The infrastructure of intelligence — built, and backed.
Jirisan Ventures runs two businesses. We build NADI, on-premises observability for NVIDIA GPU fleets. And we invest our own capital, with a focus on AI and technology.
Jirisan Ventures runs two businesses. We build NADI, on-premises observability for NVIDIA GPU fleets. And we invest our own capital, with a focus on AI and technology.
NADI reads the health of every GPU in your fleet and shows it in one place — throttling, faults, idle capacity and energy cost — without sending data outside your data centre.
Utilisation, temperature, power and error events for every GPU, by node and rack.
NVLink and network link health, so a single bad link is found before it slows a training job.
Idle GPU-hours by node and tenant — capacity you are paying for and not using.
Power drawn and cost per GPU-hour in ringgit, for billing and for planning.
A light collector runs on each GPU server inside your data centre.
Telemetry is stored on your own hardware. Air-gapped sites are supported.
Operators see each fault with its shape, colour and plain-language label.
Jirisan Investments manages the company’s own capital, with a focus on AI and technology. We invest for the long term.
Jirisan is a mountain in Korea. We named the company for the patience it takes to make what cannot be seen knowable — the same work NADI does for a GPU fleet, and the same patience we bring to investing.
Jirisan Ventures Sdn. Bhd. · Kuala Lumpur, Malaysia
One inbox for NADI demos, partnerships and investment conversations.
enquiries@jiri-san.com