VRAM AI gets $10 million GEM facility to launch Sui-based VRAM Network

By AI, Created 4:50 AM UTC, May 25, 2026, /AGP/ – VRAM AI secured a $10 million Token Subscription Facility from GEM Digital Limited and launched VRAM Network on Sui testnet as it prepares for mainnet. The network aims to coordinate decentralized LLM training with hardware-verified scoring, on-chain rewards, and privacy-preserving credential access.

Why it matters: - VRAM AI is targeting decentralized AI training infrastructure that reduces reliance on hyperscale data centers. - The launch adds another project to the growing Sui ecosystem and broader crypto infrastructure push around decentralized compute. - The protocol aims to make LLM training permissionless while still verifying work and settling rewards on-chain.

What happened: - VRAM AI announced a $10 million Token Subscription Facility from GEM Digital Limited. - The company also launched VRAM Network, a decentralized LLM training coordination network built on the Sui blockchain. - End-to-end testnet validation is complete. - VRAM Network is now preparing for an imminent mainnet launch on Sui.

The details: - VRAM Network uses hardware-attested loss evaluation through AWS Nitro Enclaves. - The protocol uses Sui Seal’s Identity-Based Encryption to protect miner credentials. - Rewards are settled through Sui Move smart contracts. - VRAM Network completed end-to-end testing across multiple miners and validators on a small LLM. - The full coordination stack runs in Rust, including gradient compression, Seal IBE credential decryption, Nitro Enclave loss evaluation, OpenSkill peer ranking, and on-chain reward distribution. - The top-performing miner captured 5.5% of reward weight in the test run based on verified gradient scoring and enclave-attested loss evaluation. - The system supports pluggable training adapters, so miners using existing distributed-training infrastructure can connect without rewriting their training stack. - The technical paper is publicly available with a permanent DOI: the full technical specification.

Between the lines: - VRAM AI is positioning the network as trust-minimized infrastructure for frontier model training, with hardware verification replacing traditional operator trust. - The structure of the protocol suggests an effort to combine privacy, verification, and incentive design in a single training layer. - GEM Digital’s funding gives VRAM AI capital to scale infrastructure before wider mainnet deployment.

What’s next: - VRAM AI plans to use the new funding to scale the miner network, complete mainnet deployment, and expand the validator set. - The company is expected to continue moving from testnet validation toward full production operation on Sui. - VRAM AI said the network is designed for secure, decentralized LLM training that shifts power from data centers to a global permissionless network.