The leading frontier model developers are moving closer together on safety standards. On September 15, 2026, Chris Lehane, Global Policy Chief at OpenAI, confirmed in Washington to Bloomberg that his company has been engaged in deep discussions with Google DeepMind and Anthropic for several weeks. The primary goal of these talks is establishing shared safety guidelines for next-generation frontier artificial intelligence models.
Lehane explicitly addressed potential antitrust concerns surrounding the initiative. He emphasized that the participating companies do not require an antitrust waiver to coordinate on safety frameworks. Drawing a direct parallel to the commercial aviation industry, Lehane noted that competing aircraft manufacturers and airlines routinely align on universal baseline safety standards without violating antitrust statutes or dampening market competition.
These corporate discussions arrive amid an accelerating political debate in Washington regarding whether to intentionally pace frontier training runs. Members of the United States Congress are moving forward with multiple legislative proposals. Bipartisan groups, including initiatives led by Representatives Lori Trahan and Jay Obernolte as well as Senators John Thune, Ted Cruz, and Amy Klobuchar, are drafting rules to oversee large-scale frontier compute.
However, the prospect of any mandated slowdown has triggered sharp pushback from prominent figures in politics and venture capital. Speaking at the All-In Summit, investor David Sacks and Donald Trump both forcefully rejected any technological pause. They argued that curbing domestic progress would undermine American leadership and cede strategic advantages to China in the global technological race.
The debate over potentially decelerated investment cycles is already impacting financial markets. According to reports from the Financial Times, tech equities have experienced heightened volatility as Wall Street reassesses the timeline for capital deployment. The trilateral dialogue among OpenAI, Google DeepMind, and Anthropic represents a clear effort to establish industry-led governance before mandatory legislative interventions take hold.