Issue #120
Morning Roost: AI compute deals, Nvidia's open models, and a hack on OpenCode
Tuesday, October 20, 2026
Today covers reported compute and chip developments from Anthropic, Nvidia, Amazon, and Tenstorrent, Microsoft's stance on AI agent security, a new Yelp calling feature, and a security breach affecting the OpenCode tool.
1 of 7 · TechCrunch AI
Nscale and Anthropic are reportedly in talks for yet another $45B compute deal
Nscale and Anthropic are reportedly in discussions over a compute deal valued at $45 billion. The agreement would add to a series of large compute arrangements Anthropic has made over the past year.
Read the full story →2 of 7 · TechCrunch AI
Nvidia launches new open models to run agents in the enterprise and on PCs
Nvidia released new open models designed to run AI agents in enterprise data centers and on local PCs. The move targets companies seeking to operate their own models and supports a push toward on-device, private AI agents.
Read the full story →3 of 7 · TechCrunch AI
Amazon is taking its biggest shot at Nvidia yet with new TorchNeuron dev tool
Amazon released a development tool called TorchNeuron aimed at making it easier for developers to run PyTorch applications on non-Nvidia chips, including Amazon's own processors. The move targets Nvidia's dominance in PyTorch, the open-source toolkit used to build most AI applications.
Read the full story →4 of 7 · TechCrunch AI
Tenstorrent hits $6.2B valuation to scale open-weight AI chips
Tenstorrent raised $1.5 billion at a $6.2 billion valuation to continue developing its open-weight AI chips. The company, positioned as an Nvidia competitor, had drawn acquisition interest earlier this year but chose to remain independent.
Read the full story →5 of 7 · TechCrunch AI
Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’
Microsoft is advising enterprises to monitor AI agents as closely as they would insider threats, citing concerns that frontier models cannot be fully trusted. The guidance focuses on agents that employees use with systems outside the control of AI labs such as OpenAI and Anthropic.
Read the full story →6 of 7 · TechCrunch AI
Yelp can now call on your behalf to check wait times at a restaurant
Yelp is adding AI features to its app, including the ability to call restaurants on a user's behalf to check wait times and generate personalized recommendations based on saved preferences.
Read the full story →7 of 7 · TechCrunch AI
Hackers hit AI coding tool OpenCode, running up charges on users’ AI accounts
Hackers compromised part of OpenCode's online platform and stole some users' API keys, using them to run up charges on the victims' AI accounts. OpenCode is an open source AI coding agent that has gained popularity as an alternative to Claude Code and Cursor.
Read the full story →That is today's roundup. We will be back tomorrow morning.
Morning Roost Issue #120