The Operating System and the Machine Room
Why Chris Blair’s NZ-EOS framework is right about the system — and why Invercargill, not Auckland, shows what “getting it built” actually looks like
Chris Blair has done something genuinely useful.
His New Zealand Economic Operating System framework — published on his site in March — lays out nine pillars across three layers that describe the system New Zealand needs to become an intelligence-driven export economy. High-tech exports. Sustainable energy at scale. Local capital feedback loops. A research-to-industry bridge. Intelligence-native organisations. Sovereign data. An AI-ready workforce. Custom AI software. Innovation-driven boards.
A taxonomy that makes sense, aimed squarely at the right audience — directors, policymakers, investors, cross-agency leaders. And the core insight is one New Zealand desperately needs to internalise: the problem isn’t initiatives, it’s the absence of a coherent system. We’ve had a chronic disease of siloed policy, agency fiefdoms, and strategy documents that read like wishlists without wiring diagrams. Chris has tried to give us the wiring diagram.
I read through it nodding. Most of it landed.
And then I got to the energy pillar, and the pragmatist in me started asking uncomfortable questions.
A framework worth building on
Before I get into the critique, let me say what Chris has got right — because in a country where we love to tear ideas down before they’ve had a chance to breathe, it matters.
He’s treating this as a systems problem, not a project problem. That’s unusual for New Zealand. We are incredibly good at announcing “transformational initiatives” that live and die inside a single agency, a single minister’s tenure, or a single budget cycle. Chris is naming the thing that actually matters: the connective tissue between the pieces. The fact that high-tech exports can’t scale without local capital feedback loops. That sovereign data is meaningless without an AI-ready workforce to act on it. That innovation-driven boards aren’t nice-to-have — they’re the control surfaces for the entire machine.
He’s also treating Māori data sovereignty as a first-order concern inside the sovereign data pillar, not as a footnote or a political concession. That’s a deliberate choice, and it’s the right one. Any national framework that doesn’t embed Te Mana Raraunga into the core architecture isn’t a national framework — it’s a colonial transplant.
And he’s putting energy as economic infrastructure front and centre. In an AI economy, compute needs power, and New Zealand’s renewable generation is a strategic asset we’ve consistently undervalued. Chris is right to put it on the board.
So — thumbs up. Genuinely.
But here’s where I want to add my own lens. Because the NZ-EOS is a map of what should exist. It’s less clear on how we actually build it, who pulls it off, and what we stop doing to free up the capacity. That’s the gap I want to fill in.
The least worst choices problem
I’ve spent thirty-plus years in IT operations and enterprise architecture. One of the core lessons — and I’ve written about this before — is that real architectural decisions are rarely about “best practices.” They’re about the least worst choices given the system you actually have, not the one you wish you had.
Chris’s framework assumes a delivery capability that, in several places, we don’t possess.
Take the energy pillar. “Abundant, low-cost energy resilient to global disruption” is a beautiful sentence. It’s also sitting on top of an industry structure that was deliberately designed in the 1990s to promote competition rather than coordination. We have five major generators, a state-owned grid operator, twenty-nine regional distribution businesses — or fiefdoms, depending on how charitable you’re feeling — and four vertically integrated gentailers that between them control roughly 85% of the residential retail market.
That structure works fine for incremental investment. A wind farm here. A geothermal expansion there. Some solar. But ask it to coordinate the $100 billion decarbonisation investment that Transpower, BCG and the Infrastructure Commission all agree is needed by 2050, and it stalls. No single entity is responsible for generation adequacy. The 2024 winter crisis — wholesale prices over $800/MWh, hydro lakes critically low, industrial plants closing “permanently” according to Transpower — was the structure doing exactly what it was designed to do. It’s not broken. It just isn’t pointed at the problem Chris’s framework wants it to solve.
You can see this fragmentation in smaller, almost comic ways too. Retailers are busy cross-selling internet and mobile phones, yet can’t share basic data with each other to ensure a good experience for something as fundamental as power delivery. That’s not an energy problem. That’s an operating model problem. And no amount of high-tech export ambition fixes it.
