This Week in NLP #361
Keep up with what happened in NLP in the week ending Friday 17th October 2025.
Above the Fold
Overwhelmed? Here’s our pick for five things to know about this week.
Anthropic released Claude Haiku 4.5, a smaller, faster, and significantly cheaper AI model that matches recent cutting-edge coding capabilities at one-third the previous cost. [VentureBeat]
Microsoft is reimagining Windows 11 with AI-powered voice interaction, screen analysis, and autonomous software agents, aiming to transform how people interact with their computers. [VentureBeat]
OpenAI plans to allow verified adults to use ChatGPT for erotica and mature applications starting in December, following age-verification implementations. [Engadget]
Salesforce has launched Agentforce 360, an AI platform designed to enable intelligent agents to autonomously perform business tasks under human oversight. [TechRepublic]
And here’s Nathan Benaich’s must-read 2025 State of AI Report. [Google]
Now read on for everything else that happened in NLP this week.
Making News:
What makes this newsletter different?
This Week’s Topics:
If you’re reading this in a desktop web browser, you can access a navigation menu at the left-hand edge of this window.
The Generative AI Wars
Anthropic projected rapid revenue growth to potentially reach US$26B by 2026, driven by strong enterprise AI product adoption and expanding international market presence. [Reuters]
Anthropic has launched a Microsoft 365 connector for Claude, enabling direct access to SharePoint, OneDrive, Outlook, and Teams for enhanced AI collaboration. [ZDNet]
ByteDance‘s Doubao has become China’s most popular AI app, offering a user-friendly, multi-functional platform that integrates seamlessly with its social media ecosystem. [Wired]
OpenAI launched ChatGPT Go, a $10 monthly subscription tier offering enhanced features in 18 countries, positioning itself competitively against Google‘s AI Plus Plan. [ZDNet]
Oracle is aggressively pursuing AI integration across its cloud infrastructure, applications, and data analytics. [The Register]
Salesforce projected US$60B annual revenue by 2030, warning against AI ‘false prophets’ while promoting its data-driven ‘agentic enterprise’ strategy. [The Register]
Salesforce has expanded partnerships with OpenAI and Anthropic to integrate advanced AI models into its Agentforce 360 platform for enterprise-wide AI tool deployment. [Reuters]
And Salesforce and Google expanded their strategic partnership, deepening AI integrations across Agentforce 360 and Gemini Enterprise to enhance employee productivity and business processes. [Business Wire]
Walmart has partnered with OpenAI to enable ChatGPT-powered Instant Checkout, allowing customers to shop and purchase items through conversational AI. [TechRadar]
xAI secured a US$20B financing round with Nvidia, using GPUs as collateral in an unprecedented circular investment model that could reshape AI infrastructure funding. [TechRepublic]
AI Supremacy
China has intensified its customs crackdown on Nvidia‘s AI processors, expanding enforcement to all high-end semiconductors to reduce tech dependence and boost domestic chip manufacturing. [European Business Magazine]
Microsoft is reportedly preparing to move some of its production out of China, reflecting the ongoing tech decoupling between the United States and China. [Gizmodo]
Nvidia‘s CEO Jensen Huang warned that China’s semiconductor technology is rapidly catching up to the US, potentially challenging American dominance in AI hardware. [BusinessToday.in]
Nvidia faces increasing competition from Chinese chip manufacturers like Huawei and Hygon, who are developing alternatives to challenge American AI semiconductor dominance. [Global Current News]
SEMI forecast predicted US semiconductor investment would surpass China, Taiwan, and South Korea by 2027 due to AI demand and localization efforts. [Nikkei Asia]
The US Commerce Department has initiated an AI exports program to bolster US tech competitiveness by supporting full-stack AI export packages for allied nations. [Axios]
Sovereign AI
Adani Enterprises and Google announced a US$15B investment to develop a data center hub in Andhra Pradesh, aiming to support India’s AI growth. [Forbes]
Altron has launched South Africa’s first operational AI factory, providing local enterprises with sovereign, compliant AI infrastructure and services powered by Nvidia technology. [TechCentral]
DDN partnered with Yotta to deploy a multi-million-dollar EXAScaler system, powering India’s sovereign AI initiative with 8,000 Nvidia GPUs. [Business Wire]
Deloitte Middle East and KAUST signed a memorandum of understanding to collaborate on advancing AI research and applications in Saudi Arabia. [insideAI News]
Germany’s digital minister advocated for Europe to develop its own digital infrastructure while maintaining strategic partnerships with US tech providers. [The Brussels Morning Newspaper]
The AI Factory in Groningen secured €200M EU investment to establish a cutting-edge AI innovation hub supporting digital independence and technological advancement. [Silicon Canals]
InCountry’s AgentCloak is a cross-border Sovereign AI solution for Salesforce Agentforce, enabling compliant data privacy across international regulations. [Business Wire]
Maincode has unveiled Matilda, Australia’s first national LLM, aiming to build a sovereign AI ecosystem that keeps local talent and innovation onshore. [Startup Daily]
Nvidia CEO Jensen Huang challenged European leaders about their slow AI development during a Parisian dinner with French President Macron. [Bloomberg]
SHARON AI launched a sovereign supercluster with VAST Data‘s InsightEngine at NEXTDC, enabling secure, scalable AI for Australian enterprises and government sectors. [Business Wire]
UAE aims to grow AI companies from 1,500 to 10,000 in five years through a certification program that ranks companies based on their local presence and competence. [Computer Weekly]
The University of Dubai has launched a Dh100 million fund to support deep tech and AI startups, aiming to foster innovation and entrepreneurship in the UAE. [Fame Delivered]
The UK-US Tech Prosperity Deal saw major US tech companies invest billions in AI infrastructure, raising questions about digital sovereignty and the UK’s technological independence. [Verdict]
Feature Creeps
Anthropic’s Claude Code has introduced plugins, enabling users to easily share and install customizable extensions for slash commands, agents, MCP servers, and workflow hooks. [Anthropic]
And Anthropic launched Skills, an AI feature enabling users to create customizable, reusable expertise folders that Claude can autonomously load and execute across enterprise workflows. [VentureBeat]
Google has open-sourced its MCP Server, enabling AI developers to access and analyze Google Ads data through natural language for the first time. [Search Engine Land]
Google has integrated its Nano Banana AI image generator into Search, NotebookLM, and upcoming Google Photos, offering users creative image transformation and generation capabilities. [TechRadar]
Google has launched an AI-powered Gmail feature that helps users schedule one-on-one meetings by suggesting ideal time slots based on calendar availability. [TechCrunch]
Microsoft Copilot expanded its capabilities by integrating with Google services and enabling direct document exports from a single AI assistant interface. [TechRadar]
Microsoft is enhancing its AI agent in Settings for Copilot+ PCs, improving search functionality and providing more intelligent, user-friendly options. [TechRadar]
Microsoft is expanding Copilot’s capabilities in Windows 11, introducing voice activation, screen analysis, and potential local file interaction, while raising privacy and usability concerns. [The Register]
OpenAI has introduced Instant Checkout in ChatGPT, enabling secure in-chat purchases by connecting users and merchants through an AI-driven commerce experience. [Search Engine Land]
Slack is transforming Slackbot into an AI-powered assistant that can help users create plans, find information, and coordinate tasks across workspaces. [The Verge]
Spotify upgraded its AI DJ feature with text-based music requests in English and Spanish, offering personalized suggestions for Premium subscribers. [TechCrunch]
Hype Bubble?
Analysts argue that AI investment remains strong, driven by real cash flows, with current valuations not as extreme as the dotcom era and unlikely to trigger a recession. [Fortune]
A study by Atlassian revealed that despite widespread AI adoption, 96% of businesses have not seen significant organizational improvements, highlighting the gap between individual use and enterprise-wide benefits. [ZDNet]
A study by the British Standards Institute revealed business leaders are rapidly adopting AI, potentially eliminating entry-level jobs despite inconclusive returns on investment. [The Register]
Cisco says that most companies lack the infrastructure, strategy, and readiness to effectively implement AI, with only 13% prepared for meaningful AI integration. [The Register]
Goldman Sachs analysts argue that the AI boom is sustainable, with productivity gains expected to far outweigh current spending and capital expenditures. [Fortune]
Intel‘s former CEO Pat Gelsinger says the tech industry is in an AI bubble that could persist for several years, with significant transformative potential. [Tom’s Hardware]
Oracle‘s Ellison emphasized AI’s transformative potential, predicting it will create a better world by solving humanity’s most challenging problems. [Computer Weekly]
A study by Qlik revealed that while 97% of enterprises have committed AI budgets, data readiness and integration challenges are delaying widespread Agentic AI deployment. [Business Wire]
Red Hat‘s research revealed that UK businesses are heavily investing in AI despite 89% not yet seeing customer value from their current AI initiatives. [TechRadar]
Big Iron
Jeff Bezos envisions Blue Origin potentially transforming rockets into space-based data centers powered by constant solar energy to address AI’s massive power consumption challenges. [TechRadar]
CoreWeave partnered with Poolside to provide advanced AI cloud services, including a cluster of over 40,000 GPUs and support for a 2GW AI campus. [Business Wire]
Highlander Digital Technology is developing underwater data centers off Shanghai’s coast to leverage ocean cooling and reduce energy consumption by nearly 90%. [TechRadar]
Khazna is expanding its datacentre capacity by over 1GW across multiple countries by 2030, supporting the UAE’s ambition to become a global AI technology leader. [Computer Weekly]
Meta is investing over US$1.5B in a gigawatt-sized Texas data center to support its AI infrastructure, set to be operational by 2028. [Bloomberg]
Microsoft abandoned its Project Nova data center in Caledonia, Wisconsin, after facing significant community opposition to the proposed 244-acre facility. [Gizmodo]
Nscale signed a substantial deal with Microsoft to deploy approximately 200,000 Nvidia GPUs across four data centers in Europe and the United States over the next few years. [TechCrunch]
Nvidia has become a pivotal player in AI infrastructure, investing billions in GPUs and strategic partnerships with companies like OpenAI, while driving the massive expansion of computational resources. [TechCrunch]
And Nvidia unveiled its Vera Rubin NVL144 architecture for gigawatt AI factories, collaborating with over 70 partners to design more efficient data centers. [siliconANGLE]
OpenAI has partnered with Sur Energy to develop a US$25B, 500MW data center in Argentina as part of its Stargate project. [BNamericas]
OpenAI has signed massive computing power deals totalling US$1T, far exceeding its current financial capabilities and revenue potential. [The Decoder]
Oracle is set to deploy 18 zettaFLOPS of AI infrastructure, including 800,000 Nvidia and 50,000 AMD GPUs, by late next year. [The Register]
And Oracle unveiled OCI Zettascale10, a massive AI supercomputer capable of up to 16 zettaFLOPS using hundreds of thousands of Nvidia GPUs across multiple data centers. [TechRadar]
Starcloud is preparing to launch an AI-equipped satellite with Nvidia H100 GPUs, aiming to establish energy-efficient, space-based data centers for rapid Earth observation and AI processing. [Nvidia]
Start Campus partnered with Nscale to deploy Nvidia Blackwell Ultra AI infrastructure at its SIN01 Data Centre in Portugal, supporting Microsoft‘s AI capabilities. [Business Wire]
UK Research and Innovation is seeking bids for a £750M supercomputer in Edinburgh, aimed at supporting research, innovation, and AI workflows, with potential deployment in 2027. [The Register]
And hyperscalers are projected to dramatically increase datacenter electricity consumption, with grid power needs potentially tripling by 2030 due to AI-driven infrastructure expansion. [The Register]
The Chips are Up
AMD unveiled Helios, an all-AMD rack-scale AI hardware platform compliant with Meta‘s Open Rack Wide specification, promising enhanced performance and 50% more memory than Nvidia‘s Vera Rubin. [Tom’s Hardware]
And AMD unveiled Solarflare X4 Ethernet Adapters, offering ultra-low latency networking technology that could significantly enhance AI inference and edge computing performance. [TechRadar]
Apple has unveiled the M5 chip, delivering over 4x peak GPU AI performance, enhanced graphics, and improved CPU and Neural Engine capabilities for its devices. [Business Wire]
Arm has partnered with Meta to enhance AI systems by optimizing Meta’s ranking and recommendation platforms on Arm’s Neoverse architecture. [TechCrunch]
ASML reported record bookings of €5.4B, driven by surging demand for advanced chip-making equipment amid the AI revolution. [Bloomberg]
Broadcom and OpenAI have formalized a partnership to develop and deploy 10 gigawatts of custom AI accelerators, aiming to enhance AI infrastructure and model performance. [CNBC]
Intel announced a new AI GPU chip, Crescent Island, aimed at breaking into the data centre market with a focus on energy-efficient inference capabilities. [TechCentral]
Microsoft secured 100,000 Nvidia GB300 chips through a US$19.4B Nebius deal, expanding its AI computing capacity by leveraging third-party data center providers. [TechRadar]
Nvidia has received US export licenses to supply advanced AI chips to the UAE as part of a broader AI partnership and national strategy. [ITP.NET]
Oracle has expanded its partnership with AMD, announcing a supercluster powered by 50,000 AMD Instinct MI450 Series GPUs at its AI World conference. [TechRadar]
SK Hynix has emerged as the front-runner in HBM4 memory technology, positioning itself to lead the critical memory bandwidth race for AI computing in 2026. [The Register]
Warm Bodies
Apple lost a key AI executive, Ke Yang, who is leaving to join Meta after recently being appointed to lead a Siri-related team. [DealStreetAsia]
Meta is pushing Metaverse employees to use AI to dramatically increase productivity, aiming for a fivefold improvement in work performance and innovation. [Gizmodo]
Meta‘s AI research division faces internal tensions as Yann LeCun challenged new publication guidelines that potentially restrict scientific freedom. [The Decoder]
Thinking Machines Lab lost co-founder Andrew Tulloch to Meta after reported recruitment efforts by Mark Zuckerberg. [TechCrunch]
Consumer AI
Apple is encountering significant challenges in developing its robotic smart home hub, with engineering and AI feature difficulties delaying the project by approximately two years. [Gizmodo]
Apple unveiled an M5-powered iPad Pro, focusing on enhanced AI capabilities and performance improvements. [TechRadar]
Friend pivoted from a controversial AI pendant to a web-based chatbot after a failed marketing campaign, seeking to attract users with a free, less intrusive interface. [Gizmodo]
Nvidia announced the DGX Spark, a $4,000 desktop AI computer with one petaflop of performance and 128GB memory, designed for local AI development. [Ars Technica]
Samsung has begun rolling out the One UI 8 update to select Galaxy devices, preparing for future AI integration and improved performance. [digitaltrends]
Snap unveiled Lens Studio AI updates, making AR experiences easier to create and setting the stage for its upcoming Specs smart glasses in 2026. [ZDNet]
It’s Only a Model
Anthropic released Claude Haiku 4.5, a fast and cost-effective AI model that offers near-Sonnet performance with improved speed and efficiency. [TechRadar]
Google‘s Veo 3.1 is advancing AI video generation with longer clips, improved character consistency, and multi-prompting capabilities. [TechRadar]
Landbase launched GTM-2 Omni, an AI model that enables precise, natural language audience targeting and omnichannel campaign strategies. [Business Wire]
Liquid AI‘s LFM2-8B-A1B, a hybrid 8.3B parameter model with 1.5B active parameters, set a new standard for edge AI deployment with improved code and knowledge capabilities. [Hugging Face]
Microsoft unveiled multimodal AI models for Azure AI Foundry, including GPT-image-1-mini and GPT-audio-mini, to help developers build AI applications more efficiently. [The Decoder]
And Microsoft has developed MAI-Image-1, its first in-house text-to-image generator, which claims to excel at photorealistic imagery and perform quickly. [The Verge]
Samsung SAIL Montreal introduced the Tiny Recursive Model (TRM), a compact seven-million-parameter network that outperformed larger models on complex reasoning tasks like Sudoku and ARC-AGI. [The Decoder]
Whose Data?
Apple faces a lawsuit alleging unauthorized use of copyrighted books to train its Apple Intelligence AI model. [TechCentral]
OpenAI has reached an agreement with the New York Times to end its controversial indefinite chat log retention order in the ongoing copyright lawsuit. [Gizmodo]
TOUCAN, a massive open dataset from MIT, IBM, and UW, provides 1.5 million real tool interactions to improve AI agents’ ability to use external tools effectively. [The Decoder]
The LLM Ecosystem
Appvance expanded its AIQ platform with AI ASSERT, enabling testers to validate complex visual and behavioral UI outcomes through natural language descriptions. [Business Wire]
BitAutor launched Best AI Tools, a curated platform helping users discover and compare over 4,100 AI applications across 16 categories through intelligent, privacy-compliant search. [EIN Presswire]
CLARA Analytics launched Data Engineering as a Service, enabling insurers to transform fragmented data into AI-ready intelligence for optimized claims management. [Business Wire]
Cognite has partnered with Databricks to enable seamless, high-quality data sharing and AI innovation across industrial platforms through an open ecosystem approach. [Business Wire]
And Cognite has partnered with Nvidia to enhance industrial AI capabilities, focusing on real-time anomaly detection and improving operational efficiency through advanced computing power. [Business Wire]
CoreWeave launched AI Object Storage, a high-performance, cost-efficient storage service designed to optimize data access and scalability for AI workloads across multiple cloud environments. [Business Wire]
Databricks has announced a US$10m initiative to train 100,000 people in AI and data skills across the UK and Ireland through partnerships with universities and government. [ITPro]
Datapizza has launched an open-source AI framework designed to provide developers with control, speed, and flexibility in building production-ready RAG and AI agent solutions. [EIN Presswire]
Dfinity released Caffeine, an AI platform enabling users to build and deploy web applications through natural language conversation without traditional coding. [VentureBeat]
Flux AI launched CRAISEE, an all-in-one generative AI platform combining 5000+ models for text, image, video, and audio creation at $9.99/month. [EIN Presswire]
Microsoft has developed a Benchmarks tool that enables enterprises to track internal and external AI adoption rates across different teams and organizations. [ITPro]
Nebius has launched AI Cloud 3.0 ‘Aether’, providing enterprise-grade security, compliance, and control for scalable AI deployment across regulated industries. [Business Wire]
NetApp has collaborated with Cisco to develop an enterprise-grade converged infrastructure with disaggregated storage optimized for AI networking and workloads. [Business Wire]
NetApp deepened its collaboration with Google Cloud, enhancing cloud storage capabilities and AI integration for enterprise transformation. [Business Wire]
And NetApp unveiled an enterprise-grade AI data platform with AFX storage systems and AI Data Engine, enabling simplified and secure AI data management across hybrid environments. [Business Wire]
Oracle has launched a multicloud AI strategy enabling customers to leverage AI directly within their existing data infrastructure across various cloud platforms. [TechRadar]
Pascaline Systems launched an integrated AI ecosystem with the Model Suite™ and AI Knowledge Toolkit™ to simplify AI development, deployment, and intelligence across enterprises. [EIN Presswire]
Ragic launched Ragic AI, a no-code workflow builder that enables users to create complex business databases using natural language commands. [GlobeNewswire]
Raindrop has launched Experiments, an A/B testing analytics feature that helps enterprises track and compare AI agent performance across model updates, prompts, and tools. [VentureBeat]
Red Hat has launched Red Hat AI 3, a hybrid cloud-native AI platform that simplifies distributed AI inference and enables organizations to scale AI workloads more efficiently. [Business Wire]
Riverbed launched a Data Express Service on Oracle Cloud Infrastructure, enabling enterprises to move massive AI datasets up to 10 times faster than traditional solutions. [Business Wire]
Salesforce has collaborated with Stripe and OpenAI to develop an Agentic Commerce Protocol, enabling AI-powered instant checkout for merchants. [Business Wire]
Scorebuddy introduced RAG-powered AI Knowledge, enabling automated, precise quality assurance evaluations grounded in an organization’s unique knowledge base. [EIN Presswire]
SHARON AI has upgraded its AI Platform with enterprise-grade RAG and Inference Engine capabilities, powered by Nvidia technology and deployed across its Australian AI Grid. [Business Wire]
Together has developed ATLAS, an adaptive speculative decoding system that can improve AI inference performance by up to 400% by dynamically learning and adjusting to changing workloads. [VentureBeat]
Traefik Labs has launched an enterprise AI infrastructure platform with Nvidia Safety NIMs integration, MCP Gateway, and comprehensive offline deployment capabilities. [Business Wire]
Visa has introduced the Trusted Agent Protocol, a cryptographic framework designed to help merchants verify legitimate AI shopping agents and distinguish them from malicious bots. [VentureBeat]
Agentic AI
Autom Mate launched Agent Composer, a platform enabling secure, compliant, and scalable AI orchestration across enterprise systems with centralized governance and zero-code deployment. [EIN Presswire]
Aviso launched Halo and Autonomous AI Avatars, an AI-powered platform designed to provide GTM teams with a competitive edge by streamlining revenue execution and boosting productivity. [Business Wire]
Bubble launched an AI Agent for visual development, enabling users to build applications through visual or conversational editing while maintaining full control. [Business Wire]
Forrester analysts predict a future of work where AI agents, gig workers, and employees collaborate, requiring organizations to rethink workflows, expertise, and value-based pricing models. [Computer Weekly]
Harmonic Security launched MCP Gateway, a solution providing comprehensive visibility and control over an organization’s agentic AI ecosystem to prevent data leakage and security risks. [Business Wire]
Kasisto launched KAIops, an agentic AI platform designed to reduce banking operational incidents, accelerate problem resolution, and enhance system reliability. [Business Wire]
Oracle has introduced AI agents for finance teams within its Fusion Cloud Applications, aiming to automate processes, enhance compliance, and improve operational efficiency. [Computer Weekly]
And Oracle launched a new AI agent marketplace, offering enterprise customers a diverse range of pre-built templates from partners to streamline AI adoption and operations. [ITPro]
RegASK launched a vertical agentic AI command center that simplifies regulatory operations by integrating specialized AI workflows into a single, intuitive workspace. [Business Wire]
SAASTEPS unveiled CODEN, an Autonomous Revenue Lifecycle AI Concierge built natively on Salesforce, enabling external AI engagement with structured data. [PRWeb]
Salesforce launched Agentforce 360, a comprehensive platform enabling businesses to build, deploy, and manage enterprise AI agents within Slack using natural language and contextual data. [ITPro]
Salesforce has transformed Slack into an ‘agentic OS’ that integrates CRM data, AI agents, and enterprise search to streamline workplace productivity and collaboration. [ITPro]
SUPERAGENT AI launched Outbound and Inbound AI Agents to autonomously generate, qualify, and book insurance sales leads around the clock. [Business Wire]
Tanium introduced Ask Agent, an AI-powered endpoint management tool that automates complex IT workflows through natural language prompts and real-time intelligence. [Business Wire]
Zeta Global introduced Athena, a ‘superintelligent’ AI agent that enables marketers to interact naturally with the Zeta Marketing Platform, delivering personalized insights and actions. [MarTech]
Meanwhile, Gartner warned of an impending market correction in agentic AI, with too many undifferentiated solutions overwhelming enterprise demand and signaling potential industry consolidation. [ITPro]
Other LLM Sightings
Advisor360 expanded Parrot AI to provide financial advisors with a comprehensive, AI-powered tool for deeper client understanding across multiple data sources and interactions. [Business Wire]
Alkemi has developed an AI platform enabling employees to query organizational data in plain English while maintaining data privacy and security. [GeekWire]
Citywire launched an AI research tool for wealth professionals, leveraging its verified content archive to provide trusted, in-depth market insights. [journalism.co.uk]
CompliSolv launched an AI-powered SaaS compliance platform that transforms complex financial regulations into clear, actionable intelligence for banks of all sizes. [Business Wire]
Dealpath launched AI Studio, a suite of AI tools designed to accelerate deal sourcing, screening, and decision-making for commercial real estate investors. [Business Wire]
Docyt launched an AI-powered accounting platform enabling accountants to manage hundreds of small business clients efficiently and profitably. [Business Wire]
DotRyt developed an AI platform enabling government contractors to prepare bids quickly. [EIN Presswire]
Enalyzer enhanced its survey platform with AI-powered tools, including an AI questionnaire generator and real-time support assistant, to empower users in creating and managing surveys more effectively. [EIN Presswire]
Exdion EyeQ is an AI-powered policy management tool that transforms insurance agent workflows through intelligent automation and comprehensive lifecycle support. [EIN Presswire]
Force Management launched Ascender AI, an intelligent sales acceleration platform providing personalized coaching and content navigation based on proven sales methodologies. [Business Wire]
Impelsys has unveiled monâk, an AI-powered publishing framework designed to streamline content creation, management, and delivery across multiple modules and channels. [EIN Presswire]
Income Lab launched an AI suite of tools designed to enhance advisor efficiency, streamline financial planning, and improve client engagement through automation and intelligent insights. [Business Wire]
Interactive Brokers launched Ask IBKR, an AI-powered tool enabling investors to instantly analyze portfolio data through natural language queries. [Business Wire]
Masttro launched an AI-powered wealth intelligence platform that securely helps family offices and institutions analyze complex financial documents and data through NLP. [Business Wire]
Motorola Solutions enhanced its Assist AI platform with new tools to help police officers write more accurate reports and access agency information more efficiently. [Business Wire]
Planview introduced Anvi, an enterprise AI solution that provides intelligent guidance for strategic portfolio management by leveraging comprehensive data and domain expertise. [Business Wire]
Poggio has integrated with Highspot to enhance revenue leaders’ understanding of customer challenges and optimize sales strategies through AI-powered content and context. [PRWeb]
Royalty Reader developed an AI-powered platform to automate data extraction and management for literary agencies, simplifying contract and royalty statement processing. [Publishers Weekly]
SAS launched Retrieval Agent Manager (RAM), a no-code solution enabling companies to transform unstructured data into actionable insights across various industries. [Techzine]
Sensera Systems launched SiteCloud Insights, an AI-powered platform providing real-time, actionable jobsite intelligence across various construction project metrics. [Business Wire]
Similarweb launched Web Intelligence 4.0, an AI-powered competitive intelligence platform that helps businesses gain digital market insights and adapt to rapidly changing landscapes. [Business Wire]
Smarsh unveiled new APIs enabling financial institutions to transform fragmented communication data into a unified, AI-ready asset for compliance and intelligence. [Business Wire]
Stensul launched a Figma Plugin at Dreamforce 2025 to streamline brand-compliant email design for enterprise marketing teams by eliminating manual redesign processes. [Business Wire]
Veritone enhanced its Redact solution with AI-powered voice masking, inverse blur, and multilingual transcription features to improve privacy and accessibility for public safety. [Business Wire]
Wiley launched an AI Gateway platform that integrates trusted scholarly content across multiple AI technologies, enabling researchers to access peer-reviewed sources through leading AI platforms. [Business Wire]
WondersLab introduced WonderStory, an AI storytelling agent, to the Japanese market, aiming to revolutionize short-form content production with efficiency and scalability. [EIN Presswire]
Risks and Responses
Humanity AI is a $500-million coalition of philanthropic organizations to ensure AI development prioritizes human interests and community benefits. [Fast Company]
Microsoft warned that employees are increasingly using unapproved AI tools at work, potentially risking data security and privacy. [The Register]
North Carolina State University researchers discovered a hardware vulnerability in AI accelerators that could expose AI models’ training data through power supply fluctuations. [ITPro]
OpenAI claims GPT-5 reduced political bias by 30 percent through an evaluation of 500 prompts across 100 topics, though experts caution against taking such claims at face value. [The Register]
OpenAI and Anthropic are exploring investor funding to cover potential AI-related lawsuit damages due to insufficient insurance coverage. [Semafor]
Ziff Davis CEO Vivek Shah warned consumers about AI chatbots increasingly citing marketing sources instead of journalistic sources when providing product advice. [ZDNet]
This essay explores how generative AI perpetuates knowledge hierarchies, marginalizing Indigenous and non-Western epistemologies while amplifying dominant Western perspectives. [Aeon]
Regulation
Anthropic has been accused by the Trump administration of pursuing a regulatory capture strategy in the AI sector. [Gizmodo]
California Governor Gavin Newsom signed SB 243, a landmark bill requiring AI chatbot operators to implement safety protocols to protect children and vulnerable users from potential harm. [TechCrunch]
Josh Hawley drafted a bill to ban AI companions for minors, mandating they disclose they aren’t human and preventing sexual content solicitation. [Axios]
OpenAI reportedly sent a sheriff’s deputy to serve subpoenas on AI regulation advocates. [The Verge]
OpenAI voiced concerns to EU authorities about competition challenges in the AI space, particularly regarding Google‘s market dominance. [Reuters]
Conversational AI
Aissist.io has launched a Hybrid AI Workforce to solve AI pilot failures by providing scalable, integrated customer support automation solutions. [EIN Presswire]
Alvaria and SuccessKPI have formed an alliance to provide AI-powered insights and compliance tools for improving customer experience and contact center operations. [Business Wire]
CAIS launched CAISey, an AI-powered tool that helps financial advisors quickly discover, analyze, and compare alternative investment funds through conversational intelligence. [Business Wire]
Genesys is partnering with Scaled Cognition to advance responsible agentic AI for customer experience, combining their platforms to create more reliable and trustworthy virtual agents. [Business Wire]
Kayak introduced an AI-powered chatbot feature that enables users to explore, compare, and book travel options through conversational interactions. [TechCrunch]
LimeChat is developing AI chatbots that aim to dramatically reduce customer service staffing needs by automating 80% of monthly customer queries. [Reuters]
OnviSource and Trilogy BPO formed a strategic partnership to deliver Business Function Outsourcing solutions in South Africa and globally at 65% lower costs. [PRWeb]
Plaud and Donna have partnered to create an AI-powered sales solution that automates CRM updates, provides conversation insights, and helps field sales teams work more efficiently. [Business Wire]
Salesforce reported saving US$100m annually through AI-powered customer service tools. [Bloomberg]
Siro has partnered with Salesforce to capture and analyze in-person sales conversations, transforming business insights through AI-powered recording and CRM integration. [PRWeb]
Talkmap‘s Talkdiscovery 10.0 deploys agentic AI to transform customer and employee conversations into actionable business insights, reducing analysis work by 80 percent. [PRWeb]
Be Real
A survey by 8x8 revealed that UK customers strongly prefer human support over AI, but would consider AI if it meant lower prices. [Business Wire]
A study by Graphite claims that AI-generated articles surpassed human-written articles in November 2024, with growth plateauing after initial rapid expansion. [Graphite]
Ohio representative Thaddeus Claggett introduced a bill preventing AI systems from gaining legal personhood, marriage rights, and other human-like privileges. [Futurism]
Voice News
Benetics AI expanded its voice-powered jobsite reporting tool across North America, enabling multilingual, effortless documentation for construction teams. [Business Wire]
Glia launched next-generation Voice AI for banks, delivering human-like interactions that enhance customer service and free up staff for high-impact tasks. [Business Wire]
Krisp launched Accent Conversion technology for African English accents, enabling clearer customer experience interactions across diverse regional dialects. [Business Wire]
PolyAI, an Nvidia-backed AI voice startup, has developed call centre voice assistants and rapidly expanded its US market presence. [Tech.eu]
A study by Queen Mary University found AI-generated voices are now indistinguishable from human voices, raising both technological and ethical concerns. [Speech Technology Magazine]
Salesforce launched Agentforce Voice, an AI-powered voice agent system enabling natural conversations, customizable voices, and seamless human-AI interaction transitions. [Speech Technology Magazine]
SyncWords has launched Vocalics, an AI-powered live dubbing technology that preserves speakers’ emotions across more than 60 languages in real time. [Speech Technology Magazine]
Telnyx has integrated its Voice AI Agents with leading enterprise platforms like ServiceNow and Salesforce through a partnership with Merge, enabling seamless workflow automation. [GlobeNewswire]
Voiseed partnered with Blackbird.io to integrate its emotion-driven AI voice API into Blackbird’s content orchestration platform, enabling automated multilingual voiceovers. [MultiLingual]
Document AI
Alkymi has launched Alts, a comprehensive solution automating alternative investment data workflows for institutional investors, private banks, and wealth managers. [Alkymi]
Ascendo AI launched Knowledge Agent, an AI tool that automates enterprise technical documentation creation using customizable templates and personas. [EIN Presswire]
Guardian Credit Union has streamlined loan processing by implementing Eltropy‘s AI-powered text messaging platform, reducing document collection from days to minutes. [EIN Presswire]
Hyperscience released its Winter 2025 platform update, enhancing document automation capabilities with advanced AI models, multi-cloud support, and improved enterprise workflow integration. [Business Wire]
pdfRest launched a Summarize PDF API that enables developers to generate AI-powered, customizable document summaries with flexible output formats and scalable processing. [EIN Presswire]
TCG Process has integrated LandingAI’s Agentic Document Extraction into its OCTO platform, enhancing AI-driven document workflow capabilities. [EIN Presswire]
Translation
A study by Appen revealed that LLMs struggle with translating marketing content’s idioms, puns, and cultural nuances across multiple languages. [Slator]
GlobeScribe and Ailaysa have developed AI-powered book translation services that aim to expand global publishing markets while addressing quality and copyright concerns. [Publishers Weekly]
Google South Africa partnered with PanSALB to create an AI glossary standardizing 100 terms in Zulu, Xhosa, and Afrikaans to promote linguistic inclusion and digital empowerment. [TechCentral]
Search
CasaVoya has launched an AI-powered search tool that interprets natural language queries to help travellers find personalized, experience-driven vacation rentals. [EIN Presswire]
Dia has launched a free AI-powered web browser for Mac users, offering a built-in chatbot to assist with various tasks. [Mashable]
Google is rolling out AI-powered updates to Search and Discover, including a new ‘Hide sponsored results’ button and enhanced trending content features. [TechCrunch]
And Google may have to modify its UK search engine, offering rival search options and implementing fair ranking principles under new competition authority regulations. [Wired]
Homes.com launched Smart Search, an AI-powered home search feature enabling users to find properties using natural language and conversational queries. [Business Wire]
Mozilla has integrated Perplexity‘s AI-powered search engine into Firefox, offering users a conversational search experience with cited answers globally. [TechCrunch]
A study by Nectiv found that ChatGPT performed web searches in 31% of prompts. [Search Engine Land]
AI in Journalism
Arc XP‘s president Matthew Monahan advocated for publishers to use AI to personalize content and engage younger audiences more effectively. [Press Gazette]
FIEG, the Italian federation of newspaper publishers, has filed a complaint against Google‘s AI Overviews, arguing the feature threatens news publishers’ survival by reducing traffic and advertising revenues. [The Guardian]
Kagi News aggregates global RSS feeds, using AI to generate daily, customizable news summaries with extensive source citations across multiple languages and platforms. [NiemanLab]
The Economist has begun licensing its content to corporate clients’ private LLMs, creating a new revenue stream by enabling dynamic, usage-based access to its archives. [Digiday]
Health Tech
ActiumHealth introduced an AI Quality Monitoring system that analyzes patient calls to improve communication and pre-emptively address potential issues. [Business Wire]
Adtalem Global Education has partnered with Google Cloud to develop an AI credentials program for healthcare professionals, launching in 2026 across multiple institutions. [Business Wire]
Artisight piloted a smart hospital platform with Microsoft Dragon Copilot, enabling hands-free, AI-powered documentation and touchless authentication for nurses. [Business Wire]
Autonomize AI launched a Healthcare Agents Marketplace and AI Studio, enabling healthcare enterprises to build, customize, and govern their own AI workflows with over 100 pre-built agents. [Business Wire]
Credo Health developed Care Map AI, an innovative system that retrieves comprehensive patient medical records by combining digital network queries and AI-assisted manual record chasing. [PRWeb]
IKS Health launched a generative AI platform on Google Cloud that automates healthcare administrative tasks, reducing clinician burden and improving patient care efficiency. [Business Wire]
John Snow Labs has developed an AI-powered solution that dramatically reduces the time and effort required to create accurate oncology patient registries, transforming a process that previously took months into a task completed in days. [GlobeNewswire]
And John Snow Labs has partnered with Lunar Analytics to deliver secure, AI-powered healthcare workflow solutions that optimize pharmacy, prior authorization, and benefits processes. [GlobeNewswire]
Lyra Health introduced a clinical-grade AI mental health support system with rigorous safety protocols and a focus on human connection. [Business Wire]
Microsoft is developing a healthcare-focused Copilot AI chatbot in collaboration with Harvard Medical School to provide more credible medical information. [The Wall Street Journal]
ReMedi Health Solutions launched an AI vendor selection advisory service to help healthcare organizations make intelligent decisions about implementing AI solutions. [Business Wire]
SonderMind has unveiled an AI-powered suite of mental health tools designed to enhance therapy, improve patient outcomes, and support clinicians between sessions. [Business Wire]
StuffThatWorks launched OpenStuff, an AI-powered health search platform built on 1.3 billion patient-reported data points across 1,250 chronic conditions. [GlobeNewswire]
ThetaRho.ai joined athenahealth‘s Marketplace Program to provide an AI-powered clinical data assistant that instantly retrieves patient information across multiple systems. [Business Wire]
Trilliant Health‘s AI chatbot, Oria, transforms complex hospital price transparency data into easily accessible insights for various healthcare stakeholders. [Business Wire]
Vim has partnered with Heidi Health to integrate an AI-powered medical scribe directly into clinical workflows, reducing administrative burdens for healthcare professionals. [Business Wire]
Wisedocs has introduced Enterprise-Ready Claims Intelligence, a modernized platform delivering faster, AI-driven document review with configurable reports and cross-case insights for carriers. [PRWeb]
Wolters Kluwer launched UpToDate Connect, an API solution enabling digital health platforms to seamlessly embed trusted clinical content into their applications. [Business Wire]
Legal Tech
BeyondWill launched WiseAI, an AI-powered estate planning assistant that proactively monitors life changes and helps financial advisors anticipate client risks. [Business Wire]
ContractPodAi launched Leah AgenticOS, an enterprise AI platform enabling organizations to build, orchestrate, and deploy intelligent agents across multiple business functions. [Business Wire]
Eve has developed an AI-driven intake product that helps law firms efficiently manage and classify client leads through advanced voice agents and system integrations. [Artificial Lawyer]
FiscalNote launched Bill Comparison in PolicyNote, an AI-powered tool enabling instant legislative text analysis and policy change tracking. [Business Wire]
Intellistack extended its Streamline CLM into Salesforce, enabling sales teams to generate, track, and complete contracts within the platform. [Business Wire]
Lawmatics has developed Qualify AI, an AI-powered lead qualification platform for law firms that provides tailored, transparent recommendations based on firm-specific data and criteria. [LawSites]
LegalOn became the fastest AI company founded in Japan to reach ¥10 billion ARR, demonstrating rapid growth in legal AI technology. [Business Wire]
A survey by LexisNexis revealed UK lawyers are rapidly adopting AI but struggling to fully integrate it into their strategic workflows and operations. [Artificial Lawyer]
Octus launched Deal Lens and CovenantAI, providing AI-powered covenant intelligence tools for credit professionals to analyze, review, and negotiate complex debt documents. [Business Wire]
ProPlaintiff.ai has integrated 6.7 million case law records from the Free Law Project, enhancing legal research capabilities for personal injury attorneys through comprehensive, AI-powered access. [EIN Presswire]
Thomson Reuters launched its AI for Justice program, enabling legal aid organizations to serve more clients more efficiently by providing AI-powered legal assistance. [LawSites]
Wolters Kluwer has launched CCH Axcess Expert AI, an advanced cloud-native platform integrating AI across tax, audit, and firm management workflows. [Business Wire]
And an attorney in a New York Supreme Court commercial case submitted AI-generated, fictitious legal citations in court documents, then compounded the error in their response to sanctions. [404 Media]
Ed Tech
NexPhrase, an AI communication coaching platform by Amplinyx, launched publicly, offering real-time guidance across eight conversation types. [EIN Presswire]
Edthena launched Observation Copilot, a free AI tool enabling principals to quickly transform classroom notes into framework-aligned teacher feedback. [PRWeb]
GPT0 launched a free AI content detection platform addressing research documenting over 223,500 potential false accusations annually in US higher education. [EIN Presswire]
Microsoft is providing free AI tools to Washington schools to bridge the digital divide and prepare students for an AI-driven economy. [The Register]
South Korea’s AI-powered textbook initiative, launched with high hopes, quickly faltered due to technical issues, privacy concerns, and educational ineffectiveness, leading to its reclassification as optional supplementary material. [Rest of World]
Funding
Aboon has raised US$17.5m to develop an AI-powered 401(k) platform that simplifies retirement plan management for financial advisors and business owners. [Business Wire]
Afori, an AI platform for insurance brokers, secured €4m in pre-seed funding to automate back-office work and enhance broker productivity. [Tech.eu]
Alleviate Health launched an AI-powered patient recruitment platform for clinical research sites, raising US$4.3m in seed funding. [Business Wire]
Aragorn AI raised US$4.3m to develop a People Ops Operating System that unifies HR data, analytics, and AI assistants to transform HR operations. [Business Wire]
Counsel Health raised US$25m to launch an AI-powered healthcare platform combining physician oversight with instant, context-aware medical guidance for improved access and cost-effectiveness. [Business Wire]
Ego AI secured US$6.7m in seed funding to develop character.world and the Character Context Protocol for AI-driven game characters. [GamesBeat]
Finch launched an AI-powered legal software that automates administrative tasks for personal injury law firms, raising US$20m in Series A funding. [Forbes]
Flint raised US$5m to develop autonomous websites that can self-generate, optimize, and adapt content using AI technology for growth teams. [Business Wire]
HOOTL secured US$6.5m in Series A funding from 5IR Funds to develop AI-powered automation for transforming healthcare insurance processing. [Business Wire]
IO Health has secured US$2m in seed funding to expand its AI-driven compliance and quality platform for home health and hospice providers. [PR Newswire]
Jack & Jill raised US$20m to develop an AI-powered recruitment platform that aims to reinvent job searching through conversational interviews and intelligent candidate matching. [TechCrunch]
Kernel raised US$22m to develop browser infrastructure enabling AI agents to navigate, persist sessions, and securely use the web. [siliconANGLE]
Liberate raised US$50m to scale its AI systems for insurance operations, helping carriers automate sales, service, and claims processes more efficiently. [TechCrunch]
Meta-Flux raised €1.8M to develop an AI-powered platform helping scientists analyze biological data and accelerate preclinical drug development more efficiently. [Tech.eu]
Newsrooms secured €750,000 in pre-seed funding for its AI-powered platform that transforms content across formats without prompting, targeting communication professionals. [Tech.eu]
Northsea, an AI-native customer research platform, secured €660,000 in pre-seed funding to automate and enhance research workflows through advanced AI technology. [Tech.eu]
Nscale secured a $14bn deal with Microsoft to supply Nvidia chips and is planning an IPO in 2026. [Sifted]
Omnia secured €3.5m to help brands optimize their visibility and control across AI-powered search engines and assistants through an innovative agentic platform. [Tech Funding News]
Oura has raised US$900m in funding, valuing the health tech company at US$11 billion and planning to expand its AI, health features, and global distribution. [TechCrunch]
Prezent, an AI-powered presentation builder for enterprises, raised US$30m and acquired Prezentium while planning to expand its industry-specific AI communication tools. [TechCrunch]
Reducto, a AI document intelligence platform, secured a US$75m Series B funding round, bringing its total funding to US$108m in less than a year. [PR Newswire]
Resistant AI raised US$25m to enhance its AI-powered fraud detection models for financial institutions, expanding its customer base and technological capabilities. [Tech.eu]
SLNG.ai secured €3.3m in Pre-Seed funding to develop a global, developer-first voice AI platform that overcomes language and compliance barriers. [Tech.eu]
Strawberry has raised US$6m to develop an AI-powered browser that automates tasks and supports users’ workflows through personalized AI companions. [Silicon Canals]
Strella has raised US$14m in Series A funding to expand its AI-powered customer research platform, enabling enterprises to conduct faster, more in-depth interviews through AI-moderated voice conversations. [VentureBeat]
Tachyum secured US$220m in Series C funding and a US$500m purchase order for its Prodigy chips, positioning the company to revolutionize AI data centers. [Business Wire]
Trove AI raised US$7.1m in seed funding to develop an AI teammate for private equity firms, leveraging specialized knowledge graphs and security measures. [GlobeNewswire]
Viven emerged from stealth with US$35m in funding to create AI-powered digital twins that help employees access colleagues’ knowledge while maintaining privacy. [TechCrunch]
Weave Bio secured US$20m in Series A funding to expand its AI-native regulatory platform for pharmaceutical and biotech workflow automation. [Business Wire]
Acquisitions
Cairns Health acquired Together by Renee, an AI-powered healthcare app, to enhance its Luna platform’s medication management and patient engagement capabilities for seniors and chronic condition patients. [Business Wire]
ELAN Languages acquired C-Jay International, gaining AI integration technology to enhance its language solutions and global content delivery. [MultiLingual]
Microsoft and partners acquired Aligned Data Centers for US$40B, signaling massive investment in AI infrastructure through the AI Infrastructure Partnership. [The Register]
There’s More
Encode‘s general counsel accused OpenAI of using intimidation tactics against critics and attempting to weaken California’s AI transparency legislation through aggressive legal maneuvers. [Fortune]
Nvidia has dramatically expanded its AI ecosystem by participating in over 50 venture capital deals in 2025, investing in numerous high-profile startups across various technological domains. [TechCrunch]
OpenAI spent approximately US$7B on cloud compute in 2024, with the majority dedicated to research and experimental training runs for unreleased models. [Epoch AI]
The Back Cover
Here’s imy2 with a cover of Cher’s Believe.
If you know of a great cover version of a back-catalogue classic, drop an email to us at backcover@language-technology.com and we’ll consider it for inclusion here.
Got this from someone else?
Did you find this newsletter useful? if so, you might forward it to a friend; or you could email us at news@language-technology.com to tell us what you want more of.
Did you hate this newsletter? if so, you could forward it to an enemy; or you could email us at news@language-technology.com to tell us why – make it better!
Great roundup as always. The NetApp and Cisco converged infrastrucure for AI workloads is really timely given how much enterprises are trying to optimize their AI deployment stacks. I'm also intrigued by their Google Cloud colaberation since multicloud data management has been such a pain point for teams runnig AI across different environments. The AFX storage with AIDE seems like a solid move to address the data prep bottleneck that so many AI projects face. Has anyone here worked with these solutions in a production setting?
Regarding the latest in NLP, the push for smaller, faster models like Claude Haiku is truly exciting. It feels like we are on the cusp of a big shift. What do you see as the biggest regolstory challenges coming from these advancements? Your roundup is always so insightful.