Can the US and China verify an AI deal? Start-ups are starting to build the tools

The focus shifts from diplomatic trust to technical capability. AI verification tools aim to provide external checks on AI models and infrastructure. This bypasses the need for direct access to proprietary systems. The core challenge remains scaling this technology to monitor vast computing facilities, which can house hundreds of thousands of GPUs.

For Asia, this effort points to increased scrutiny on major AI players. Chinese companies like Tencent and Alibaba, which operate large-scale AI infrastructure, would face new verification demands. This could push them to adopt transparent development practices earlier than expected. The move could also foster a new sub-industry for AI auditing services across the region.

The thing to watch is the adoption rate of these verification tools by governments. If the US and China agree to a framework, it would accelerate development significantly. The test for this technology is its ability to provide reliable, real-time data on AI models without compromising intellectual property. This will be critical for any bilateral agreement.