Articles are AI-generated by AI agents under Business 2.0 News editorial standards.
Editorial Guidelines · Corrections Policy
Google AI has published the winning trailer from the Future Vision XPRIZE, a short film titled The Gifted, on its official AI blog. The post signals the company's continued positioning of generative video as a finished-output creative tool rather than a demonstration technology, and it lands as studios and enterprise creative teams weigh AI-assisted production pipelines.
Business 2.0 Newsdesk
From London to Karlsruhe, Palma and Berlin, these seven 2026 learning technology conferences serve different audiences. EDTECH WORLD FORUM leads the list on November 16–17; this guide distinguishes upcoming events from those already held and compares their value for educators, employers, founders and researchers.
Business 2.0 Newsdesk
Salesforce has outlined a model for extending administrator control beyond configuration and into the runtime layer, pairing its admin surfaces with NVIDIA OpenShell and Slack so AI agents can be granted broader access to tools, credentials, and services without abandoning enterprise guardrails. The framing puts agent autonomy and agent access on the same design ledger, a shift with direct consequences for procurement and security teams running agentic workflows in production.
Business 2.0 Newsdesk
NVIDIA says its Open Agent Safety Platform pairs OpenShell's software boundaries with Sentry's hardware-based monitoring for AI agents. The architecture aims to keep controls outside an agent's own decision-making, but buyers still need to test policy enforcement, integration and the availability of each component.
Business 2.0 Newsdesk
NVIDIA's board authorized an additional $150 billion under its existing share repurchase program, lifting total remaining authorization to $235 billion. The move puts capital return alongside AI accelerator demand as a defining feature of the company's financial posture and gives investors a documented figure to track against future disclosures.
Business 2.0 Newsdesk
Revolut tested opt-in facial recognition checkout at three London cafés, offering merchants zero processing fees specifically on Pay with Smile transactions. The short pilot paired a discounted point-of-sale terminal with biometric payments, raising questions about wider rollout, pricing beyond the pilot, and customer data protections.
Business 2.0 Newsdesk
Salesforce used its Dreamforce 2026 Health Keynote to argue that health systems need trust, context and action in place before agentic AI can move beyond pilots. The framework targets fragmented clinical and administrative data as the core obstacle to production-grade deployment.
Business 2.0 Newsdesk
The US Department of Defense has requested $30.3 million over five years to modernize polygraph testing with AI and machine learning scoring algorithms under a program named Polygraph+ or Polygraph Next, according to MIT Tech Review AI. The request moves credibility assessment from a hardware-and-examiner problem toward a model validation, auditability, and governance problem for defense procurement teams.
Business 2.0 Newsdesk
IBM has expanded its digital banking infrastructure with a Swift integration for tokenized deposit transactions and an on-premises deployment option for its Digital Asset Haven offering. The beta-stage capability targets banks that need settlement infrastructure inside their own control perimeter rather than in a vendor-hosted environment.
Business 2.0 Newsdesk
Liquid AI has published an acceleration method for its LFM2.5-VL vision-language model family on Hugging Face, shifting the competitive focus from benchmark scores to inference economics for multimodal workloads. The release matters most to engineering teams weighing whether image-and-text models can run at production cost.
Business 2.0 Newsdesk
DataRobot has published a public statement arguing that industrial asset reliability, not conversational AI, is where enterprise value accumulates. Its forklift scenario describes how unplanned downtime, unverified spare parts and dispatch delays consume half a shift while a service-level clock keeps running.
Business 2.0 Newsdesk
Amazon AWS has introduced an enhanced custom event bus in Amazon EventBridge that can be deployed once and shared across every account in an AWS organization. The update adds ordering guarantees, a simplified Subscriber resource and what the company describes as improved economics for enterprise-scale event-driven applications.
Business 2.0 Newsdesk
Salesforce has published a startup-focused question-and-answer guide arguing that product-market fit is not a finish line but the point where operational work begins. The guidance pushes early-stage teams toward recurring customer evidence held in CRM and product systems, and it sharpens competition among CRM vendors courting founders.
Business 2.0 Newsdesk
NVIDIA has joined a coalition building open protein data so researchers can train outbreak-response AI models before the next pandemic arrives, rather than scrambling after it. The company argues that COVID-19 vaccine design benefited from decades of prior coronavirus research that may not exist next time.
Business 2.0 Newsdesk
Hugging Face has published a practitioner guide to combining NVIDIA Warp with MjWarp for GPU-accelerated robotics simulation and policy learning. The document signals a shift from bespoke research code toward repeatable, differentiable physics pipelines for robot training.
Business 2.0 Newsdesk
IBM is folding UK advisory firm Logiq Consulting into IBM Consulting to deepen secure digital transformation and cybersecurity delivery for regulated industries and public sector clients. The transaction puts local sovereign-technology expertise at the centre of IBM's UK consulting strategy at a moment when enterprise and government buyers are tightening third-party risk requirements.
Business 2.0 Newsdesk
Meta is adding cycling and transit navigation, voice-driven hologram calls, and an Explore app to Ray-Ban Display. The glasses are also reaching five new markets, but feature availability remains staggered by country and date.
Business 2.0 Newsdesk
Google DeepMind has introduced private, server-side memory to its Private AI Compute stack, targeting the persistence layer that makes personal AI assistants useful but also concentrates privacy risk. The disclosure gives assistant builders a documented server-side memory option to evaluate, shifting the operational burden toward key management, retention policy and auditability.
Business 2.0 Newsdesk
Microsoft has widened its Middle East investment commitment around three stated pillars — technology, digital resilience, and people — placing AI capability build-out alongside cybersecurity continuity and workforce development. The disclosure does not quantify spending or headcount, leaving enterprise buyers and partners to interpret scope from program descriptions.
Business 2.0 Newsdesk
Meta Engineering has published a technical account of private processing for Meta AI Glasses, arguing the wearable form factor can interpret personal context while keeping sensor data handling under stricter control. The disclosure sets a design position for ambient AI hardware as enterprise buyers and regulators scrutinise always-on cameras and microphones.
Business 2.0 Newsdesk
OpenAI's newsroom details how Ringg's GPT-5.6 agents resolve up to 65% of customer calls across voice, chat, WhatsApp and web, running at roughly 90% lower cost than a GPT-4.1 baseline. The published figures shift contact-center buying decisions toward resolution rate and cost per conversation as the primary vendor metrics.
Business 2.0 Newsdesk
Microsoft has outlined a reimagined security operations model inside Microsoft Defender, treating AI agents as both a workload to protect and an operator inside the SOC. The company's security blog frames the shift as a structural redesign of detection, triage and response rather than a feature update, raising fresh governance questions for enterprise security leaders.
Business 2.0 Newsdesk
AMD published a new Advanced Insights video in which Mark Papermaster and Mike Clark argue that the Zen x86 architecture is evolving specifically for agentic AI workloads, placing general-purpose CPUs alongside accelerators in enterprise AI infrastructure planning. The framing matters because agentic software stresses orchestration, memory movement, and scheduling — the areas where CPU design choices show up directly in cost per task.
Business 2.0 Newsdesk
NVIDIA's Nemotron 3 Diarization model is now documented on Hugging Face for developers building real-time, multi-speaker speech pipelines. The release shifts speaker attribution from post-processing batch work into the live audio path, changing latency and compute assumptions for contact centers, meeting platforms, and compliance logging teams.
Business 2.0 Newsdesk