In May, Catalyst Cloud hosted events in Wellington and Auckland, titled “Look to New Zealand: The State of Aotearoa Cloud in 2024.” This three-part series explores the presenters key insights and future potential of New Zealand's cloud technology.

After discussing the state of cloud in New Zealand (part one) and its application at Catalyst Cloud (part two), the question must be asked “how are these advancements being utilised by New Zealand-based companies?” We had fantastic speakers from local New Zealand-owned businesses who wanted to answer this question: Finlay Thompson and Henry Macdonald from Dragonfly Data Science, as well as Kandy Wahanui-Peters from the soon-to-be-launched Tohu Media.

Dragonfly Data Science: Embracing Cloud for Data Innovation

Dragonfly is a data science company that’s been innovating in the New Zealand market since 2008. Finlay and Henry witnessed some of the changes mentioned by Don and Paul first-hand. Years ago, running algorithms on large datasets with underpowered desktop computers was slow and painful, and often just not practical. Equally, building a high-powered computing network was expensive and tough to manage. Cloud computing turned out to be the game-changer for Dragonfly - it allowed them to run models on powerful VMs for short periods, making data science much more efficient and cost-effective.

Multi-cloud solution

For Dragonfly, cloud computing is a no-brainer. They use both AWS and Catalyst Cloud, picking the best cloud setup for their applications' needs. But, for Finlay, computing was one just one part of the process and encourages other businesses to adopt cloud-based mentalities too. For Finlay, this was CI (Continuous Integration). CI means repeatedly and constantly integrating new code into your existing code base, tearing down and rebuilding as needed.

Empowering developers is another big deal

By ditching the old-school, slow approval processes for computing resources, developers can run the VMs they need whenever they need them. This approach is faster and more cost-effective, letting developers work without unnecessary delays.

Dragonfly’s use of GPU-backed instances from Catalyst Cloud to run AI/ML algorithms quickly and efficiently shows how well this approach works. Running on Catalyst Cloud turned out to be 43% cheaper and 32% faster than using equivalent AWS instances. Plus, running local Large Language Models (LLMs) in Aotearoa offers predictable, lower costs, the ability to fine-tune datasets, and ensures complete data sovereignty.

Dragonfly is a perfect example of how a New Zealand company can lead in cutting-edge tech, delivering great results for clients and developers alike.

Check out the full case study for Dragonfly using Catalyst Cloud.

Tohu Media: Data Sovereignty and Legal Jurisdiction

Tohu Media has a clever way of using cloud technology while keeping everything local. Kandy Wahanui-Peters, Technical Founder of Tohu Media, highlighted the importance of data sovereignty but stressed that it's not the only reason to keep applications running in New Zealand. Legal jurisdiction over data is paramount. For Kandy, data pertaining to New Zealanders, especially Māori, should only be subject to New Zealand laws. This data is considered taonga (treasure), needing protection and must never be controlled by global corporations.

In a nutshell the Tohu solution works like this:

• Upload video content to the Tohu platform.
  
• A series of locally-hosted AI models get uploaded to Catalyst Cloud.
  
• The models run their processes, like speaker recognition, speech-to-text, Local Contexts assessment, and more.
  
• The models are spun down, and the data is downloaded and removed from the cloud.
  

This setup ensures two key things:

1. The data and models aren’t stored away from the local server for long periods.
   
2. Resources are only used and paid for when needed.
   

If we think back to Don’s presentation on Federated IT models, it sounds like Tohu Media is putting this into practice. They run GPU servers when needed for tasks that can’t be handled on local infrastructure, while mixing this with locally-hosted data and AI models. This respects the data’s taonga and context.

This way, New Zealand customers get the best of both worlds: the scalability and technical advances of the cloud, combined with a Federated IT approach that keeps Data Sovereignty paramount. This is not just about Māori Data Sovereignty, it’s about building the technological sovereignty of New Zealand.

The federated cloud approach is achievable

Cloud technology development hasn’t stopped, and the responsibility hasn’t just been handed over to US-owned businesses. Instead, New Zealand can take cues from others on decentralising IT—like the Federated IT approach Don spoke about—and put Kiwi businesses at the forefront of innovation. This innovation is cutting-edge, AI-powered, and driven by Data Science. It respects what makes New Zealand unique and honours our status as Treaty partners. It’s about taking global tech advancements and applying them within a New Zealand context.

What do we see when we Look to New Zealand?

By choosing local cloud providers, you can adopt a federated cloud approach – to all the New Zealand businesses reading this, the tools and support you need to innovate, secure your data, and lead the global market are already in New Zealand and right at your fingertips.