Congress Probed Kimi. Musk Paid $60 Billion.

On July 8, CNBC reported Republican House chairs are still probing Cursor's Kimi-built Composer 2 while Beijing blacklisted Claude Code and weighed export curbs on Qwen — the same week SpaceX agreed to buy Anysphere for $60 billion and SpaceXAI may ship its first joint model with Cursor. Congress is investigating the Chinese stack inside the tool Musk just bet the farm on.

What's new:

• CNBC's July 8 report revives an April joint probe by the House Homeland Security and China Select committees into U.S. firms adopting PRC-developed AI.
• Chairmen Andrew Garbarino (R-NY) and John Moolenaar (R-MI) sent letters to Anysphere CEO Michael Truell and Airbnb CEO Brian Chesky demanding details on Chinese model use, internal security reviews, and communications with PRC providers.
• Garbarino told CNBC: "The Chinese Communist Party is no longer just nipping at our heels in artificial intelligence; it is racing to close the gap in some of the exact capabilities that will shape the future of cybersecurity."
• Cursor declined to comment. Airbnb said its AI activity runs "overwhelmingly on U.S.-origin models" and limited China-origin open-source use runs only through approved U.S.-based providers.

Why July 8 matters:

• The probe landed the same week Beijing's MIIT flagged Claude Code as a national backdoor risk and Reuters reported Beijing is weighing curbs on overseas access to Alibaba's Qwen, ByteDance's Doubao, and Z.ai's GLM.
• TNW reported SpaceXAI and Cursor may release their first jointly trained model as soon as Wednesday, citing an internal memo from The Information that benchmarks against Anthropic's Opus 4.8 and OpenAI's GPT-5.5.
• Musk rebranded xAI as SpaceXAI on Monday and folded the unit into SpaceX five months after the February merger.
• Washington and Beijing are now mirroring each other's suspicion — one capital probes American firms for using Chinese weights; the other blacklists American coding agents.

The Kimi problem:

• Semafor first reported the Anysphere letter focuses on Composer 2, which Cursor disclosed was built atop Kimi — Moonshot AI's open-weight model out of Beijing.
• The Hill reported lawmakers named Moonshot alongside DeepSeek and MiniMax as firms running "coordinated campaigns to extract advanced capabilities" from American AI systems.
• Garbarino's letter warned: "When capabilities are stripped out through distillation and repackaged without equivalent safeguards, the resulting models may become available to hostile state actors, terrorist organizations, and criminal enterprises."
• Moolenaar told Semafor: "The AI models these companies use are trained by China's censorship regime and introduce hidden vulnerabilities that put Americans' data and businesses at risk."

The $60 billion contradiction:

• SpaceX announced June 16 it would acquire Anysphere in an all-stock deal worth $60 billion, expected to close in Q3 2026 subject to regulatory approval.
• CNBC noted Cursor built Composer 2 on Kimi even as SpaceX poured $12.7 billion into AI last year — more than triple its space spend, per TNW.
• Truell posted on X in June he was "excited to partner with the SpaceX team to scale up Composer." TNW reported he told customers Cursor trained a model from scratch on Colossus, xAI's supercomputer.
• A committee aide told CNBC the probe also asks whether Washington has a viable open-weight strategy so firms are not "forced to choose between expensive or restricted U.S. models and cheap, capable PRC-developed alternatives."

What lawmakers want next:

• Disclosure of every PRC-developed model deployed, plus independent security evaluations and internal risk analyses.
• In-person briefings from employees who chose Chinese models.
• A broader answer to Rep. Andy Ogles' June warning: "When the cheap, capable, easy option for an AI model is Chinese, the rest of the world will build on it."
• Brookings fellow Kyle Chan told CNBC federal procurement bans are the likeliest lever — but banning open weights is "ultimately impossible" because model files sit freely online and could raise First Amendment issues.
• CNAS senior fellow Daniel Remler said the administration may fear action "could harm start-ups that use these models, or chill support for open models generally."

Convina's view: Congress opened a national-security file on Kimi in April. Musk agreed to a $60 billion bet on Cursor in June. Beijing blacklisted Claude Code on July 8. None of these moves are about model benchmarks — they are about who owns the plumbing. Open weights made Chinese AI cheap enough for American startups to build on; distillation politics made that choice toxic; and now both capitals are racing to gate access faster than either can gate code. The Kimi probe will not stop the SpaceX deal. It will force every enterprise buyer to ask whether their coding stack runs on weights a House committee already flagged — and whether Musk's Colossus training run fixes that provenance or just buries it under a bigger logo.