Category Definition: Novel algorithms, datasets, or architectures.
Examples in NZ: AI radiology model, freight optimisation engine, Pacific languages speech recognition.
Why it Matters: Many AI and software ventures fall into a grey zone when it comes to government R&D rules. If you’re creating new algorithms, datasets, or architectures — not just configuring existing tools — you can unlock the same Callaghan grants and R&D Tax Incentive as deep tech companies.
This pathway helps ensure you:
Do This First: Run “Is my R&D eligible?” test and then book a free session with your Regional Business Partner (RBP)
(⚠️ RBP registration is required for most grants.)
| Stage | What to Do | Programs & Links | Why It Matters |
|---|---|---|---|
| 1. Setup & Entry | Confirm R&D eligibility, formalise basics, connect with advisors | - “Is my R&D eligible?” test |
⚠️ Note: Several Callaghan programs (New to R&D, Career/Experience, Ārohia) are transitioning to MBIE administration in 2025. Always confirm availability before applying.
<aside>
Government R&D eligibility is about whether your work meets a legal/technical definition — solving scientific or technological uncertainties with a systematic approach.
Venture capitalists, on the other hand, use deep tech to mean something narrower:
Deep Tech (VC lens) = high-bar science and engineering companies (e.g. biotech, advanced materials, robotics, climate science, aerospace). They usually require long development timelines, capital intensity, and defensible IP.
R&D (Gov’t lens) = much broader. A novel AI algorithm, a new data-processing approach, or an advanced manufacturing process can all count as R&D — even if they wouldn’t be pitched as “deep tech” to VCs.
Why it matters:
<aside>
</aside>
Pathway 1: R&D YES Pathway 2: AI & Software – R&D YES Pathway 3: AI & Software – R&D NO Pathway 4: R&D NO
<aside>
</aside>
<aside>
</aside>
