OneTriangle Analysis
What is OneTriangle?
OneTriangle provides an inference engine that uses frontier research on KV cache transfer between models to reduce inference costs by 20% and time-to-first-token by 40% versus current standards. It prefills on a small model and decodes on a large one, delivering large-model-quality answers faster and cheaper. The company is backed by Y Combinator (S26) and was founded by MIT CSAIL graduates with backgrounds at DeepMind and Jane Street.
Location
San Francisco, California, United States
Employees
1-10
Founded
2026
Industry
B2B, Infrastructure
AI Pattern
AI infrastructure / tooling
Builds Hardware
Software only
Business Model v4
Developer infrastructure
Delivery Model
Developer platform / API / infra
Extruct Category
DevTools
Founders LinkedIn
- Hannah Chung: https://www.linkedin.com/in/hannahchung1/
- Medha Venkatapathy: https://www.linkedin.com/in/medhav/
Sub-cluster
inference engine
YC Partner
Ankit Gupta
Product Description
OneTriangle offers an innovative inference engine designed to enhance the efficiency of AI model deployments. Their main product, the TrustAI inference service, utilizes cutting-edge research on KV cache transfer between models, resulting in significant improvements in performance metrics.
Key features of OneTriangle's offerings include:
- Cost Reduction: The inference engine reduces inference costs by 20%, making it a cost-effective solution for AI engineering teams.
- Speed Enhancement: It decreases the time-to-first-token by 40%, allowing for faster response times in applications.
- Model Prefilling and Decoding: The system pre-fills on a smaller model and decodes on a larger one, which enables the delivery of high-quality answers akin to those generated by larger models, but at a fraction of the cost and time.
- Managed Open-Weight Models: OneTriangle provides access to managed open-weight models like DeepSeek V4 Flash, which is priced at $0.15 for input and $0.35 for output, targeting AI engineering teams and ML infrastructure professionals.
The benefits of using OneTriangle's services include improved operational efficiency, reduced costs, and enhanced performance for companies deploying AI solutions. This makes OneTriangle a valuable partner for organizations looking to optimize their AI infrastructure.
See something that needs updating? Suggest edits to this profile.