Is Frontier AI Becoming a Geopolitical Resource?

In April, Anthropic released its most capable model, Claude Mythos Preview, not to the public but to a closed consortium of partners under Project Glasswing. The White House intervened to block an expansion of that list, making it the first known case of the US government actively restricting an advanced AI model rollout.
In his article Cut Off, Anton Leicht claims that this signals a trend toward restricted frontier access. Leicht identifies 3 forces that will further constrain the availability of advanced AI models in the coming future: security concerns, compute scarcity, and US government involvement.
In this group discussion, we'll debate around these questions:
Is frontier AI access really becoming scarce and selective?
Should frontier AI be governed like a strategic resource? And if so, by whom?
What does this mean for Spain and the EU, who are neither frontier developers nor currently at the table?
No background required; only familiarity with the article Cut Off (~10-minute read).