The rising cost and opacity of centralized clouds have made scalable compute harder to access. Fluence Network addresses this gap by making enterprise-grade infrastructure available through a decentralized marketplace with transparent pricing and verifiable reliability.
Co-founder Evgeny Ponomarev explains how the company evolved from its Web3 roots into a global, enterprise-ready compute platform and how Fluence’s network delivers GPU-enabled infrastructure with transparent pricing and decentralized governance, building a neutral and permissionless foundation for scalable and reliable compute.
What problem did Fluence originally set out to solve?
Developers were struggling with two core issues: unpredictable cloud bills and over-reliance on centralized providers. Fluence was designed to give teams predictable, transparent compute where pricing is clear, access is open, and no single vendor controls supply. We started with CPU virtual servers and expanded into GPU compute to help AI builders access enterprise-grade capacity at up to 85% lower cost than hyperscalers.
You call Fluence “enterprise-grade decentralized compute.” What gives it that quality?
Enterprise-grade means reliability, consistency, and accountability. Fluence connects businesses with professional infrastructure providers operating Tier-3 and Tier-4 certified data centers across multiple regions. These providers meet strict standards for hardware, connectivity, uptime, and security, including GDPR, ISO 27001, and SOC 2 compliance.
The result is a decentralized network of audited data centers where users deploy workloads within seconds with full administrative control. It delivers the same dependability enterprises expect from top-tier cloud providers.
How does the marketplace ensure quality among many providers?
On Fluence, compute providers commit capacity to the marketplace and prove continuous readiness via on-chain Proof of Capacity. Rented capacity is paid in USDC, while idle but ready capacity earns FLT. We partner with professional operators like PikNik and Kabat to anchor reliability, and thousands of blockchain nodes already run on Fluence infrastructure.
What does “cloudless VMs” mean for users?
It delivers the full flexibility of traditional cloud infrastructure while removing vendor lock-in and centralized control. You can spin up VMs across multiple independent providers and still get unified coordination for placement, monitoring, and cost visibility. GPU containers, VMs and bare metal are live, so AI workloads can match the right hardware profile every time.
How do you handle cost transparency and egress fees?
Predictability is everything. Fluence lists all costs up front and eliminates egress fees that typically surprise users. Our decentralized model lets teams estimate expenses confidently. Across our network, customers have already saved over $3.8 million in cloud costs compared to hyperscaler equivalents.
What role does the FLT token play?
FLT powers the economic engine of Fluence. It secures the network, coordinates activity between providers and users, and drives value back to participants. Providers stake FLT to commit capacity, earning yield when their hardware remains online and ready, and receive USDC for active rentals.
FLT also governs the Fluence DAO, where holders shape decisions on upgrades, treasury use, and long-term network direction. Beyond governance, FLT benefits from real protocol revenue, buybacks that reduce token supply over time, and new financial utilities like pFLT staking and RWA-linked yield. As the network scales, FLT becomes both the coordination layer and a store of economic activity across real compute usage.
You recently revealed a Real-World Asset roadmap. What is the vision there?
Fluence sees compute power as a productive on-chain resource. By tokenizing and verifying real workloads, providers can earn measurable returns from their hardware contributions. The RWA roadmap aims to unlock liquidity and financial tools around this model.
We are developing an FLT-collateralized stablecoin so providers can fund operations without selling their tokens, introducing lending markets for unvested FLT, and exploring tokenization of hardware revenue. It deepens liquidity and scales capacity while keeping costs competitive.
What are the first steps in that RWA plan?
The first pilots include a pFLT-backed stablecoin that can also be used for compute payments. The next phase brings lending and staking mechanisms that tie capital efficiency directly to verified performance. It is all about aligning economic incentives with real-world compute delivery.
How is Fluence positioned for the current AI boom?
We are addressing the bottleneck of access. Fluence enables instant provisioning of GPUs for inference, fine-tuning, and model serving while keeping costs predictable. We are also exploring confidential GPU computing through TEEs and encrypted memory to support privacy-sensitive inference. AI builders want speed and sovereignty. That is where our network fits.
What is your perspective on centralization risk after recent cloud outages?
Cloud outages are still treated as rare incidents when they are structural risks. Each major failure exposes the fragility of concentrating orchestration in a handful of providers. DePIN architecture is built to tolerate these failures by design, ensuring no single operator can bring the system down. In Fluence, resilience is engineered into the network itself, allowing workloads to keep running even when parts of the infrastructure fail.
Where is Fluence expanding next?
We are increasing supply in regions with strong AI and Web3 demand, integrating decentralized storage networks such as Filecoin and Arweave for modular AI pipelines, and rolling out specialized GPU models globally. The aim is a neutral compute-data stack that users can deploy end to end under one open protocol.
What milestones can users look forward to?
More verified GPU providers, container runtime upgrades, privacy-preserving compute pilots, expanded staking pools, and seamless data-stack integrations. Every step moves us closer to simple access and predictable, upfront pricing for high-performance compute.
Read more interviews here.
