Trust Link Inc., operator of AGGX.IO, has closed a $5,000,000 Strategic Infrastructure Financing Round to accelerate the AGGX-DePIN Infra-Engine — an AI-native model suite that automates compute routing, workload verification, and real-time GPU asset token pricing across global heterogeneous GPU clusters.
To process terabytes of real-time telemetry, continuous workload parallelization, and zero-knowledge compute verification across thousands of distributed GPUs, AGGX.IO is training three core proprietary models.
AI-Driven Intelligent Workload Offloading Model
Ingests continuous telemetry streams — VRAM latency, CUDA core thermal metrics, network jitter, and matrix-multiplication throughput — from globally aggregated nodes.
Processes over 1.5 Trillion time-series telemetry tokens to continuously train an RL-driven Mixture-of-Experts (MoE) model capable of dynamic, microsecond-level AI task partitioning across heterogeneous hardware.
Distributed Execution Verification
Uses high-throughput lightweight LLMs to audit and verify that distributed worker nodes execute full-precision matrix operations without spoofing or result manipulation.
RWA Algorithmic Pricing
Continuously analyzes global spot GPU rental pricing, electricity cost curves, and compute-token liquidity pools to output real-time algorithmic yields for tokenized GPU assets.
Strictly placed in an escrow account dedicated to High-Density Data Center Co-location and NVIDIA Blackwell HGX B200 / NVL72 Node Procurement.
The AGGX-DePIN Infra-Engine relies on a 200B+ MoE gating network. The 180GB HBM3e memory bandwidth and 1.8 TB/s NVLink bi-directional topology of Blackwell B200 nodes are technically required to prevent inter-node communication bottlenecks during multi-modal time-series matrix pre-training.
Trust Link Inc. (Operator of AGGX.IO) — a premier DePIN GPU compute aggregation and RWA protocol. Strategic Infrastructure Financing Round led by an Asian commercial investment group.