I was delighted when I learned that Thrive 2050 had been accepted into NVIDIA Inception.
The programme had been recommended to us by someone within the technology industry, so it was one we genuinely wanted to join. Being accepted is an encouraging step for Thrive and gives us access to a community, knowledge base and set of technical resources that can help us explore how artificial intelligence could be applied to carbon project development.
More than anything, it gives us greater confidence in the direction we are taking—and better tools with which to pursue it.
What NVIDIA Inception gives us
NVIDIA Inception is a programme designed to support technology startups as they develop and grow. Members can access technical training, developer resources, tools, partner offers, community activities and networking opportunities.
For us, the practical value lies in being closer to the people, research and technologies shaping the development of AI.
We are already receiving information about new research, technical developments and opportunities within the programme. Over time, that should help us learn faster, make better-informed technical decisions and understand which technologies are most relevant to the problems we want to solve.
It also connects us with a wider community of startups working on real-world applications of AI. That matters because many of the questions we are considering will benefit from knowledge and experience beyond our own company.
What we mean when we talk about AI
Public discussion about AI currently focuses heavily on generative AI and large language models. These technologies are important, and they may have useful roles within Thrive, but they are only part of what we mean when we talk about AI.
We are thinking more broadly about machine learning and other computational approaches that can help us interpret information, identify patterns and understand data at scale.
That is particularly relevant to carbon projects.
Project developers already work with information from many different sources. This can include field observations, maps, monitoring reports, satellite imagery and remote sensing. As sensors and connected devices become more widely used, the amount and frequency of that information will continue to increase.
Collecting more data is useful, but only if we can understand it and act upon it. At a certain scale, relying on people to manage and interpret everything manually becomes slow, expensive and increasingly impractical.
That is where we see a valuable role for AI.
Making more ambitious questions practical
The possibility I find most exciting is future scenario modelling.
A project developer might want to understand how a project could develop under different conditions, what effect a particular intervention could have, or how decisions made today might influence outcomes over the next five, ten or twenty years.
Carbon would be only one part of that picture. A project may also need to consider biodiversity, water, resilience, permanence, community outcomes, financial viability and other environmental or social benefits.
Any one of these variables can be modelled. The difficulty comes when they need to be considered together, updated as the project changes and revisited whenever a new decision must be made.
A spreadsheet or one-off model might answer a particular question at a particular moment. Six months later, however, the data may have changed, the original assumptions may need to be reconstructed, and much of the analysis may need to be performed again.
We see an opportunity to turn that kind of one-off work into a continually evolving capability.
Over time, this could begin to resemble a digital twin: a digital representation of a real project that is updated using observations from the physical world. Digital twins can bring together monitoring, simulation and forecasting to help people understand the current state of a system and explore possible future outcomes.
For a carbon project, that could eventually mean combining project records with monitoring data, remote sensing, environmental observations and other evidence. A developer could then compare scenarios and ask complex questions that would be difficult to investigate repeatedly today.
That is a longer-term direction, but it is an exciting one.
From one-off analysis to a living capability
A useful model will never simply be built once and forgotten. It must be tested, maintained and updated as new evidence becomes available. Project conditions change, methodologies develop and models need to be checked against what is actually happening.
The opportunity is to manage that continuing work through a reusable platform, rather than requiring every project developer to assemble the same specialist capability independently.
This could make sophisticated analysis more practical for a wider range of organisations. It could also allow teams to return to important questions throughout a project’s life, instead of treating modelling as an isolated exercise undertaken only when time and budgets allow.
AI cannot remove all uncertainty from natural systems or predict the future with complete confidence. Its value is in helping us make uncertainty more visible, compare possible outcomes and recognise patterns that might otherwise be missed.
The resulting information should support professional judgement, not replace it. Deterministic calculations, carbon methodologies and human expertise will continue to be essential. The opportunity is to give the people making decisions a broader and more useful evidence base.
Creating a virtuous cycle
At the heart of this is a simple idea: better information can support better decisions.
Better decisions can improve the prospects of stronger project outcomes across carbon, biodiversity, water and wider social or environmental benefits. Stronger projects can create greater environmental and commercial value. That value can then help support further projects, creating a virtuous cycle in which successful work enables more work to take place.
This is why the ability to model future scenarios feels potentially transformative. The objective is not merely to complete an existing calculation more quickly. It is to make new and more complex questions practical to ask—and practical to revisit as a project develops.
That could help project developers understand risks earlier, evaluate possible interventions and make better use of the growing volume of information available to them.
Where we go from here
Joining NVIDIA Inception gives us an opportunity to develop this thinking within a community that understands the technologies involved.
We now have access to resources, training and technical knowledge that can help us turn broad ideas into focused experiments. From there, we can learn what works, what is technically realistic and where these capabilities could create the greatest value for carbon project developers.
I am genuinely excited that Thrive has been accepted into the programme. It is an important step for us, but it is also the beginning of a much more interesting piece of work: exploring how AI can help people understand complex projects, make better decisions and ultimately create more positive environmental impact.