Blackouts and Robotaxis: What Real-World AI Failures Teach Us About Innovation

A recent power outage in San Francisco resulted in Waymo’s autonomous robotaxis stalling at intersections.

While the episode was widely interpreted as a technological failure, I see it as a systemic innovation issue. The vehicles behaved according to their design constraints; what failed was the broader innovation system in which they were embedded.

This incident reminds us that technical capability is only one component of high-velocity innovation. Infrastructure resilience, organizational design, governance, regulation readiness, and the adaptation of accountability systems, norms, and behaviors are equally necessary. Innovation, especially with AI, is as social and institutional as it is technological.

From this perspective, several foundational questions deserve attention when deploying AI in real-world environments:

1. Who is accountable when AI-enabled systems fail, not only individually, but through their interdependencies?
2. How do we govern situations where systems create new operational regimes outside “normal conditions” (for example, autonomous systems operating when traditional guardrails and infrastructure cease to function)?
3. How do we align responsibility models for distributed agency across humans, machines and infrastructures?

The introduction of non-human agents into human social systems is not merely a technical upgrade. It is a civilizational design challenge. It requires rethinking innovation processes as living systems, where learning, failure, adaptation, and responsibility must be intentionally and dynamically designed.

For innovators and leaders, the question is no longer whether AI systems can perform under ideal conditions. The real question is whether our innovation models can also absorb ambiguity, failure, and systemic stress.

Learning about AI strategy today means learning how entire systems respond to disruption, not just how models perform in controlled environments; from tools and models, to systems thinking, techno-anthropology, and governance literacy.

Innovation does not fail at the algorithm. It fails at the seams.

https://techcrunch.com/2025/12/21/waymo-suspends-service-in-san-francisco-as-robotaxis-stall-during-blackout/

hashtag#ArtificialIntelligence hashtag#AIStrategy hashtag#InnovationSystems
hashtag#SystemsThinking hashtag#ResponsibleAI
hashtag#Leadership hashtag#Learning hashtag#AIGovernance