The promise of 5G was never just about speed. It was about architecture — moving intelligence out of centralized, purpose-built telecom hardware and pushing it to the edge of the network, closer to where data is actually generated and consumed. That shift has created something genuinely new: a distributed, software-defined network infrastructure that enterprises and governments can slice, customize, and control in ways that were simply impossible with 4G. It has also created a security problem that the industry is still working through.

A question that keeps coming up in conversations we’re having with customers and at industry events: how do you actually secure the infrastructure behind autonomous robots? Not just the device — the whole backend. The control plane, the model serving pipeline, the OTA distribution, the telemetry ingestion. The part that decides what robots do in the physical world.

DC just published their Confidential Computing study, surveying 600 IT leaders across 15 industries globally. The findings are striking—but not surprising to those of us working in this space. 75% of organizations are already deploying or piloting Confidential Computing. Only 6% have no plans at all. This is no longer an emerging technology. It's becoming standard practice.

The promise of trustless computing was never about decentralization. It was about verification. We've been addressing the wrong issue. For the past decade, blockchain has led discussions about trustless systems. The idea was appealing: what if you could perform computations without a single party in control? What if you could remove the need to trust anyone? But the industry made a mistake: we confused decentralization with trustlessness. We thought that distributing computation across thousands of nodes was the only way to eliminate trust. It wasn't. This was just the first approach we explored.