opened a research funding call worth up to $10 million to study how AI agents behave when many of them interact at scale. The money targets safety risks that surface only when independent agents built by different organizations work together, a gap most current safety testing ignores. Academic and independent researchers worldwide can apply by August 8, 2026, with award decisions expected in autumn 2026.

Artificial Intelligence Sneaks Into the World Cup Thanks to Google Gemini

Google has made its Gemini AI a main global sponsor of Argentina's national football teams, placing the Gemini logo on training kits and building the technology into the defending champion's preparation for the 2026 FIFA World Cup. Argentina's coaching staff and players use Gemini for tactical analysis, injury prevention, and breakdowns of their own and opponents' statistics. It is one of three national-team deals Google struck, alongside France and the United States, to put Gemini in front of a global audience during the tournament.

Uncovr raises $7M to build AI infrastructure for surgery

Uncovr, a surgical AI startup based in New York and Paris, raised $7 million in seed funding led by Index Ventures to build software that documents what happens inside the operating room. The product reads surgical and endoscopic video in real time and turns it into structured clinical records, draft operative reports, and billing code suggestions before the surgeon leaves the room. Every output requires surgeon review and sign-off. The company already has a deployment pipeline of more than 400 operating rooms across the United States and Europe.

Legora to open Paris, Milan and Madrid offices and London engineering hub

Legora, a Swedish legal AI company, is opening offices in Paris, Milan and Madrid and adding an engineering hub in its existing London base. The firm aims to more than double its workforce across Europe, the Middle East and Africa to over 700 people within the next 6 to 12 months, up from about 325 today. The move follows a Series D round that reached $600 million and valued the company at $5.6 billion in April 2026.

Meta signs first AI data center deal in India with Reliance

Meta has signed its first dedicated AI data center deal in India, leasing a 168-megawatt facility in Jamnagar, Gujarat, that Reliance Industries will design, build, and operate. The site will run on renewable power and use desalinated seawater for cooling, with Meta paying the full cost of the energy and water it consumes. The data center is expected to be ready within two years and is built to expand over time, giving Meta secured computing capacity in one of the fastest-growing AI markets.

5 ways to automate Meta's Conversions API tool with Zapier

Zapier published a guide showing five no-code ways to feed customer actions into Meta's Conversions API, the server-side tool that sends marketing signals straight to Meta without relying on a browser pixel. The workflows automatically pass events like purchases, form submissions, CRM stage changes, and new email subscribers from apps such as Stripe, Shopify, HubSpot, Salesforce, Typeform, and Klaviyo into Meta so ad targeting stays current. Setup runs through Meta's Events Manager, where Zapier is offered as a partner integration. For tools without a ready-made connection, Webhooks by Zapier sends custom data instead.

How to automate Claude with Zapier

You can connect Anthropic's Claude to the rest of your business tools through Zapier, the no-code automation platform that links more than 9,000 apps and supports over 30,000 actions. Instead of typing one prompt at a time, you set up automated workflows, called Zaps, where Claude reads, writes, summarizes, or analyzes content the moment something happens in another app. Zapier offers one-click templates for common jobs such as drafting replies, summarizing feedback, and routing approvals, so non-technical teams set them up without code. Claude Opus 4.8 is available inside Zapier today, alongside Anthropic's newer Mythos model class.

Rotomate raises €2.1M to automate industrial reliability analysis

01Health secures $15M to scale specialist healthcare platform

UK healthtech company 01Health has raised $15 million in Series A funding, equal to about £11.2 million, led by Gresham House Ventures. The money supports a wider rollout of its platform that lets local clinics deliver specialist care normally confined to large hospitals, plus a planned expansion into the United States. Founded in 2022 by former NHS doctor Dr Sonia Szamocki, the company has now raised $25 million in total and employs more than 100 people.

Which AI models can you automate on Zapier? (Opus 4.8, Gemini 3.5 Flash, and more)

Zapier now lets business teams plug a wide menu of AI models into their automated workflows, spanning OpenAI, Anthropic, Google, and specialty providers like Mistral and DeepSeek. To help users pick the right one, Zapier ranks each model with AutomationBench, its own test of how well models complete multi-step business tasks rather than one-off prompts. Anthropic's Claude Fable 5.0 leads the overall leaderboard, with Claude Opus 4.8 and Google's Gemini 3.5 Flash close behind.

How a two-person SEO shop is building an engine to run twelve clients in thirty minutes a month

Adrian Martinez, who runs the Toronto agency Hire Adrian with his wife, is building an AI delivery system to serve roughly twelve SEO and answer engine optimization clients while keeping the team at two people. The goal is to shrink hands-on work from 10 to 15 hours per client each month down to about 30 minutes, without hiring account managers. He pairs Claude for planning and decisions with Zapier MCP, which connects the agent to WordPress, reporting tools, and image generation so plans turn into real production work.

How Gourmet Ads uses Zapier MCP to turn Salesforce and Atlassian into a weekly growth report

Gourmet Ads, an 18-year-old digital advertising business, wired Salesforce and Atlassian Confluence together through Zapier MCP so an AI assistant produces a weekly growth report on its own. The report bundles six parts and five recommended actions, each scoped to take under 20 minutes, and it has already caught fake traffic, a broken link, and a sales pipeline gap. President Benjamin Christie says the setup lifted his personal output by 30 to 40 percent and turned a multi-day engineering job into something he tests from his desk.

Startup's nuclear-inspired cooling system could make data centers more sustainable

Ferveret, a startup founded by two MIT nuclear engineering researchers, has built a liquid cooling system for AI data center chips that uses zero water and reports a 15% gain in computational power efficiency over current liquid cooling. The company borrows a heat-transfer method called subcooled boiling from inside nuclear reactors, submerging servers in a special liquid to pull heat off chips far faster than air. Paired with its power-control software, the setup lets AI models generate 35% more tokens from the same amount of power. Early testing partners include CleanSpark, FuriosaAI, and Switch.

The 17 best AI marketing tools in 2026

Zapier's 2026 roundup picks 17 AI marketing tools and sorts them by the job a marketer needs done, from making visuals to tracking whether a brand shows up in ChatGPT answers. The core message is that most marketers do not need every AI feature on the market. The author, who tested the tools, argues the bigger win comes from matching a few tools to real workflows and connecting them so they work together instead of in silos.

Google just fired a warning shot in the AI subscription price wars

Google cut the monthly price of its budget AI subscription, Google AI Plus, from $7.99 to $4.99 and doubled the included cloud storage from 200GB to 400GB. The change brings a price war that had been building in emerging markets such as India squarely to American consumers. At $4.99, Google now offers the lowest-priced AI subscription from a top-tier provider, a move analysts read as a sign that raw AI capability is becoming a commodity.

From data to decisions: how LSEG is scaling trusted AI

London Stock Exchange Group (LSEG), one of the world's largest financial markets infrastructure firms, deployed OpenAI's ChatGPT Enterprise and OpenAI APIs across its global business, giving an initial 4,000 employees access to AI tools inside a secure environment. The shift cut many product release cycles from three to six months down to two weeks, and trimmed some customer delivery from as long as nine months to roughly four weeks. LSEG built governance in from the start, including model evaluation, human review, and strict data controls.

I tried Siri AI, and so far it actually works

Apple's rebuilt Siri AI, shown at WWDC 2026, finally works for everyday multistep tasks, according to a hands-on test by The Verge. The new assistant adds a batch of events from an email or a messy flyer to your calendar in one shot, builds shopping lists, sets reminders, and answers practical questions by reading your own email, messages, and calendar. The Verge's writer tried each scenario and confirmed it worked, while noting the feature set still trails what Google's Gemini has done on Android for two years. The upgrade runs partly on a custom Google Gemini model hosted on Apple's privacy servers.