Antseed has launched a peer-to-peer AI inference marketplace that gives developers more affordable, flexible, and decentralized access to AI models.

The launch comes alongside a $2.4 million token round for the nonprofit Antseed Foundation led by Spark Capital, with participation from Collider, DCG, North Island Ventures, Reciprocal Ventures, Relay Capital, and Venice.ai.

The company says its open marketplace can provide access to leading AI models at prices up to 97% below official API rates by allowing independent providers to compete directly for inference requests.

A Decentralized Marketplace for AI Inference

AI inference is the process by which a trained artificial intelligence model receives a request and generates a response. As AI-powered applications, coding platforms, and autonomous agents become more widely deployed, inference is becoming a significant operating expense for developers and businesses.

Antseed aims to address that cost by creating an open marketplace where providers can offer AI inference capacity and set their own prices.

Instead of routing requests through a centralized platform, Antseed uses a locally hosted router that connects developers with hundreds of participating providers. Users can configure routing based on factors including price, latency, model capability, provider reputation, and privacy preferences.

Providers can operate models on their own computing hardware or offer access to compute capacity they already control.

“What BitTorrent did for files, Antseed is doing for AI inference, with no central server, and no company deciding who gets access or what it costs,” said Antseed co-founder Amos Meiri.

AI Inference Costs Continue to Grow

Antseed is entering the market as spending on AI infrastructure accelerates.

According to Gartner, worldwide spending on AI-optimized cloud infrastructure is projected to increase 96% to $42 billion in 2026. That growth is making the infrastructure connecting developers with AI models increasingly valuable.

Developers typically must either maintain separate accounts and billing relationships with individual AI providers or use centralized model aggregators that manage access to multiple services.

Antseed is proposing another option: a decentralized network where providers compete to process individual requests.

Providers are verified for the models they claim to offer. Those that fail verification can see their reputation scores decline, reducing their priority within the network.

Four Features Behind Antseed’s AI Marketplace

Antseed says its network is built around four primary advantages.

First, providers compete on pricing rather than operating under a single centrally determined rate structure. Developers can therefore route requests toward providers that meet specific budget and performance requirements.

Second, the users do not need a central account, email address, or platform-issued API key to route requests. The router operates locally, and developers can select TEE-verified providers when hardware-based prompt confidentiality is available.

Third, the marketplace gives developers broader control over their choice of AI models and providers, rather than letting one gateway determine which services can be accessed.

Finally, Antseed operates on a pay-per-request model rather than requiring users to commit to subscription tiers.

Antseed said the network had processed nearly 150 billion tokens across 202 active sellers, with more than $285,000 settled through September 2026.

Creating an Open Market for AI Compute

Spark Capital Co-Founder and General Partner Santo Politi compared inference costs to fuel costs for airlines, arguing that AI companies increasingly need greater control over one of their most important operating expenses.

Antseed’s model could become especially relevant as autonomous AI agents dramatically increase the number of model calls applications generate.

By creating a peer-to-peer AI inference marketplace, Antseed is betting that competition among independent compute providers can lower costs while giving developers greater control over performance, privacy, and infrastructure choices.

Developers who are facing rapidly increasing AI infrastructure expenses, that combination could make decentralized AI inference an increasingly attractive alternative to traditional centralized platforms.