The constraint facing AI infrastructure developers is legal exposure. Without a defined liability shield, the cost of model failure remains an open variable in the build-out economics. Treasury Secretary Scott Bessent told CNBC's "Squawk Box" that the administration will not grant AI leaders a liability shield. This decision removes a proposed regulatory buffer from the equation, leaving the legal framework for AI deployment to be shaped by existing tort law and future legislation rather than executive action. The mechanism here is straightforward: by refusing to create a safe harbor, the government keeps the financial risk of AI errors with the operators and their insurers. This places the burden on the private sector to price that risk into their capital structures. Bessent made these comments while discussing AI safety concerns. He also addressed the upcoming summit between Chinese President Xi Jinping and President Donald Trump. The juxtaposition of domestic regulatory stance and foreign policy engagement highlights the dual nature of the current AI policy environment. On the domestic front, the absence of a liability shield means that companies building large-scale AI systems must navigate a legal landscape where negligence standards are still being defined by courts. There is no specific unit or metric provided in the release that quantifies the potential financial impact of this decision. The statement serves as a clear signal that the White House is not prepared to intervene directly in the liability question. For the tech sector, this means the interconnect between technological capability and legal responsibility remains tight. The specific unit that drives the economics here is the legal defense cost, which will vary by case but is now explicitly not capped by federal policy. Bessent's position aligns with a broader trend of the administration favoring market-driven solutions over regulatory protections for emerging technologies. The summit with President Xi offers a contrast to this domestic approach. International negotiations on AI safety and standards will likely proceed in parallel with the domestic legal vacuum. The outcome of those talks could influence how US companies perceive their competitive position against Chinese counterparts. However, the immediate takeaway for domestic stakeholders is the rejection of the liability shield. This is a concrete policy decision that affects how AI companies structure their risk management. The lack of a shield does not mean the absence of regulation, but it does mean the specific protection sought by industry leaders is not forthcoming. Bessent's remarks were part of a broader discussion on AI safety. The administration's stance is now on the record. Companies must plan their operations under the assumption that liability will be adjudicated on a case-by-case basis. This adds a layer of uncertainty to the deployment of new AI models. The legal team becomes a critical node in the product development stack. The cost of compliance and legal review is now a fixed line item in the operational budget. This decision by Bessent sets the tone for how the administration will handle similar requests from other sectors. It signals a preference for existing legal frameworks over new, sector-specific protections. The impact will be felt most acutely by companies at the frontier of AI development, where the risks are highest and the legal precedents are least developed. The refusal to provide a shield is a deliberate choice to maintain the status quo of legal liability.