January 2026
Joined Nvidia Inception Programme for AI Startups
Doses AI has been admitted to the Nvidia Inception programme, which supports cutting-edge AI startups with hardware access, technical training, and go-to-market support. The programme gives Doses AI access to Nvidia's DGX Cloud infrastructure and preferred pricing on GPU compute.
About Nvidia Inception
Nvidia Inception is a free programme designed to help AI startups evolve faster through access to technology, expertise, and go-to-market support. The programme has nurtured over 18,000 startups globally, providing a path from early research through to scaled deployment. Members gain access to Nvidia's hardware ecosystem, deep learning training resources, and a network of venture capital partners.
For Doses AI, the programme addresses a core bottleneck in their development pipeline. Training ternary language models — where weights are constrained to {-1, 0, +1} — requires intensive matrix operations during the knowledge distillation phase. The continual pretraining stage of their distillation pipeline, which converts full-precision fp16 models into 1.58-bit ternary architectures, is computationally demanding and benefits directly from Nvidia's GPU acceleration.
Nvidia Inception supports AI startups with preferred pricing on Nvidia hardware and cloud instances, access to DGX Cloud for large-scale training, technical workshops with Nvidia engineers, and co-marketing opportunities. The programme is free to join and has no equity requirement.
Programme Benefits
Nvidia Inception provides Doses AI with resources that directly accelerate the development of the Doses Mortar model family:
- DGX Cloud access — Nvidia's fully managed cloud infrastructure for training large AI models, providing multi-node GPU clusters optimised for distributed training workloads
- Preferred GPU pricing — Discounted access to A100 and H100 GPU instances through Nvidia's cloud partners, reducing the per-experiment cost of training ternary model variants
- Technical training — Access to Nvidia's Deep Learning Institute courses and hands-on workshops covering CUDA optimisation, mixed-precision training, and TensorRT deployment
- Go-to-market support — Co-marketing opportunities and introductions to Nvidia's venture capital partner network for future fundraising