Golem is one of blockchain’s original “big vision” projects — launched in 2016 with a $8.6M crowdsale (one of the largest at the time), Golem Network proposed to build a global decentralized supercomputer by connecting idle CPU and GPU resources into a peer-to-peer marketplace where anyone could rent compute power for any computationally intensive task. The original Golem Application Registry and Brass Golem (focused on CGI rendering) were replaced by a rebuilt Yagna protocol in 2021, which powers the current network with a more general compute marketplace supporting Docker-based workloads, AI inference, and scientific computing. GLM (formerly GNT — Golem Network Token) is used to pay providers for compute; requestors pay GLM and providers earn GLM for completed tasks. Despite never reaching the mainstream adoption originally envisioned, Golem remains one of the longest-operating decentralized compute protocols.
| Stat | Value |
|---|---|
| Ticker | GLM |
| Price | $0.13 |
| Market Cap | $128.91M |
| 24h Change | +2.3% |
| Circulating Supply | 1.00B GLM |
| Max Supply | 1.00B GLM |
| All-Time High | $1.32 |
| Contract (Ethereum) | 0x7dd9...6429 |
| Contract (Energi) | 0xf3ff...e9a8 |
How It Works
Yagna protocol:
Golem’s rebuilt protocol (launched 2021) uses a decentralized marketplace:
- Requestor — publishes a “demand” (compute task specification, max GLM price per unit)
- Provider — publishes an “offer” (compute specs, GLM price per unit, availability)
- Market matching — Demand and offers matched on Golem’s decentralized network
- Task execution — Work runs in an isolated virtual machine (VM or Docker) on provider hardware
- Payment — Requestor pays GLM upon task completion; Golem’s payment system uses Ethereum L2s for low-fee micropayments
Task types:
- CGI/blender rendering (original use case)
- Docker workloads (data processing, machine learning inference)
- AI inference (GPU task support)
- Scientific simulations and batch processing
Payment channels:
GLM payments use state channels and Polygon for microtransactions — avoiding Ethereum mainnet fees for small per-task payments.
Reputation system:
Providers and requestors build reputation scores based on task completion rates and payment reliability.
Tokenomics
| Metric | Value |
|---|---|
| Max Supply | 1,000,000,000 GLM |
| Distribution | Initial ICO (2016) + Golem Foundation reserve |
| Migration | GNT → GLM (1:1 migration completed 2021) |
| Use | Purely utility — payment for compute |
| No staking | GLM has no staking mechanism; pure payment token |
Use Cases
- Compute payments — GLM used to pay providers for completed compute tasks
- Provider income — Computer owners earn GLM by providing idle CPU/GPU capacity
- Ecosystem reserve — Golem Foundation holds GLM for development grants
History
- Nov 2016 — Golem’s $8.6M crowdsale; one of the first major Ethereum ICOs
- 2018 — Brass Golem testnet (CGI rendering focus); significant delays from original roadmap
- 2020 — Clay Golem mainnet (alpha); GNT rebrands to GLM; migration begins
- 2021 — “New Golem” rebuilt from scratch as Yagna protocol; mainnet V1 launches
- 2022 — Golem integrates GPU support; begins targeting AI inference workloads
- 2023–2024 — Golem positions in the DePIN/AI compute narrative alongside io.net and Render
- 2024 — 8th anniversary; one of the longest-running Ethereum projects still operating
Common Misconceptions
“Golem failed and is dead.” Golem’s original architecture was rebuilt entirely in 2021 and the network has been operational since. Task volume is lower than initial projections, but the network functions and processes real workloads — calling it “dead” mischaracterizes a functional but growing slowly network.
“GLM is a governance token.” GLM has no governance utility. It is a pure payment token — used to compensate providers for compute, nothing more. Golem Network’s development is directed by the Golem Factory team, not token-holder governance.