Wednesday — June 17, 2026

SpaceX announced today that it has entered into a definitive agreement to acquire Anysphere, the company behind the Cursor AI coding editor, for approximately $60 billion in an all-stock transaction. The deal represents the largest AI acquisition of 2026 and signals a dramatic shift in how the aerospace company views its role in the broader AI ecosystem. Cursor has rapidly become the go-to AI coding assistant for developers, known for its deep codebase understanding and agentic capabilities. By combining Cursor with xAI's Grok models and SpaceX's massive computational infrastructure, the company aims to create what it calls 'the most advanced AI coding agent ever built.' The acquisition also positions SpaceX to compete directly with Microsoft/GitHub Copilot, Amazon CodeWhisperer, and Anthropic's Claude Code in the developer tools market. For developers currently using Cursor, expect integration with Grok models in the coming months, along with potential pricing changes as the tool scales under SpaceX's ownership. The deal is expected to close in Q3 2026, pending regulatory approval.

On June 15, 2026, Anthropic officially deprecated Claude Sonnet 4 and Opus 4, meaning any API calls using these specific model identifiers will now return errors. This is a hard deprecation—not a soft sunset—so developers who haven't migrated will see immediate failures in production. The recommended replacements are claude-sonnet-4-5 and claude-opus-4-7, which offer improved reasoning, longer context handling, and better instruction following. Anthropic has provided migration guides, but the window is tight. In related news, Anthropic paused a planned change that would have moved automated Agent SDK usage onto a separate monthly credit. This pause gives developers continued flexibility in how they use agents, particularly for those running Claude Code and Managed Agents at scale. The deprecation comes as Anthropic doubles down on its newer model families and prepares for the next generation of Claude models. For teams running production workloads, this is an urgent wake-up call: update your model identifiers, test your prompts, and validate that agent behavior hasn't changed with the new versions.

Salesforce announced the acquisition of Fin for $3.1 billion, adding the company's AI customer service agents to its existing suite of enterprise tools. Fin has built a reputation for deploying AI agents that can handle complex customer service interactions using its proprietary Apex model, which is designed specifically for support workflows. The acquisition is a clear signal that Salesforce sees AI-powered customer service as a critical battleground in the enterprise software market. By integrating Fin's technology, Salesforce can offer businesses pre-built, customizable AI agents that handle everything from password resets to complex billing disputes without requiring extensive AI expertise. For Fin customers, the transition means deeper integration with Salesforce's CRM, Service Cloud, and Data Cloud platforms. However, it also raises questions about data sovereignty and vendor lock-in, as Fin's agents will increasingly rely on Salesforce's infrastructure. The deal is expected to close in the coming weeks, and Salesforce has indicated that Fin's agents will remain available as standalone products for the near term.

Two major policy developments today signal a significant shift in how the US government approaches AI regulation. First, President Biden signed National Security Presidential Memorandum 11 (NSPM-11), which establishes a comprehensive policy framework for AI use across the national security enterprise. The memorandum accelerates adoption of AI in defense and intelligence while calling for voluntary cybersecurity assessments with frontier AI developers. It emphasizes secure, governable AI deployments and sets expectations for builders in defense-adjacent spaces. Second, a bipartisan group of lawmakers released a discussion draft of the Great American AI Act, which proposes binding obligations for large frontier developers—defined as companies with over $500 million in annual revenue that train frontier AI models. The act would require safety testing, transparency reporting, and adherence to federal standards. Together, these developments indicate a clear trend toward increased regulatory oversight for powerful AI systems. For AI companies, this means compliance teams need to be staffed up, and for enterprises using AI, it means due diligence on your vendors' regulatory posture will become standard practice.

Google announced that the Gemini CLI will reach end-of-life on June 18, 2026—just one day from now. After this date, the CLI will no longer receive updates or support, and Google strongly recommends migrating to the new Agentic 2.0 CLI. The Agentic 2.0 CLI offers improved agent orchestration, better integration with Google Cloud services, and enhanced security features. For teams using the Gemini CLI in CI/CD pipelines, automated scripts, or development workflows, this is an urgent migration. The new CLI has a different command structure and configuration format, so migration will require updating scripts and testing pipelines. Google has provided migration documentation, but the timeline is extremely tight. This deprecation is part of Google's broader strategy to consolidate its AI tooling around the Agentic platform, which also includes the recently released Agentic 2.0 API and managed agent services.