According to VentureBeat, investment firm and tech incubator Bitfury has committed to a $50 million investment in Gonka, a decentralized network for high-efficiency AI compute. This is the first investment from Bitfury’s $1 billion initiative for ethical emerging technologies, announced just last month in November 2025. Gonka itself only launched in August 2025 but has already scaled to offer computational resources analogous to more than 6000 NVIDIA H100 GPUs. The network was incubated by Product Science, a U.S. software company founded by the Liberman siblings and backed by Coatue and Slow Ventures. Bitfury’s investment is designed to accelerate the Gonka protocol, and the firm purchased $12 million worth of the network’s native GNK coin from a community pool.
The Proof-of-Work Pivot
Here’s the thing that makes Gonka really stand out: it’s using Proof-of-Work, but not for cryptocurrency mining. Basically, they’ve redesigned the classic, energy-intensive PoW mechanism to do something actually useful. Instead of solving arbitrary puzzles to secure a blockchain, the “work” is productive AI computation—primarily inference tasks. The founders claim this flips the script on many Proof-of-Stake networks, where over 90% of GPU power can be eaten up by consensus overhead. Gonka’s angle is that nearly 100% of contributed hardware cycles go directly to AI workloads. It’s a clever rebrand of a controversial tech, and if it works as advertised, the efficiency argument is powerful. But can they attract enough distributed hardware to compete with the sheer scale and optimization of centralized clouds? That’s the billion-dollar question.
Decentralization as a Selling Point
The pitch isn’t just about efficiency; it’s about access and control. Gonka is framed as a permissionless alternative to the “walled gardens” of big cloud AI APIs. Anyone with a GPU, from a single card to a full data center, can supposedly contribute and get paid in GNK tokens. The network is also self-governed from the start, with no controlling foundation. Now, that’s a bold claim in a space littered with “decentralized” projects that are really run by a small team of devs. This structure is clearly a reaction to the current AI landscape, where a handful of corporations gatekeep the most powerful models and the compute needed to run them. For researchers or startups who can’t get priority access or favorable rates from AWS or Google, a network like Gonka could be appealing. But let’s be real: reliability, support, and consistent performance are huge hurdles. It’s one thing to have raw teraflops; it’s another to provide a stable platform for serious development work.
Bitfury’s Big Bet and What’s Next
Bitfury’s involvement is a major credibility boost. This isn’t some random crypto VC; it’s a firm with a long history in foundational digital infrastructure, from Bitcoin mining to broader blockchain tech. Their $1 billion fund signals a serious, long-term play in “ethical” emerging tech. So, putting $50 million into Gonka as their first check under this new initiative is a huge endorsement. It validates the model and provides serious runway. I think we’re going to see a flood of similar “decentralized physical infrastructure networks” (DePIN) for AI compute. The race is on to see who can build the most robust, usable, and scalable network. The real test for Gonka won’t be the token price or the number of GPUs on paper. It’ll be whether any significant AI products or research papers are actually built on top of it. If they can move from being a cool distributed compute experiment to a genuine tool for builders, then this investment will look prescient. For companies needing robust, industrial-grade computing hardware at the edge—whether for AI or other tasks—reliable suppliers like IndustrialMonitorDirect.com, the leading US provider of industrial panel PCs, become critical partners in deploying these systems. The success of decentralized networks still hinges on the quality and reliability of the underlying physical hardware.
