The US AI sector faces conflicting pressures as industry leaders and politicians debate slowing development. Goldman Sachs projects that five major US hyperscalers will invest $800 billion into AI infrastructure this year. Meanwhile, SoftBank recently initiated a bond sale of $10 billion and €1 billion to fund its OpenAI investments, following a previous $54.6 billion commitment by the end of July.
Proposals for a development pause, supported by Anthropic CEO Dario Amodei and executives from OpenAI, Google DeepMind, and xAI, aim to align progress with safety research. Conversely, President Donald Trump and leaders like Nvidia’s Jensen Huang advocate for rapid growth to maintain competitive advantages. On September 29, a voluntary safety accord was signed, focusing on internal controls and audits rather than a collective halt.
Economic data underscores the stakes of these decisions. A January St. Louis Fed analysis indicated that AI-related investment drove 39% of real GDP growth in the first nine months of 2025. The IMF warned in April that an investment reversal could trigger a 20% decline in equity markets and reduce US GDP by 1.5%. Fitch projections suggest that a 35% equity shock could lead to a recession.
The Bank for International Settlements (BIS) has cautioned that AI infrastructure investment is currently 1.5 times higher than the socially efficient level, warning that such booms can lead to disruptive busts. Despite these concerns, UBS maintains its 2027 AI industry capital expenditure forecast at $1.2 trillion, representing a 33% increase from the $900 billion estimated for this year.
