Launch: Latest AI News

Replica Cyber Ships Isolated AI Agents With Runtime Controls for Regulated Firms

The product makes containment, intervention and post-incident reconstruction part of the runtime environment. Its unanswered test is whether those controls hold when agents connect to the systems that make them valuable.

Aug 29, 2026Infrastructure

Last week

16 stories

Aug 30, 2026Startups

Launch Startups

Jeff Dean Leaves Google to Build Discovery Loop, a Science AI Startup Backed by Radical Ventures

The new public-benefit company has an unusually experienced founding team and support from both Radical Ventures and Alphabet. Its central technical question is still unanswered: what system will turn the ambition to automate research into a usable product?

4 min

Nvidia Rolls Out Vera Rubin as AI Competition Shifts to Data Traffic

The architecture packages compute with storage and networking hardware, betting that efficiently moving data through large AI systems is becoming its own competitive layer.

Aug 27, 2026Products

Oracle Adds AI Studio Skill to Fusion 26C, Deploying Pro-Code Agents Inside Fusion Apps

The new pro-code path is built around where an agentic application runs: Oracle is tying developer tools to Fusion’s existing data, workflow, security, and audit infrastructure.

Aug 27, 2026Products

Perplexity Puts Computer Behind MCP, Opening 400+ Services to AI Clients

The endpoint turns Computer into an execution layer for compatible assistants, but clients must preserve task state and surface logins, questions and sensitive-action approvals.

Aug 27, 2026Products

Harness Launches Agent-Ready Repository and AI Review for High-Volume Code Changes

The products pair high-volume repository operations with inherited permissions and merge-time checks, but Harness still places the final decision to ship with people.

Aug 26, 2026Infrastructure

Estuary Makes Rust Runtime Default, With Exactly-Once Delivery Still Conditional

The change is built to keep related database updates together as they reach agents and operational systems. But a complete end-to-end delivery guarantee still relies on the receiving destination and connector.

Aug 26, 2026Products

AWS AgentCore Evaluations Uses OpenTelemetry to Score Agents Across Frameworks

The service offers one set of agent-quality evaluators across SDK choices, but its portability depends on specific telemetry conventions, scope names and complete message logs.

Aug 26, 2026Models

Launch Models

Google Puts Gemini 3.5 Transcribe in Preview With Live and Recorded-Audio APIs

The release turns speech recognition into a Gemini product layer for developers and Google surfaces, but a product-specific price remains unavailable for teams weighing production use.

Aug 26, 2026Products

Nutanix Enterprise AI 2.8 Adds Agent Token Quotas and MCP Access Controls

Nutanix is putting controls around what agents can reach and consume, but says secure backend infrastructure, API keys and roles remain essential.

Aug 26, 2026Models

Launch Models

China Media Group Launches AI Media System Push With Telecom, Cloud and Chip Partners

The effort links a broadcaster’s production workflow to telecom networks, cloud computing, chips and security systems, but leaves the eventual model and deployment timetable undefined.

Aug 25, 2026Products

Perplexity’s Portable Computer Runs Agents Locally—but Needs a 24GB Nvidia GPU

The product turns a difficult DIY local-AI setup into one application, but its hardware floor and Linux-only launch keep it aimed at well-equipped paid users.

Aug 25, 2026Products

Google Cloud Launches Gemini Enterprise Agents for Finance and Legal, Starting in Preview

The new packages put specialized skills, data connectors and partner agents around regulated workflows. Their first test is whether a broad enterprise platform can become useful inside capital-markets, banking and legal operations.

Aug 25, 2026Products

Anthropic Connects Claude Science to 60+ Databases and Tools for Enterprise Work

The company’s challenge is to prove that integrations, traceability and controlled agents can become a durable advantage even as enterprises retain control of their sensitive systems.

Aug 24, 2026Infrastructure

NVIDIA Adds BlueField-4 Scale-In for 800 Gb/s AI-Factory Infrastructure Offload

The new layer targets a growing bottleneck around GPU clusters: moving, securing and governing data without consuming the host resources assigned to AI workloads. Its performance case remains NVIDIA-supplied.

Aug 24, 2026Infrastructure

Nvidia Puts Groq 3 LPX Into Production for Faster AI-Agent Responses

The first Nebius deployment will test Nvidia’s argument that specialized token generation can make long, tool-using AI sessions more responsive while leaving GPUs at the center of the system.

Aug 24, 2026Products

EverestLabs Puts AI Agents Near Recycling Controls, but Keeps Operators in Charge

Navigator is meant to turn camera data into diagnoses and proposed equipment settings without requiring a plant retrofit. Its commercial case now rests on whether operator-approved recommendations deliver the company’s projected throughput gains in live facilities.

Aug 22, 2026Infrastructure

August 2026

1 story

AWS Wants Data Agents to Write Pipelines, Then Stay Out of Production

The reference architecture uses models to draft data-engineering work, but makes human-reviewed code and policies—not live agent decisions—the default production layer.