Montana passes Right to Compute act (2025)

Original Article ↗ Hacker News Discussion ↗

Montana’s new “Right to Compute” law, framed as protecting individuals’ ability to own and use computational resources, is widely seen by commenters as primarily shielding AI companies and data center operators from future regulation, zoning limits, and local opposition. Critics argue the bill’s AI “safety” provisions are largely symbolic and that its real effect is to preempt environmental, noise, and infrastructure concerns around large data centers, while doing nothing for software freedom, right-to-repair, or user control over personal devices. A minority view sees it as a modest but constructive safeguard against overreaching state controls on computing and AI use.

Scope and intent of the law

  • Thread consensus: the “Right to Compute” label suggests individual user rights, but the text mostly constrains government regulation of “computational resources” and sets a light requirement for AI in critical infrastructure.
  • Several note this aligns with positioning Montana as an AI/data center hub and contrasts it with more restrictive states.

Perceived beneficiaries and regulatory capture

  • Many argue the real goal is to make it harder for state/local governments to block or tightly regulate large data centers and AI platforms.
  • The name is widely criticized as PR/doublespeak: framed as a civil right while primarily aiding hyperscalers and investors.
  • Some see it as classic “regulation written by incumbents”: weak safety obligations that large firms can easily meet, while preempting stronger local rules.

Rights framing vs. actual protections

  • One camp thinks it modestly strengthens individual rights by:
    • Requiring any restriction on lawful compute use to meet a “compelling government interest” standard.
    • Potentially making it easier to challenge future compute/AI restrictions (compared to arguing from general free-speech principles).
  • Others counter that:
    • It explicitly carves out broad “compelling interests” (fraud, deepfakes, datacenter nuisances, etc.).
    • It may actually expand state justification to intervene by enumerating new “compelling” areas.
    • It does nothing about corporate control over devices (DRM, locked bootloaders, app-store power).

AI safety / critical infrastructure clause

  • The law requires deployers of AI-controlled “critical infrastructure” to create a risk management policy referencing standards (NIST, ISO).
  • Earlier drafts apparently included a mandatory shutdown mechanism; commenters note this was removed and survives only in the title.
  • Many call the requirement “toothless”:
    • Policy can be written after deployment.
    • Federal-compliance plans automatically count.
    • No clear enforcement or substantive safety constraints.

Datacenters, externalities, and local opposition

  • Debate over whether blocking data centers is reasonable:
    • Critics cite noise, water and power use, pollution, higher utility prices, and loss of local control.
    • Supporters argue concerns are exaggerated or NIMBY, and that predictable rules and investment outweigh downsides.

Language and missed opportunities

  • Side thread on “compute” as noun vs verb and language evolution.
  • Several lament that “Right to Compute” could have been used for genuine user-computing rights (repair, modifiable hardware/software, anti-DRM, anonymous use) but is instead applied to protect AI/data center buildout.