Skip to product information
Core LLM Ai Supercomputer in a box (NVIDIA GPU CORE)
$100,000.00
YOUR OWN SUPERCOMPUTER + LARGE LANGUAGE MODEL (LLM)
We will build for your organization a core LLM ai supercomputer. This includes both the hardware and software. Capable of over a trillion calculations per second, and trained on over a Trillion total parameters. Bring the power of a full scale LLM into your organization, with Foundry Funnel' Supercomputer and Core LLM in a box.
For AI model training and fine-tuning, our NVIDIA solution is vastly superior. For inference or creative work, the Apple M4 Max's higher memory bandwidth and media accelerators may be advantageous.
Why run your own organizational supercomputer and core LLM?
- Complete control over custom training within your own organization. LLM's are based on predictive computing, meaning by training on your own datasets and scenarios, you can have LLM outputs tailored to as you like. (AKA, be better then anyone at the results you want)
- Data security (connect to your custom CORE LLM Foundry Funnel Supercomputer, within your own organization's firewalls and gap security protocols).
- Expandable as your needs grow, with industry leading ultra high-speed connections (you can link together multiple supercomputers in a box, and run them in tandem, for the ultimate supercomputer data solution.)
Hardware:
-
NVIDIA GB10 Grace Blackwell superchip
-
1 PFLOPS of FP4 AI performance
-
128GB of coherent, unified system memory
-
ConnectX-7 Smart NIC
-
4TB NVME.M2 with self-encryption
-
150mm L x 150mm W x 50.5mm H
Software:
We will load onto your core ai accelerated GPU powered supercomputer hardware, a Mixture-of-Experts model with 32 billion activated parameters and 1 trillion total parameters. It achieves state-of-the-art performance in frontier knowledge, math, and coding among non-thinking models. It goes further — meticulously optimized for agentic tasks, your custom LLM does not just answer; it can act.
-
Base: The foundation model, a strong start for researchers and builders who want full control for fine-tuning and custom solutions.
-
Instruct: The post-trained model best for drop-in, general-purpose chat and agentic experiences. It is a reflex-grade model without long thinking.
With Foundry Funnel, advanced agentic intelligence is more open and accessible than you could imagine. We can’t wait to provide you with your very own organization's own Ai accelerated custom supercomputer, with Core LLM included.
Related benefit: By using your organizations own ai supercomputer with custom core LLM, you no longer have to pay externally for tokens. You can generate your own, at a fraction of the cost, and with total control over security, content, context, and internal data.
Related benefit: By using your organizations own ai supercomputer with custom core LLM, you no longer have to pay externally for tokens. You can generate your own, at a fraction of the cost, and with total control over security, content, context, and internal data.