Tools

Google's I/O 2026 conference unveiled Gemini Spark, a 24/7 personal AI agent in the cloud, a Search overhaul, and a new $100 Ultra plan. Here's what shipped and what's still months away.

Dell announced Deskside Agentic AI at Dell Technologies World 2026, offering local AI agent workstations that run open-weight models on-premises and cut cloud API costs by up to 87% over two years.

ClickUp is running 3,000 AI agents for 1,300 employees , a 3-to-1 ratio that its CEO calls the new normal of enterprise work. Here's how they actually built it, and what's hard about keeping it running.

Intercom renamed itself Fin on May 12, 2026, putting its AI customer agent's name on the parent company. Then, three days later, it launched a second AI agent whose only job is managing the first one.

OpenAI launched personal finance tools for ChatGPT Pro users, letting them connect bank accounts via Plaid. Here's what it actually does, what it can't do, and the privacy questions worth thinking through.

xAI launched Grok Build on May 14, a terminal-native coding agent with an 8-parallel-agent architecture and 2 million token context window. Here's what developers need to know.

Apple is designing a new compliance framework that would allow AI agents to operate inside App Store boundaries, with details expected at WWDC on June 8 alongside a major Siri overhaul.

Amazon retired Rufus and launched Alexa for Shopping, an AI assistant that monitors prices, schedules purchases, and buys things on your behalf, even on websites Amazon doesn't own.

Meta revealed plans for 'Hatch,' a personal AI agent, and a shopping tool for Instagram that completes purchases from Reels without leaving the app. Internal testing targets end of June; shopping tool aims for Q3 2026.

Microsoft says Agent 365 is generally available, giving enterprise teams a unified control layer for AI agents across security, identity, and workflow systems.

Mistral’s new remote agents in Vibe point to a larger shift in how coding assistants are used: less pair-programming at the keyboard, more parallel cloud execution with human review at decision points.

Anthropic says 6% of sampled Claude conversations involve personal guidance requests, a behavior shift that forces product teams and enterprises to rethink AI trust, safety policy, and governance controls.

OpenAI says it has already secured 10 gigawatts of U.S. AI compute capacity, years ahead of its 2029 goal, signaling that power contracts and infrastructure access are now central to AI competition.

Lens says its new Lens Agents platform applies policy, identity, and audit controls across AI agents running in cloud stacks and employee desktop tools, signaling a governance-first shift for enterprise AI operations.

AWS's April 27 roundup puts Bedrock AgentCore CLI alongside major ecosystem signals, showing how enterprise AI agents are moving from pilots to repeatable delivery workflows.

Stripe says merchants will soon be able to sell inside Google AI Mode and the Gemini app, a move that could shift AI shopping from demo behavior into measurable transaction flow.

Choco and OpenAI shared new production metrics, including 8.8 million annual orders and a reported 50% drop in manual entry. The update shows agent systems moving from pilots into daily distributor operations.

OpenAI launched workspace agents in ChatGPT on April 22, 2026, putting shared cloud-run automation with admin controls into team workflows. Here is what changes for enterprise rollout decisions.

ONNX Runtime v1.25.0 landed on April 20, then surged on GitHub Trending by April 24. Here is what that timing means for inference reliability, cost control, and enterprise deployment strategy.

Meta will add tens of millions of AWS Graviton cores, highlighting a broader market shift: CPU-heavy orchestration is now a first-order planning factor for enterprise AI infrastructure.

Meta and Broadcom extended their custom AI silicon partnership through at least 2029. The move signals a longer planning horizon for compute capacity, networking design, and enterprise AI cost control.

Cyera’s acquisition of Ryft highlights a deeper shift in AI security. Enterprise teams now need controls over how agent systems assemble and move sensitive data, not just guardrails around model prompts and outputs.

Microsoft plans to invest A$25 billion in Australian AI and cloud infrastructure by 2029. We break down what it changes for capacity access, procurement strategy, and delivery risk.

At Cloud Next on April 22, 2026, Google introduced TPU 8t for training and TPU 8i for inference. The bigger story is how this split changes enterprise AI infrastructure decisions on cost, latency, and governance.

VAST Data closed a Series F round at a $30 billion valuation. The bigger signal is where AI infrastructure budgets are moving, from raw GPU supply toward data-and-execution control planes.

The ChatGPT Images 2.0 launch points to a more practical phase of image AI, where text fidelity and workflow fit matter more than novelty screenshots.

Google launched Agentic Data Cloud to help AI agents act on enterprise data. Here is what data teams should change first to scale safely and avoid governance bottlenecks.

ServiceNow and Google Cloud announced new enterprise AI agent integrations focused on autonomous operations. The move gives IT and ops leaders a concrete signal on what to change in rollout planning now.

At Cloud Next 2026, Google introduced TPU 8t for training and TPU 8i for inference. The split points to a new infrastructure playbook for AI teams that need speed in model development and lower latency in production.

Google moved Gemini Enterprise from a collection of tools into one agent platform, and that changes how IT leaders should manage deployment risk, observability, and workflow ownership.

Google introduced prepay billing in AI Studio, a change aimed at making Gemini API spend easier to predict for teams moving from prototypes to heavier usage.

Vercel announced Workflows general availability, giving teams a framework-native way to run long-lived execution paths that many AI and automation products require.

Google updated Chrome AI Mode with side-by-side browsing, signaling a push toward persistent assisted navigation instead of one-pane answer experiences.

Cloudflare launched AI Search as a native primitive for agents, aiming to reduce indexing friction and make retrieval-first workflows easier to run in production.

OpenAI says Codex now supports broader computer workflows beyond code generation, signaling a faster move toward agentic software tasks that span planning, execution, and verification.

Perplexity is expanding Personal Computer, its always-on Mac mini agent setup, which signals a bigger shift from chat interfaces toward persistent software that executes multi-step work.

Anthropic introduced Claude Design in research preview on April 17, 2026. The new product moves Claude deeper into design and prototype workflows, with direct implications for product, design, and engineering teams.

OpenAI says ChatGPT now sees about 8.4 million weekly advanced science and math messages from roughly 1.3 million weekly users, a signal that AI tools are moving deeper into day-to-day research workflows.

Cloudflare introduced Project Think, a new layer for its Agents SDK focused on durable execution, sub-agents, persistent sessions, and sandboxed code execution for long-running agent workloads.

Google introduced Android CLI, Android Skills, and an Android Knowledge Base to help AI coding agents handle Android workflows outside Android Studio with lower token usage and faster setup.

Cloudflare launched Browser Run for AI agents with live view, human handoff, CDP access, session recordings, and a 4x increase in concurrency limits.

GitHub now allows enterprise admins to enable Copilot cloud agent for selected organizations, including policy targeting via custom properties.

Google AI Studio now supports prepay credits for Gemini API usage, starting in the US for new billing accounts and rolling out globally in the coming weeks.

Google launched Skills in Chrome, letting users save and rerun Gemini prompt workflows with one click. The feature includes a starter library and confirmation safeguards for sensitive actions.

Anthropic says Claude Code routines are in research preview with schedule, API, and event triggers, and they run on Anthropic’s web infrastructure instead of a local laptop session.

Developers are reporting higher Claude Code usage burn after cache TTL changes, while Anthropic staff argue the shift can reduce cost in many sessions. The gap now is workload shape.

Meta and Broadcom expanded their custom AI chip partnership through 2029 with an initial 1 gigawatt deployment commitment. The deal sharpens Meta’s internal AI infrastructure strategy.

GitHub released a Secure Code Game focused on agentic AI failure modes. The new training path helps teams practice prompt-injection defense, safer tool use, and stronger review loops before production incidents.

New reporting says Google may be testing an Agent workspace in Gemini Enterprise. The signal matters because it suggests a broader desktop and task-orchestration strategy.

Leak-focused reporting says OpenAI may add web browsing and new workflow surfaces to Codex. We separate confirmed product direction from claims that still need official verification.