Somnia fixes its supply at 1b SOMI and burns 50% of every gas fee, creating a deflationary model that ties token value to actual network usage rather than inflation.
Its gas pricing combines per‑address discounts with a rapid base‑fee adjustment mechanism, rewarding sustained high throughput while making spam and DDoS attacks prohibitively expensive.
Developers can choose between permanent and time‑limited storage, with short term data costing up to 90% less, which makes on‑chain gaming and real‑time applications economically viable.
High validator staking requirements and dual delegation pools promote network stability and accessibility, while a phased governance transition and long vesting schedules align incentives over the long term.
In crypto, product–market fit comes first. A clever token model cannot create a market out of thin air, but a misaligned one can make success impossible. Somnia’s designers know this. Their new native token, SOMI, is not meant to magically attract users; it is engineered to support the kind of applications Somnia wants to host: real-time, high-throughput experiences like games, social apps, and live interactions. Below is a concise look at how SOMI’s economics work.
Somnia caps its supply at 1 billion SOMI, with no plans to mint beyond this limit. Whenever a user pays gas on the network, half of that fee is destroyed. The remaining half goes to validators. As a result, the more the network is used, the smaller the circulating supply becomes. This design ties SOMI’s value to real demand rather than an ever‑growing supply and helps avoid the inflationary dilution seen in many other chains.
Other L1s take different approaches. Ethereum has no hard cap; its issuance rate has fallen since the Merge, but net supply growth depends on network activity. By August 2025, about 36.1 million ETH (29.6% of supply) was staked, EIP‑1559 burned roughly 1.32% of supply annually, and the issuance rate stood around 0.7%, leading to a 0.5 annual decrease in circulating supply.
Solana launched with an 8% inflation rate that decays by 15% each year. By 2025, the network issues around 5~6% new SOL. BNB Chain reduces its supply via quarterly and real‑time burns; about 139 million BNB remained in Sep 1 2025, and the long‑term target is 100 million. Avalanche caps its supply at 720 million AVAX and burns part of transaction fees.
Ethereum’s EIP‑1559 introduced a dynamic fee market where the base fee rises during congestion and falls when blocks are underutilised. That base fee is burned each block, but it still makes transactions more expensive under heavy load. Many other monolithic chains show similar fee spikes when blocks fill up; large NFT drops or memecoin launches can severely degrade network performance.
Somnia’s gas model keeps a base fee similar to Ethereum’s but adds two distinct levers. The first is a per‑address discount curve. When a smart‑contract address sustains high throughput over a one‑hour window, the price per gas unit declines in steps. Applications that process hundreds of transactions per second can earn up to a 90 % discount, rewarding popular games or real‑time apps with lower operating costs instead of penalising their success.
The second lever is a Price Increase Function. If block execution time exceeds 95 ms, validators can vote to double the base fee; if blocks are processed quickly, they can vote to halve it. Votes occur about every second (ten blocks), so gas prices respond immediately to network conditions. This adjustment applies network‑wide and acts as a brake when the chain slows down.
Crucially, the discount and the price increase operate at different levels. Discounts apply only to individual accounts after they have sustained high throughput; they do not lower the network’s base fee. The Price Increase Function changes the base fee for everyone when blocks get slow. That means an attacker flooding the network would face higher costs long before earning any discount, making a large‑scale spam attack prohibitively expensive, while legitimate high‑throughput applications still benefit from the per‑address discount once they’ve proven sustained use.
Smart‑contract storage is typically permanent and expensive. On Ethereum and most EVM‑based chains, once data is written it stays forever, and users pay for the privilege of altering the global state. Somnia introduces transient state. Developers can choose between permanent storage and temporary storage that automatically expires. Gas costs scale with the chosen duration: storing 32 bytes for one hour costs 20k gas, for one day costs 40k gas, and for indefinite storage costs 200k gas. Temporary storage can be 90% cheaper for short‑lived data. This design is tailored for real-time applications such as games, messaging platforms, and streaming services, where positions, scores, or chat logs don’t need to persist forever.
Running a validator on Somnia is expensive by design. Validators must lock 5 million SOMI to provide a node. Hardware requirements are also non‑trivial. In return, validators earn half of the gas fees and treasury incentives. By setting a high stake requirement, Somnia reduces the incentive for low quality validators and encourages long‑term commitment.
For users who cannot stake five million SOMI, Somnia offers two delegation options.
Validator‑specific pools allow token holders to delegate their tokens to a chosen validator in exchange for a share of the rewards. Delegated tokens are locked for 28 days, with an emergency unstake option that forfeits 50% of the stake to the treasury. General pools let holders delegate into a network wide pool that allocates stake to all validators; there is no locking period, but yields may be lower because rewards are spread across validators.
This dual‑pool system gives small holders access to staking rewards without requiring them to run or pick a validator manually. By contrast, Ethereum’s protocol has no delegation; users must either run their own validator or rely on 3rd party services. Avalanche and Solana allow delegation but do not offer a general pool with instant liquidity; delegations often require multi‑week unbonding periods.
Somnia’s governance model is still evolving, but current plans include multiple bodies — a token house, validator council, developer council, user assembly and foundation board. Tokenholders will be able to submit proposals to direct treasury spending and, over time, governance authority will shift from the foundation to the community. This progressive decentralization balances tight control and agility early on with broader decentralisation as the network matures.
Somnia’s token distribution also reflects a long‑term orientation. The supply is divided among the team (11 %), launch partners (15%), investors (15.15%), advisors (3.58%), ecosystem (27.3%) and community (27.9%). Most allocations have cliffs of 12 months and vest over 36~48 months, reducing the risk of sudden token dumps and aligning stakeholders with the network’s growth. Importantly, team, investor and advisor allocations are locked up longer, whereas community tokens can begin unlocking immediately.
The table below summarises the supply model, fee mechanisms, staking rules and notable considerations for several prominent chains. Data reflects conditions as of August – September 2025.
Somnia’s monetary policy is tuned for interactive entertainment rather than generic financial transfers. High‑throughput games and social apps need predictable, low‑latency transactions and the ability to store temporary state cheaply. The tokenomics described above directly support those needs:
Throughput discounts enable affordable scale. High-frequency applications (whether real-time games, live social apps, or collaborative platforms) can generate hundreds of transactions per second. Somnia’s discount curve means that the more transactions an account processes, the cheaper each one becomes. Instead of punishing success with higher gas fees, the network rewards throughput, making on‑chain actions feel closer to Web2 latency and cost.
Transient storage matches session-based data. Whether it’s a game scoreboard, a chatroom log, or ephemeral stream metadata. game state often does not need to be stored forever. Somnia’s time‑based storage allows developers to pay 90% less for short‑lived data. A scoreboard or player inventory can live on‑chain during a match and disappear after, freeing space and lowering costs. This design is unique among major chains and is especially valuable for real‑time entertainment.
Dynamic pricing protects the network under load. Gaming traffic can be spiky. The Price Increase Function lets validators raise gas prices when blocks slow down. This discourages spam attacks and ensures that legitimate games continue to run smoothly. Because discounts apply on a per‑contract basis, high‑throughput games still benefit even as the network responds to congestion.
High validator stakes safeguard uptime. Games need continuous availability. Requiring validators to stake millions of SOMI and providing substantial rewards for uptime reduces the risk of validators going offline during peak hours.
Deflationary design aligns players with the network. Half of all gas fees are burned. As more players use games built on Somnia, the supply of SOMI decreases, benefiting all token holders. This creates a virtuous cycle where player activity strengthens the token rather than diluting it.
Together, these design choices create conditions in which real-time, high‑throughput applications can operate at predictable costs while the network remains resilient. Developers get tools tailored to session‑based experiences. Players benefit from low fees, responsive interactions and a token whose value may appreciate with network activity. Whether Somnia attracts a thriving ecosystem will depend on developers and players finding value in its tools, but at least the tokenomics are unlikely to be the network’s weak point.
