01 OpenAI and Microsoft amend partnership; exclusivity removed
OpenAI and Microsoft announced an amended agreement that the companies say simplifies their relationship and provides long‑term clarity. OpenAI’s public blog post describes the change as supporting continued AI innovation at scale and clarifying commercial terms between the parties.
Reporting from TechCrunch and Ars Technica says the amendment gives OpenAI the right to sell products on other cloud platforms — explicitly clearing the way for models to run on Amazon Bedrock — while Microsoft receives expanded revenue‑share arrangements and other concessions. The Verge and other outlets characterize the update as ending a previous exclusivity around cloud deployment.
Microsoft will remain OpenAI’s primary cloud partner and OpenAI products will continue to ship first on Microsoft’s infrastructure, according to coverage, but the contract no longer bars OpenAI from using additional cloud providers. The change also removed a longstanding clause tied to an AGI‑focused commitment that had shaped earlier terms of the partnership.
Industry observers say the amendment shifts commercial freedom toward OpenAI while preserving a deep commercial and financial relationship with Microsoft; TechCrunch frames the deal as resolving legal and commercial peril tied to OpenAI’s previously announced arrangements with Amazon.
- OpenAI and Microsoft issued an amended agreement that removes exclusivity and clarifies the partnership.
- OpenAI can now sell products to run on other cloud providers, including Amazon Bedrock.
- Microsoft remains the primary cloud partner and gains expanded revenue‑share terms.
02 Musk v. Altman: jury selection opens in lawsuit over OpenAI’s future
Jury selection began in the lawsuit Elon Musk filed in 2024 challenging OpenAI’s governance and alleging the company abandoned its founding nonprofit mission, according to reporting from The Verge and Ars Technica. The trial is framed as a high‑stakes contest over the direction and control of one of the tech sector’s leading AI organizations.
The Verge reports that Musk’s suit asks the court to compel changes that could affect OpenAI’s corporate structure; Ars Technica notes that Musk’s shifting public stance on AI safety could complicate the trial’s narrative. Both outlets say the litigation will probe internal decisions made as OpenAI moved from its original nonprofit roots toward a capped‑return model and large commercial partnerships.
The proceedings are likely to surface testimony about board decisions, investor agreements, and the company’s evolving mission. Coverage highlights that the outcome could set legal and practical precedents for how mission commitments are enforced when nonprofit origins meet venture‑scale commercial funding.
Observers say the trial’s factual record may influence investor and partner behavior for AI companies beyond OpenAI, regardless of the ultimate ruling, because it will clarify how courts view nonprofit‑to‑commercial transitions and governance promises.
- A jury is being selected in Elon Musk’s suit that challenges OpenAI’s governance and mission.
- Reporting indicates the trial will examine past board decisions and investor agreements.
- The case could set precedents for how courts treat mission commitments when organizations commercialize AI.
03 Over 600 Google employees ask leadership to block classified Pentagon use of models
More than 600 Google employees signed a letter to CEO Sundar Pichai asking the company to deny the Pentagon access to Google AI models for classified purposes, The Verge reports. Organizers say many signers work in DeepMind and include principals, directors, and vice presidents.
The letter — reported by The Washington Post and summarized by The Verge — demands that Google set firm limits on classified military use of its AI technology, arguing that such uses conflict with employee ethical concerns and the company’s values. The organizing employees want leadership to publicly commit that Google will not permit classified use.
Google’s internal debate comes as other major AI players navigate contracts with government and defense agencies; the employee action underscores rising pressure inside tech firms to set explicit bounds on military and classified applications of advanced models.
The episode adds to an ongoing conversation about corporate responsibility, employee influence, and how large AI providers balance commercial opportunities with ethical and national‑security considerations.
- Over 600 Google employees, including senior DeepMind staff, asked CEO Sundar Pichai to refuse classified Pentagon use.
- The letter seeks a public commitment from Google to block classified military applications of its models.
- The demand highlights internal pressure at tech firms over ethical limits on government AI use.
What moved around the edges
業務ソフトで新しい論点が浮上
OpenAI is available at FedRAMP Moderate authorization for ChatGPT Enterprise and the OpenAI API, enabling secure AI adoption for U.S. federal agencies.
OpenAI BlogDavid Silver’s Ineffable Intelligence raises $1.1B at a $5.1B valuation
TechCrunch reports the new lab led by former DeepMind researcher David Silver raised $1.1 billion in funding at an estimated $5.1 billion valuation to pursue AI that reduces reliance on human‑labeled data.
Hacker News AIGoogle and Kaggle open 'Vibe Coding' GenAI intensive for developers
Google’s AI blog and Kaggle are offering a hands‑on 'Vibe Coding' GenAI course beginning in June 2026 that focuses on building agents and generative‑AI tools for developer workflows.
Google AI BlogMicrosoft and OpenAI remove AGI clause but keep cloud partnership terms
The Verge reports the two companies have dropped a clause tied to an artificial general intelligence commitment while confirming Microsoft remains OpenAI’s primary cloud partner and that OpenAI products will ship first on Microsoft infrastructure.
The Verge AIIsItArtStudio turns AI agents loose to create and critique public art
IsItArtStudio launched a public platform where autonomous agents can register to author prompts, submit artworks, and critique submissions; the project’s corpus is public and writing access is intended for agents rather than human authors.
Hacker News AI