Any specific market resolves to "Yes" if by January 1, 2030, for the relevant entity, the criteria for "soft nationalization" or "hard nationalization" occur. That market resolves to "No" otherwise.

If it is unclear on 1st January 2030 whether the criteria above are met, resolution will be delayed by up to 30 days. If it is still unclear, the market will resolve NO.

The criteria for "soft" and "hard" nationalization are below. Below that are clarifications on the relevant entities.

Resolution will be based on information available in:

• Official government documents

• Company publications and SEC filings

• Federal Register publications

• Congressional records

• Legal proceedings documentation

• News reports from reputable sources such as AP
  
  Note that a "hard nationalisation" will resolve the respective "soft nationalization" question to "Yes"

Hard nationalization:

1. The US federal government has acquired majority ownership and voting rights in the relevant entity, including through:

• Full acquisition with purchase of all existing equity

• Forced transition to state-owned enterprise, e.g. by legislative action or national security directive

• Complete integration into a government-run agency

or

2. The government exercises comprehensive control over the relevant entity, such as if most of the following hold:

• All personnel become federal employees with security clearances

• Appointments of executives are done directly by government or require government approval

• All IP and research performed by the company is classified and government-owned

• The US government is the sole legal distributor of the relevant entity’s products

Soft nationalization:

This market will resolve YES if either:

1. The conditions for total nationalization are met as described above OR

2. OpenAI becomes subject to direct federal oversight where at least TWO of the categories (A-C) of control are implemented, where a category counts as implemented if ANY of its listed conditions are met:

A. Management & Governance

• Permanent government liaisons are embedded within the company with broad oversight authority

• Two or more board seats are held by government appointees with voting rights

• Executive appointments require federal approval

B. Operational Control

• Federal licensing requirement for training runs above specified compute thresholds

• Government pre-approval required for:

  • Any customer contract exceeding $10M annually

  • Any foreign customer regardless of contract size

  • Any deployment to government or military customers

• Compute resources exceeding specified thresholds require federal permits

• Mandatory security clearances for researchers working on specified AI capabilities

• Mandatory compliance with federal cybersecurity protocols exceeding standard industry regulations

• Mandatory sharing of technical research with federal agencies before public release

C. Financial Control

• Government holding special class shares with veto rights over major decisions

• Direct government ownership exceeding 25% of voting shares

• A profit cap for private investors or mandatory profit-sharing with the federal government

For resolution purposes, any specific entity refers to:

• The entity currently known by that name

  • If there is some way that the entity currently colloquially called one of these names technically has a different legal name, it refers to that entity

  • OpenAI refers to both the non-profit and for profit

• Any successor organization that inherits the majority of that entity's staff and intellectual property

• Any collection of successor entities that collectively inherit the above assets

• Any new entity formed by a merger, acquisition, or reorganization involving that entity's core AI development team and capabilities
  

"Frontier AI company" is any of the named companies plus any other company that develops AI models as per the definition below. "Other frontier AI company" is any company that fulfils this definition that isn't one of the named companies. Here is the definition for "other AI firm":

• Exceed 1 trillion parameters in size

• Cost more than $10m in compute to train (2024$)

• Require more than 10^24 FLOPs to train