Neysa

Neysa offers enterprise-focused AI and ML infrastructure as a service, providing tailored solutions that include dedicated consulting teams to help clients optimize infrastructure and customize models.

Originally an infrastructure provider, Neysa launched its flagship platform, Velocis, in July 2024, which offers on-demand NVIDIA GPU resources for training, fine-tuning, and prompt engineering. With elastic scaling and integrations using GlusterFS and Slurm, Velocis allows for deployment of production applications and scaling as needed.

By the end of 2024, Neysa aimed to expand its offerings to include a developer platform and inference as a service, along with enhancing infrastructure management and AI workload security through advanced observability tools. 

Among its key products, Neysa Overwatch delivers full observability across networks to improve availability, security, and performance, reducing downtime through Neysa Radar’s ML Engine, which automates corrective actions. Meanwhile, Neysa Aegis protects AI models from security threats, integrating robust defenses against model attacks like model inversion and prompt injection and enabling real-time telemetry for proactive threat management.

Neysa distinguishes itself from competitors like CoreWeave and Lambda Labs by providing adaptable, flexible models to meet varied client needs.

Neysa’s GPU training infrastructure is priced at USD 0.37 per hour (for long term contracts).

Key customers and partnerships

As of October 2024, Neysa had roughly 12 paying customers. Around 70% of its entire customer base opted for private cluster use, while the remaining 30% was on a public cloud. Types of enterprise using the platform included research institutes and AI-native firms in the media, banking, and manufacturing sectors. At the time, the company planned to enter global markets with its next round of funding (expected to close in 2025).