Spheron
Founded 2025 · 4 employees on LinkedIn · 5 known investors
Spheron operates a GPU cloud marketplace that aggregates compute capacity from certified data centers worldwide, offering NVIDIA GPUs (H100, H200, B200, A100) at 40-60% below hyperscaler pricing. The platform serves AI training teams and ML builders with on-demand GPU instances, reserved capacity, and custom cluster sourcing across multiple providers through a single dashboard.
Also known as Spheron Network
Founders & leadership
Spheron was founded in 2025 by Prashant Maurya.
Investors · 5
Company profile
researched Aug 2026Spheron operates an enterprise GPU rental marketplace that aggregates accelerator capacity from certified Tier 3/4 data centers globally and exposes it through a single account and dashboard. Users can deploy on-demand or spot instances across multiple underlying GPU cloud providers without separate contracts or vendor lock-in, with unified billing and cost tracking across providers and a single SLA standard applied to all suppliers. The company advertises more than 50 GPU models, a 99.9% uptime SLA and deployment in under 60 seconds.
The catalogue spans NVIDIA Ampere through Blackwell and announced Rubin hardware: A100 80GB, H100 80GB, H200 141GB, GH200 96GB, B200 192GB, B300 288GB, GB200 192GB, GB300 288GB, RTX PRO 6000 96GB, RTX 5090, L40S, RTX 4090, and R100 (Rubin, 288GB) listed as shipping in the second half of 2026 on a reservation basis. Instances are billed per minute with no commitment, and both spot and on-demand rates are published; as of April 2026 listed on-demand rates included $1.43/hr for A100, $2.65/hr for H100, $4.79/hr for H200 and $7.20/hr for B200, with spot discounts of roughly 17-49% depending on model. Deployments offer full root access with VM or bare-metal options and, depending on the provider, InfiniBand connectivity for multi-node distributed training.
Beyond self-serve instances, Spheron offers reserved capacity with duration commitments at lower rates and a sourcing service for custom clusters of 8 to 512 or more GPUs with specific hardware and InfiniBand configurations. In that workflow the company matches the requirement to a provider in its certified data center network, negotiates pricing and handles setup, with a stated typical turnaround of 24-48 hours.
Business model
Spheron resells and brokers GPU compute capacity sourced from third-party certified data centers, charging per-minute on-demand and spot usage, offering discounted reserved capacity for duration commitments, and arranging custom cluster deals as a supplier matchmaking service.
Usage-based hourly (billed per minute) charges for on-demand and spot GPU instances, plus reserved-capacity contracts at negotiated rates; payment accepted via stablecoins such as USDC and USDT as well as enterprise invoicing.
Traction
The company site lists customer case studies with Prem AI (H200 rental), Wafer and Compute Desk, and displays as users or partners Baseten, io.net, Replika, Exo Labs, Gonka, Eigen Labs, Humanify, Libermans Co, Cognichip, NeevCloud, Hosted.ai, Fluence, Stanford University, UC Berkeley and the University of Washington. Prem AI's CTO describes Spheron sourcing H200 capacity from a vetted Tier 3+ partner with long-term commitment terms negotiated on the customer's behalf.
▸Full profile — market position, technology, go-to-market, geography
Market position
Positions itself as a lower-cost alternative to hyperscalers and dedicated GPU clouds, stating that aggregating supply directly from certified data centers yields enterprise GPUs at 40-60% less than hyperscaler pricing. Its published April 2026 annual H100 on-demand comparison lists Spheron at $14,400 versus CoreWeave at $18,000 (25% higher) and AWS at $39,456 (174% higher).
Multi-provider access from one account with no vendor lock-in, a unified SLA and consolidated billing across suppliers, published transparent spot and on-demand rates billed per minute, breadth of hardware including current Blackwell parts and pre-reservation of next-generation Rubin GPUs, and a brokered sourcing service for large or unusual cluster configurations.
Technology
An aggregation layer over multiple GPU cloud providers and certified Tier 3/4 data centers, providing a single dashboard for provisioning, unified billing and cost management, a common SLA across suppliers, sub-60-second deployment, full root access on VM or bare-metal instances, and optional InfiniBand fabrics for multi-node training where the underlying provider supports it.
Go-to-market
Self-serve sign-up and instant deployment for on-demand capacity, combined with a sales-assisted track: prospective customers submit GPU requirements, timeline and scale, and Spheron sources, negotiates and configures capacity from its data center network. Case studies and customer logos are used as proof points on the company site.
Teams training and fine-tuning large language models across multiple providers, startups needing immediate GPU access, research labs running multi-node distributed training, production AI inference deployments requiring high uptime, and cost-optimization teams comparing GPU providers. Named users and partners cited on the company site include Prem AI, Wafer, Compute Desk, Baseten, io.net, Replika, Exo Labs, Stanford University, UC Berkeley, University of Washington, Gonka, Eigen Labs, Humanify, Libermans Co, Cognichip, NeevCloud, Hosted.ai and Fluence.
Geography
Draws on a network of certified Tier 3/4 data centers worldwide; pricing and availability are noted as varying by region.
Compiled by commissioned research from 1 cited public sources — announcements, filings, and press listed under research sources below.
Key figures
latest reportedCompany-reported or press-reported figures, each dated to when it was claimed — not independently audited.
In the news
▸Research sources · 1
primary sources listed
- Spheronspheron.network · web
1 public sources were cited for this profile; the first-party ones are listed here.
Frequently asked questions
- What does Spheron do?
- Spheron is a multi-provider GPU cloud marketplace renting NVIDIA H100, H200, B200 and other accelerators on demand.
- Who founded Spheron?
- Spheron was founded by Prashant Maurya in 2025.
- Who are Spheron's investors?
- Spheron's investors include iSeed, Arcanum Ventures, Nexus Venture Partners, Protocol Labs, Z47.