The numbers that should have triggered the alarm
Soul Machines entered receivership on February 5, 2026, with KPMG’s Leon Bowker and Luke Norman appointed to run a sale process. The Auckland-founded AI company had raised more than US$135 million, roughly NZ$225 million, across six funding rounds from 11 investors including Li Ka-shing’s Horizons Ventures, Salesforce Ventures, and Singapore sovereign wealth fund Temasek.
By March 2023, UK financial filings showed just $12.3 million in cash against $38.1 million in net outflows for the year. At that burn rate, the company had roughly four months of runway. The UK entity was deregistered in September 2024, a full 18 months before the formal receivership announcement.
No revenue figures were ever publicly disclosed. CB Insights lists no revenue data in available filings. For a company that had been operating for seven years, that silence is its own answer.
Everyone left before the lights went out
The workforce tells the story in miniature. Headcount fell from 253 in July 2023 to 70 by July 2024, and to just 45 immediately before receivership, an 82 percent reduction from peak. Both founders were gone before the end: CEO Greg Cross resigned in September 2023, and co-founder Dr Mark Sagar, the double Oscar winner whose facial animation work for Hollywood underpinned the company’s technology, stepped down as director in June 2024.
The marquee customers followed. Mercedes-Benz, ANZ, and Air New Zealand all abandoned Soul Machines’ technology. These were not peripheral accounts. They were the names that justified the valuation and attracted subsequent funding rounds. When they walked, the commercial case walked with them.
The moat that vanished overnight
Soul Machines was founded in 2016 as a spinout from the University of Auckland, building emotionally responsive digital humans for enterprise customer service. At the time, it was genuinely differentiated. Then OpenAI, Google, Microsoft, Amazon, and Anthropic arrived, and the market for AI-powered conversation became something you could build on top of an API for a fraction of the cost.
Tim Warren, founder of competing AI chatbot firm Ambit, put it plainly: “They had unique IP and technology, but suddenly they’re in a crowd with everyone else.” Warren noted the company had shifted from “a very high-end, very expensive offering” to consumer-friendly pricing but “didn’t appear to have landed anchor customers, or enough of them, to justify its valuation.”
Clare Capital managing partner Mark Clare framed it as a structural problem across the sector: “Companies who’ve had a credible standalone product for some time now find their offerings commoditised by the bigger players.”
The final indignity is illustrative. Before receivership, Soul Machines had pivoted to selling avatars built on top of ChatGPT. A company that raised $225 million on the promise of proprietary AI ended up as a reseller of someone else’s.
The lesson NZ’s tech ecosystem needs to absorb
The optimistic read, offered by The AI Corner newsletter, is that the talent and IP will not vanish: “There’s now even more great talent on the market. We’ll probably see a few of them start their own ventures.” That is true as far as it goes.
But the harder question is whether NZ’s deep-tech ecosystem has other companies running the same playbook. Large capital raises, impressive underlying technology, thin commercial revenue, and a narrative that outpaces the business. The venture capital model tolerates high failure rates, but $225 million is not a seed round that did not work out. It is a sustained, decade-long commitment of capital that produced a company with no publicly disclosed revenue and a cash position that was visibly terminal years before receivership.
The warning signs were not hidden. The UK filings were public record. The founder departures were public. The customer losses were public. The workforce collapse from 253 to 45 was public. KPMG’s statement acknowledged the company had made “the difficult decision to enter receivership after careful consideration of the company’s current financial position”, but careful consideration should have started when the burn rate exceeded the bank balance by a factor of three.
For NZ founders chasing AI-adjacent venture capital, the lesson is not that ambition is wrong. It is that revenue is not optional, proprietary technology is only a moat until someone bigger builds a bridge, and the market will eventually demand you show receipts, not just pitch decks.
Sources
- NZ Herald: AI casualty: Once high-flying Soul Machines in receivership, KPMG looks for buyer (2026-02-05)
- BusinessDesk: UK filing offers glimpse of Soul Machines’ financials (2024)
- NBR: Soul Machines in receivership (2026-02-05)
- BusinessDesk: What Soul Machines and Futureverse teach us about disruptive tech, hype and reality (2026)
- The AI Corner: Big-Funded NZ AI Startup In Receivership (2026)