AI Security Breaches, Funding, and New Open Models

Intro

AI continues to reshape software development, personal assistance, and security. Today’s highlights span from a major hacking incident to fresh funding and new open‑weight chip capabilities.

OpenCode coding tool hit by hackers

OpenCode, an open‑source AI coding agent, suffered a breach where attackers stole users’ API keys. The stolen keys were used to run AI tools that refused to process malicious code, escalating charges on user accounts. Why it matters: Developers relying on OpenCode must audit key security and monitor usage to avoid unexpected billing and data leaks. Read more: Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts

Tolan secures $30M for AI life‑navigation companion

AI companion app Tolan raised an additional $30 million after amassing more than 250,000 active users. The funding highlights growing interest in AI assistants that guide users through daily decisions. Why it matters: Founders can look to Tolan’s growth trajectory as a benchmark for scaling personal‑assistant platforms. Read more: Tolan snags $30M for its AI companion that helps users navigate life

Replika founder launches free personal AI assistant

Eugenia Kuyda’s new startup Wabi offers a downloadable personal AI assistant, a concept she pitched over a decade ago. The free model provides an accessible entry point for users seeking AI help in daily tasks. Why it matters: The move shows that free, downloadable AI can coexist with subscription models, impacting pricing strategies for developers. Read more: Replika’s founder built the AI assistant she pitched at Disrupt 12 years ago

Tenstorrent raises $1.5B to scale open‑weight chips

Tenstorrent, a challenger to Nvidia, closed a $1.5 billion round to keep its open‑weight AI chip line independent. The funding reflects investor confidence in scalable, open‑source hardware for AI workloads. Why it matters: Hardware vendors and enterprises can use Tenstorrent’s chips to reduce reliance on proprietary solutions and accelerate deployment of AI models. Read more: Tenstorrent hits $6.2B valuation to scale open-weight AI chips

Nvidia launches open models for private on‑device agents

Nvidia announced new open models designed for enterprise data centers and private on‑device AI agents. The initiative targets companies that want to run AI locally while maintaining data privacy. Why it matters: Developers can integrate these models into existing infrastructure to offer secure, offline AI experiences for customers. Read more: Nvidia launches new open models to run agents in the enterprise and on PCs

AI agents breach multiple corporate sites

A hacker used an army of AI agents to compromise more than 100 websites, including John Deere, SharkNinja, and Cummins, according to AI security firm Norforge. The attacks exploited automated, self‑learning agents that identified and exploited vulnerabilities. Why it matters: Businesses must evaluate AI‑driven attack vectors and strengthen defensive tooling against autonomous threat actors. Read more: A hacker broke into John Deere and SharkNinja with an army of AI agents

What to watch

• Open-source AI tool security practices and key management.
• The pace of funding for AI‑assistant startups and their monetization models.
• Growth of open‑weight AI chip ecosystems and their impact on hardware choices.
• The rise of AI‑driven corporate attacks and the need for adaptive defenses.