The Hardware Bottleneck AI Can’t Fix

Hardware engineering is trapped in a pre-software era of slow, manual processes—where testing an aircraft can take days just to retrieve data from a hard drive, and a single failed rocket test can cost tens of millions. Jason Haack, CEO of Nominal, argues that the real bottleneck isn’t engineering talent or materials, but the absence of software-like tooling: observability, CI/CD, and real-time feedback loops. Nominal’s platform tackles this by managing the entire hardware data supply chain—from high-frequency sensor streams and video feeds to post-test analysis and live control room dashboards—enabling teams to treat physical assets like 'pets' rather than 'cattle.' The result? A feedback loop that’s still orders of magnitude slower than software, but one that’s now being accelerated by AI agents, data lineage, and intent-based UIs. Yet, as Haack warns, we’re not ready to let AI design rockets—yet. The real frontier isn’t automation, but cultural and technical alignment: building systems where data is shared, trusted, and interpretable across every phase of development, from design to manufacturing to operations.