September 23, 2026

How do you spend $50.7 million on software and get nothing?

Portrait of a woman with facial recognition technology lines across her face on a plain background.

Seven years, fifteen managers, zero benefit

On 23 September 2026, MBIE chief executive Nic Blakely confirmed at select committee that $50.7 million of taxpayers’ money had been wasted on the abandoned Biometric Capability Update. The number is striking not because government IT projects overrun, they always do, but because of the gap between what was promised and what was delivered.

The project began in 2018 to replace ageing immigration identity systems. The original budget was roughly $10 million with a September 2020 completion date. By the time it was killed in December 2025, it had endured at least six formal change processes, 15 different managers in seven years, and more than a dozen deferred go-live dates. It delivered nothing usable.

And it may not be over. The government faces a potential $12 million liability from vendor NEC, which signalled it could charge up to $750,000 a month for missing the 2025 deadline.

The threshold that drove the whole mess

The structural culprit is worth understanding because it recurs in private-sector procurement too. Cabinet had approved a funding threshold of $35 million for whole-of-life costs. Exceeding it meant going back to Cabinet, an uncomfortable admission officials preferred to avoid.

So they didn’t. MBIE’s own review recorded staff describing financial management as ‘creative accounting’, with project budgets repeatedly merged and unmerged to obscure real spending. A total of $4.44 million was transferred out of the project between 2022 and 2025. The minister was kept in the dark entirely. In August 2026, RNZ reported that officials pressed on despite Immigration Minister Erica Stanford never agreeing to extra funding, and that two internal reports gave markedly different accounts of the project’s status.

The cost escalated in public view. In June 2026 MBIE told the minister the write-off was about $33 million. By August, Stanford revealed a further $6 million had surfaced, and said even that might not be the full extent. It has since climbed to $50.7 million.

When silence was always the safer bet

Oliver Hartwich, a Wellington think-tank director, offered the sharpest explanation in August 2026. “At each decision point, silence was safer,” he wrote. “Admitting failure created an identifiable loser, while carrying on spread the cost and accountability across budgets, years and successors.” He noted officials “gave misleading advice to two governments in succession”, so this is not a partisan failure.

The warning signs were flagged and ignored. A March 2024 internal update to Stanford noted that abandoning the project then would leave “sunk costs of approximately $25 million+, leaving Immigration New Zealand no further ahead than when the project started in 2019.” It ran for another 21 months before being cancelled, doubling the loss.

The contempt finding, and the absence of consequence

At a March 2026 select committee, Blakely was asked about the biometric programme and failed to disclose that it had already been cancelled. The Privileges Committee found the omission “was intentional” and that the ministry had committed a contempt of Parliament, adding that failing to tell the truth “is an insult to the entire public service.”

Blakely offered his resignation. Public Service Commissioner Sir Brian Roche declined it. On 23 September, Blakely told the committee: “I recognise that we have a mountain to climb to rebuild the trust and confidence of this committee and of the Parliament.”

Nearly half the big projects are failing

The scandal forced a stocktake of MBIE’s entire technology portfolio, released 22 September 2026. It covered 70 projects worth roughly half a billion dollars. Of 31 larger projects with formal controls, 9 rated ‘red’, 6 ‘amber’ and 15 ‘green’. Nearly half were failing or at risk. Deputy secretary Paul Stocks said he was “reassured by the fact that there were no surprises in these findings, but there are some things we can do better” – a sentence that reassures nobody.

Every failure mode here is recognisable in corporate IT: gaming approval thresholds, rotating owners, deferred launches standing in for honest status reporting. The difference is that in business the losses eventually crystallise on a balance sheet and someone answers for them. Here they were written off in a Budget line, and the people responsible kept their jobs.

Sources

Community

Join the discussion

Add useful context, ask a good question, or challenge an idea — keep it specific and respectful.

Create a commenter account

Enter the name you want shown publicly and your email. We will email you a password-set link; you cannot comment until you use it.

Your email is used for sign-in and account security. It is not published with comments.

Subscribe for weekly news

Subscribe For Weekly News

* indicates required