Thunder Compute Analysis: $500K Raised
What is Thunder Compute?
Thunder Compute provides cloud GPU servers with NVIDIA H100, A100 80GB, L40, and RTX A6000 options. The company says it uses proprietary GPU virtualization software to add usable capacity across cloud, on-prem, VM, and bare-metal environments. Its site claims up to 3x capacity gains for enterprise fleets and up to 80 percent lower pricing than AWS.
Location
San Francisco, United States
Employees
11-50
Founded
2024
Revenue
$900K
Investors
- Y Combinator
Product Features & Capabilities
- Self-hosting AI/ML applications, NVIDIA Tesla T4 GPU ($0.27/hr), NVIDIA A100 (40GB) GPU ($0.57/hr), NVIDIA A100 (80GB) GPU ($0.78/hr), hosted in AWS/GCP, U.S. Central servers, 7+ Gbps networking, instance templates for common tools, easy CLI for instance management.
Pricing
- 1 x NVIDIA Tesla T4: $0.27/hr
- 1 x NVIDIA A100 (40GB): $0.57/hr
- 1 x NVIDIA A100 (80GB): $0.78/hr
How much Thunder Compute raised
Seed - $500,000
July 10, 2024Lead Investor: Y Combinator, Team Ignite Ventures, Avesta Fund
Who are the founders of Thunder Compute
CP
BM
Brian Model
CTO
Business Model
Subscription-based GPU cloud services with tiered pricing for AI and machine learning applications.
Market Focus
Thunder Compute primarily focuses on the cloud computing industry, specifically providing GPU cloud services for self-hosting AI and machine learning applications.
Competitors Overview
Thunder Compute operates in a competitive landscape with several notable competitors in the GPU cloud service market, particularly for AI and machine learning applications. The main competitors include:
-
CoreWeave: Offers managed Kubernetes, virtual servers, and storage solutions, focusing on sectors requiring intensive computational power. Its managed services provide advantages in scalability and operational efficiency.
-
Modal: Specializes in serverless infrastructure for running generative AI models and large-scale applications. Its serverless architecture offers flexibility and cost efficiency, particularly for dynamic workloads.
-
NexGen Cloud: Focuses on sustainable cloud IaaS, specializing in high-performance computing and GPU infrastructure. Its commitment to sustainability may attract environmentally conscious customers.
-
Deepstributed: Streamlines machine learning experiments by simplifying GPU resource utilization, enhancing the efficiency of machine learning workflows.
-
AWS, GCP, Azure, Oracle: Major providers known for robust storage solutions and integration with existing cloud infrastructure. They offer generous credits for startups but can be expensive and complex to set up.
-
Lambda Labs: Provides a mix of enterprise and on-demand services, excelling in large-scale AI projects, though at a higher price point.
-
TensorDock: A decentralized marketplace offering GPU cloud instances at lower costs, but may face reliability issues due to its crowdsourced nature.
-
RunPod: Offers a seamless user experience for container deployment at a lower cost than Modal, but with less reliable GPUs.
-
Vast.ai: A low-cost marketplace for renting GPUs, primarily container-based, but also faces reliability issues.
Thunder Compute differentiates itself by offering ultra-low costs (up to 75% lower than AWS or GCP) and a simple user experience through its GPU virtualization technology, making it particularly appealing for indie developers and small teams focused on research and development.
See something that needs updating? Suggest edits to this profile.