# What exactly did Anthropic and OpenAI say about slowing down?

According to Bloomberg Technology, Anthropic PBC and OpenAI have both made calls to "tap the brakes on artificial intelligence development." The Bloomberg report, published September 14, 2026, frames these calls as a direct collision course with "forces in the tech industry, financial markets and the Trump administration that are likely to resist any substantial slowdown." The source material does not specify the exact wording or venue of the slowdown calls, but the framing matters. These are not academic papers. They are public positions from the two most valuable private AI labs in the world, both of which have raised tens of billions of dollars on the promise of rapid capability gains. When Anthropic and OpenAI say "slow down," they are talking to their own investors as much as to regulators. My read: this is a pre-emptive reputational move. By staking out a safety position now, both labs can claim foresight if a major AI incident occurs. If no incident occurs, they can quietly accelerate and point to competitive pressure from Google DeepMind or xAI as justification.

# Why is the Trump administration likely to resist a slowdown?

The Bloomberg report explicitly names the Trump administration as a force "likely to resist any substantial slowdown." That is a direct statement of political alignment. The current White House has positioned itself as pro-innovation, anti-regulation, and explicitly competitive with China on AI. Any public call to slow down from American labs cuts against that posture. Bloomberg Technology reported that the resistance comes from three directions: the tech industry, financial markets, and the Trump administration. That triangulation is important. It means Anthropic and OpenAI are not just facing skeptical regulators — they are facing their own ecosystem. The political math is simple. If Anthropic or OpenAI genuinely slow down, Chinese labs like DeepSeek or Alibaba's Qwen team fill the gap. The Trump administration will not accept that outcome. So the administration's likely response is not to mandate acceleration, but to ignore or publicly dismiss slowdown calls while continuing to fund and deregulate the sector.

# How do Wall Street and private markets factor into this?

Bloomberg Technology reported that "financial markets" are among the forces likely to resist a slowdown. This is the most underappreciated part of the story. Anthropic and OpenAI are private companies, but they are backed by public-market giants: Microsoft, Google, Amazon, and Nvidia. Those companies report quarterly earnings. AI revenue growth is now a material driver of their stock prices. Any public signal that Anthropic or OpenAI is slowing down would be read by Wall Street as a demand signal for AI compute — and that would hit Nvidia, Microsoft, and Amazon directly. According to Bloomberg Technology, the resistance is not hypothetical. It is structural. The capital that funds Anthropic and OpenAI comes from institutions that need exponential returns. A slowdown call is, in effect, a call against those institutions' business models. This is why I think the safety rhetoric is a hedge, not a handbrake. Anthropic and OpenAI can say "slow down" because they know the market will not let them. The statement costs them nothing and buys them reputational insurance.

Dimension	Anthropic PBC	OpenAI
Public safety posture	Constitutional AI, interpretability focus	Preparedness Framework, red-teaming
Primary capital backers	Amazon, Google	Microsoft, Nvidia
Slowdown call framing	Existential risk, caution	Safety-first deployment
Political exposure	Moderate — less consumer-facing	High — ChatGPT is a household name
Competitive pressure	Google DeepMind, Meta	Google DeepMind, xAI, Anthropic
Verdict	More credible on safety, but smaller	More exposed to political whiplash

# What does this mean for AI developers and researchers?

The practical consequence is a widening gap between what labs say and what they ship. Developers building on Anthropic or OpenAI APIs should not expect a slowdown in model releases. If anything, the competitive pressure from Google DeepMind and open-weight models like Llama will force both labs to ship faster, not slower. The real risk is regulatory whiplash. If the Trump administration publicly dismisses slowdown calls, and then a major AI incident occurs — a bioweapon synthesis, a critical infrastructure breach, a mass disinformation event — the political pendulum could swing hard. Developers who built on permissive APIs could face sudden compliance requirements. According to Bloomberg Technology, the tension is already live. The September 14, 2026 report is not a prediction. It is a description of a standoff that is happening now.

# Predictions

1. By Q1 2027, the Trump administration will issue an executive order or agency guidance that explicitly prioritizes AI competitiveness over precautionary slowdowns, effectively dismissing Anthropic and OpenAI's safety calls. 2. By Q2 2027, either Anthropic or OpenAI will publicly revise its slowdown position, citing competitive pressure from Google DeepMind or Chinese labs. 3. By Q3 2027, at least one major AI incident (defined as a publicly reported harm involving a frontier model) will trigger congressional hearings where Anthropic and OpenAI's safety statements are used as evidence of inconsistent commitment.

# Article Summary

• Anthropic and OpenAI's slowdown calls are a reputational hedge, not an operational constraint, because their funding and political support depend on acceleration.
• The Trump administration's pro-competitiveness posture makes any genuine slowdown politically untenable for US labs.
• Wall Street's AI revenue expectations are structurally incompatible with safety-first slowdowns, creating a permanent tension.
• The biggest winners from this standoff are Google DeepMind and xAI, which can position themselves as fast movers without the safety baggage.
• The biggest losers are AI safety researchers, whose credibility depends on the leading labs actually meaning what they say.