Perceptron Raises $6.5M to Build a Decentralized AI Data Network With 800K+ Nodes

The bottleneck for AI training data is no longer just compute — it's access to high-quality, real-world datasets that aren't locked behind closed partnerships or scraped at diminishing returns. Perceptron, a Dubai-based decentralized AI data network, is trying to solve that problem with a global contributor mesh that already spans over 807,000 nodes and 300,000 daily active users. On July 30–31, 2026, the company closed a $6.5 million strategic funding round from 14–15 Web3 investors, trading firms, and infrastructure partners to scale its network and launch its data-questing platform .

Who Invested in Perceptron's $6.5 Million Round

No single lead investor was disclosed for the strategic round. According to multiple reports, the confirmed participants include :

• Sigma Capital
• Selini Capital
• QCP Capital
• P2 Ventures (formerly Polygon Ventures)
• CoinDCX Ventures
• Momentum6
• DeFi Capital
• Walrus Foundation
• Aethir
• Colosseum
• GuruDev Capital
• Tempo Finance
• NewTribe Capital
• Digital Consensus Fund
• CodeCraft Capital

Some sources list 14 investors, while others name 15. The variation appears to stem from whether NewTribe Capital is counted as a lead investor or participant . The company has not publicly disclosed its valuation .

How the Decentralized AI Data Network Works

Perceptron operates a decentralized "AI data mesh" that replaces centralized data scraping with a global contributor network. Anyone can run a Perceptron Node — available as a browser extension or Android app — that shares idle bandwidth and contributes compute resources to the network . The network is currently live with over 807,000 nodes and more than 300,000 daily active users across 150+ countries .

AI companies can commission specialized datasets through Perceptron's data-questing platform. They post structured data tasks tailored to their model-training needs, and contributors complete these tasks — providing unique datasets, domain expertise, or bandwidth — in exchange for token rewards . Contributors retain ownership of their data and monetize it directly through the network, avoiding the centralized "scrape-and-sell" model .

Perceptron claims this model reduces data-acquisition costs by ~90% compared to traditional centralized providers . The contributor community includes doctors, lawyers, native-language speakers, and other domain experts who can contribute knowledge and real-world data .

Scaling Targets: Near-Term and Long-Term

The company's near-term priority is scaling the network toward 5 million nodes by late 2026 . The new capital will fund three specific priorities :

• Launching the data-questing platform at scale
• Expanding contributor tools and reward infrastructure
• Supporting network growth toward 5 million nodes

No specific hard numerical long-term targets (such as "X million nodes by 2028") have been publicly disclosed beyond the 5-million-node milestone . The company's stated long-term goal is to become the default decentralized data layer for AI training and inference by expanding the contributor mesh across idle bandwidth, unique datasets, and domain expertise .

Traction So Far

Perceptron's growth trajectory shows strong early demand. The network grew from 200,000 daily active users in its initial phase to more than 300,000 — a 50% increase . Node count similarly rose from 700,000+ to 807,000+ over the same period, with the company reporting activity across more than 150 countries .

The network merged with BlockMesh in June 2025, combining Perceptron's infrastructure with BlockMesh's network of over 700,000 devices gathering and structuring public web data .