Case study

Basecamp Research uses AI to deliver process-ready proteins faster

Basecamp Research leverages Ori's GPU Cloud to help them deliver more accurate structure predictions, more protein annotations and controllable sequence generation


Basecamp Research, a UK-headquartered, biotechnology research firm at the forefront of biotech and AI, required access to advanced computational resources to accelerate and scale their machine learning projects. They turned to Ori’s AI Native GPU Cloud to get cost-effective and predictable GPU compute access.

About Basecamp Research

Basecamp Research is a market leader in the development of proteins that meet the most challenging requirements in applications as broad as food, pharma, and bioremediation. They have sampled the most extreme and extraordinary biomes on the planet, to create a knowledge graph of natural biodiversity that is the most comprehensive and diverse in existence to deliver better medicines, better food and better products for the planet. 

They recently launched BaseFold, a breakthrough in 3D protein structure prediction of large, complex protein structures that leverages this purpose-built foundational dataset to significantly increase prediction accuracy of large, complex protein structures and small molecule interactions. BaseFold is up to six times more accurate than AlphaFold2 and offers up to a three-fold improvement in small molecule docking. Why is this awesome? Having more reliable 3D structure predictions for larger and more complex proteins will greatly accelerate AI-based drug discovery efforts!


The healthcare and life sciences industries have delivered remarkable advancements in drug discovery and patient therapies in recent years by incorporating machine learning techniques into their research. Many of these advancements have been driven by the ability to process new and more complete datasets, integrating more sophisticated and efficient machine learning models, taking advantage of the latest generative AI capabilities and of course, research labs getting access to more powerful computing infrastructure. 

Basecamp Research was no exception in being challenged with finding a partner that could deliver consistent access to the vast amounts of GPU computing power needed to train their complex machine learning models. 

“Training new models and building new architectures, particularly when your data is extremely large, is an extremely expensive business.”

Glen Gowers, CEO and Cofounder - Basecamp Research



Basecamp Research makes use of Ori GPU Cloud’s NVIDIA H100 GPU architecture deployed in a private cloud to train and benchmark their models. This has allowed their researchers to get dedicated access to the computing power they need, when they need it, while doing so in a cost-effective manner. 

Why use NVIDIA’s H100 GPUs for life sciences use cases?

Training life sciences models requires the processing of vast amounts of data and intricate simulations that can benefit immensely from NVIDIA’s H100 GPU. This processor is built on the cutting-edge Hopper architecture, boasting an impressive 80 billion transistors, and equipped with 80GB of HBM2e memory providing the bandwidth necessary to handle large datasets and complex computations quickly and efficiently. Compared to its A100 predecessor, the H100 offers significant upgrades in processing power and energy efficiency, making it ideal for the demanding needs of life sciences computations.

The GPU's enhanced Tensor Cores and support for FP8 precision accelerate AI workloads, enabling more rapid model iterations and faster time to insight. For instance, in molecular dynamics and protein folding simulations, these features can reduce computation times dramatically, allowing researchers to obtain results in hours (instead of days) with increased accuracy.


Basecamp Research is designing genomes and proteins with far greater novelty, controllability, and performance than the previous state-of-the-art. Their advances come from their unique, proprietary dataset of real organisms and the diligent care they take to teach AI as much as possible about the underlying biology. For example:

  • More accurate structure predictions than Google DeepMind's AlphaFold2, unlocking more reliable small molecule docking for larger and more complex proteins than ever before.
  • 40% more proteins annotated than all other state-of the art algorithms, including CLEAN and Google's ProteInfer, allowing us to discover and classify the most difficult dark matter sequences.
  • Controllable sequence generation that leverages our dataset's superior diversity, context, and quality to design proteins and genomic systems to best match our partners' desired function and performance.

What’s next?

  • Learn more about Basecamp Research and how they are rapidly developing AI that outperforms state-of-the art.
  • Sign up to Ori and provision your GPU computational needs on-demand, or speak with one of our experts to learn how to accelerate your life science initiatives.

Similar posts

Join the new class of AI infrastructure! 

Build a modern GPU cloud with Ori to accelerate your AI workloads at scale.