Global ideas, clearly explained.

Search

5 results for β€œenterprise”

Microsoft Rebuilds Copilot Around Home, Code and Autopilot, Betting the Work Surface Is the Moat
Technology

Microsoft Rebuilds Copilot Around Home, Code and Autopilot, Betting the Work Surface Is the Moat

Microsoft has rebuilt Copilot around three capabilities β€” Home, Code, and Autopilot β€” turning the AI assistant into a unified work surface where users chat, build apps in natural language, and delegate to persistent agents billed by usage.

Meridian Newsroom30 Sept 20265 min read
Thinkific Cuts 96 Roles in AI-First Restructuring, Targets 25% Free Cash Flow Margins
Startups

Thinkific Cuts 96 Roles in AI-First Restructuring, Targets 25% Free Cash Flow Margins

The Vancouver learning platform expects $19 million in annual savings as it pivots to enterprise customers and AI-first product development β€” 2026's startup playbook in miniature.

Meridian Newsroom30 Sept 20268 min read
Island's $400 Million Series F at $6.4 Billion Bets the Browser Can Govern AI Agents
Startups

Island's $400 Million Series F at $6.4 Billion Bets the Browser Can Govern AI Agents

The enterprise browser pioneer is recasting itself as an 'agentic control plane' β€” and investors from Sequoia to Evolution Equity Partners just priced that thesis at $6.4 billion.

Meridian Newsroom30 Sept 20268 min read
OpenAI's $70 Billion Inflection Point: The Enterprise Pivot Behind the Fastest Revenue Run in Tech
Case Studies

OpenAI's $70 Billion Inflection Point: The Enterprise Pivot Behind the Fastest Revenue Run in Tech

On September 29, 2026, Axios reported OpenAI's annualized revenue run-rate is approaching $70 billion β€” up more than 70% since July, powered by enterprise sales that more than doubled. Behind the headline is a case study in converting consumer reach into corporate contracts.

Meridian Newsroom29 Sept 20265 min read
Anthropic's Claude Sonnet 5.5 Beats the Flagship on Coding at Half the Price
Artificial Intelligence

Anthropic's Claude Sonnet 5.5 Beats the Flagship on Coding at Half the Price

The mid-tier model's 70.6% score on Terminal-Bench 4.0 β€” ahead of Opus 5.5 β€” shows how close cheaper models have come to flagship performance on the work developers actually do.

Meridian Newsroom29 Sept 20266 min read