
Metrics
About the project
What is Filecoin? It is a blockchain that pays independent operators to keep other people’s data on their disks and makes them prove, every day, that the data is still there. The operators are called storage providers. They earn FIL, the network’s token, and must lock FIL as collateral that is burned if they fail. Mainnet has paid block rewards since 15 October 2020.
Six years on, the machine works and the market around it is thin. Raw storage capacity fell from 2,934 PiB on 10 October 2025 to 1,309 PiB on 19 October 2026, according to the Filfox explorer, a drop of about 55% (our arithmetic). Users paid $69,037 in transaction fees over the 30 days to 18 October 2026, DefiLlama data show. And on 19 October the protocol changed what it rewards, four days after the founders’ tokens finished vesting.
This month: Solstice is live and Fil+ has ended
The Solstice upgrade (network version 29) was scheduled for epoch 6,470,279, 12:59:30 UTC on 19 October 2026, and the Filecoin status page marked the maintenance complete at 13:59 UTC. It carries one proposal, FIP-0118, which retires the Filecoin Plus data-subsidy programme and splits block rewards between block winners and a new service pool. Our news report covers the day the Solstice upgrade switched on. This page explains the system it changed.
What is Filecoin and what is it used for
Filecoin comes from Protocol Labs, the company Juan Benet founded in May 2014, which also built IPFS, the content-addressed file system many Web3 apps use to name and fetch files. IPFS finds a file by its fingerprint. It does not promise that anyone will keep it. Filecoin adds the missing part: a market where a client pays a provider to store specific data for a fixed period, and a chain that checks the promise is kept.
The Filecoin documentation calls FIL a utility token that rewards persistent storage. It pays for storage deals, it is the collateral providers lock to take part, and it pays the gas for transactions and smart contracts on the chain.
Since 26 March 2026 developers can also buy storage through Filecoin Onchain Cloud, a “warm” storage service with daily proofs and on-chain billing. On launch day the Filecoin blog counted 49.41 TiB stored across 478 active datasets and 81 paying wallets, with prices from $2.50 per TiB per month for each copy and two copies by default. The network’s raw capacity on 19 October was more than 27,000 times that launch-day volume (our arithmetic, different dates).
How Filecoin works: sectors, proofs and pledge collateral
Storage providers and sectors
A storage provider runs software such as Lotus or Curio on its own hardware and offers space in fixed units called sectors, of 32 GiB or 64 GiB. The consensus minimum is 10 TiB of power. Client and provider agree a deal off-chain, on price and duration, and publish it on-chain. Since March 2023 deals must last between 180 and 540 days, according to the documentation, so long-term storage means renewing them.
Sectors do not have to hold client data. A provider can fill them with random data, called committed capacity, and still earn block rewards for the space. The documentation calls this the easiest way to get started as a provider, and warns that its economics are very dependent on the price of FIL.
Proof of replication and proof of spacetime
Before a sector counts, the provider seals it: a slow computation that turns the data into an encoding unique to that provider. The resulting Proof-of-Replication (PoRep) shows the provider holds its own physical copy and not a pointer to someone else’s.
Then come the checks over time, grouped under Proof-of-Spacetime. WindowPoSt splits each 24-hour period into 48 deadlines of 30 minutes, and every sector is checked once a day; the provider answers with a compact zk-SNARK proof posted on-chain. WinningPoSt is the short proof a provider must give when it is elected to produce a block. Miss a WindowPoSt window and part of the collateral is forfeited, while the provider’s power drops.

Pledge collateral and penalties
Each new sector requires an initial pledge in FIL. On top of that, 75% of every block reward is held back and released linearly over 180 days, and that locked reward also backs the provider’s commitments (the rule comes from FIP-0004). A sector that goes offline costs a fault fee equal to 3.51 days of its expected block rewards, for each day it stays down. Ending a sector early triggers a termination fee, and consensus faults are slashed separately.
Forfeited collateral is not handed to anyone. It goes to the burn address f099 and leaves the supply for good. On 19 October 2026 at 18:17 UTC, Filfox counted 62.26 million FIL locked as pledge collateral, supply that cannot be sold while the sectors behind it are active.
Expected Consensus: power comes from proven storage
Filecoin uses neither proof of stake nor proof of work. Its consensus, called Expected Consensus, elects block producers at random in proportion to their storage power, using randomness from the drand beacon. An epoch lasts 30 seconds and about five providers win the right to produce a block in each one; their blocks are grouped into a tipset. The documentation treats a block as final after 900 epochs, about 7.5 hours.
Power comes in two kinds. Raw byte power is the plain size of proven sectors. Quality-adjusted power, the figure that decides block wins, used to give sectors with verified client deals ten times the weight of committed capacity. That multiplier was the core of Filecoin Plus. Since 19 October every new sector gets the 10x weight, whatever it holds.
The Filecoin Virtual Machine and what runs on it
Programmability arrived on 14 March 2023, when the Filecoin Virtual Machine (FVM) went live with network version 18 at epoch 2,683,348, 15:14 UTC. The FVM runs WebAssembly code, and its first runtime was the Ethereum Virtual Machine, so a smart contract written in Solidity can be deployed with familiar tools and Ethereum-style addresses.
The largest applications by DefiLlama’s count are pools that lend FIL to storage providers for their collateral: GLIF held about $26.7 million on Filecoin and stFIL about $5.6 million on 19 October 2026. DefiLlama’s total for the chain was only $2.86 million that day, against a peak of $122.2 million on 5 December 2024. The chain total is smaller than GLIF’s own figure, so DefiLlama does not count those pools in it and the numbers should not be added. Stablecoins on Filecoin came to about $204,500.
A newer layer sits on top: Filecoin Onchain Cloud with its Proof of Data Possession checks, a payment system called Filecoin Pay that streams money from clients to providers, and the Synapse SDK for developers. The 2026 network strategy puts paid, on-chain deals first among its objectives, and these are the tools meant to produce them.
FIL tokenomics: a 2 billion cap, minting and burns
Filecoin’s specification fixes the base supply at 2,000,000,000 FIL, the most that will ever be created. Because fees and penalties are burned, the documentation says the cap will never actually be reached. The genesis split, from the token allocation chapter of the spec, with release terms from the documentation:
| Allocation | Share | FIL | Release |
|---|---|---|---|
| Storage mining rewards | 55% | 1.1 billion | Block rewards, simple plus baseline minting |
| Mining reserve | 15% | 300 million | Held in reserve |
| Protocol Labs (incl. 4.5% team and contributors) | 15% | 300 million | 6-year vesting from mainnet |
| Filecoin Foundation | 5% | 100 million | 6-year vesting from mainnet |
| 2017 token sale investors | 7.5% | 150 million | 3-year vesting |
| Ecosystem development and potential future fundraising | 2.5% | 50 million | No schedule given |
The 1.1 billion FIL of mining rewards are minted in two ways. Simple minting covers 30% of it, 330 million FIL, released on a six-year half-life whatever the network does. Baseline minting covers the other 770 million and flows in full only when total storage power keeps up with a target that started at 2.5 EiB and grows 100% a year. The documentation says the whole amount would be released only if the network reached a yottabyte within 20 years. By our arithmetic that target now stands above 150 EiB, against 11.6 EiB of quality-adjusted power, so most baseline rewards are deferred. Our tokenomics explainer covers why rules like these matter more than the headline cap.
On 19 October 2026 at 18:17 UTC, the Filfox explorer gave these figures for the chain (the percentages are our arithmetic):
| Measure (Filfox, 19 Oct 2026, 18:17 UTC) | Figure |
|---|---|
| Total supply | 983.48 million FIL |
| Circulating supply | 921.26 million FIL, about 46.1% of the cap |
| Burned since launch | 43.03 million FIL, about 4.4% of total supply |
| Locked as pledge collateral | 62.26 million FIL |
| Mined per day | about 56,800 FIL |
Data sites that use their own method publish other circulating figures, so totals for FIL differ from one site to the next. This page uses Filfox, which reads the numbers from the chain.
The end of the six-year vesting
The 400 million FIL reserved for Protocol Labs and the Filecoin Foundation vested over six years from mainnet, about 66.7 million FIL a year. That schedule ran out around 15 October 2026. FilecoinTLDR, an explainer site from the Filecoin ecosystem, wrote on 4 September that the end of vesting cuts gross issuance “by roughly 75%”, leaving about 22 million FIL a year from block rewards. Filfox’s daily mining rate gives about 20.7 million FIL a year (our arithmetic), close to that estimate. Our glossary entry on the token unlock explains how schedules like this one move supply.
Lower issuance is not the same as a shrinking supply. The net change depends on how much FIL is burned through fees and penalties, and how much is locked in, or released from, pledge collateral.
What Solstice changes in the reward stream
From 19 October 2026, under FIP-0118, every block reward is split. The consensus share, paid to block winners, starts at 95% and falls in a straight line to a floor of 50% over about nine quarters of 262,974 epochs each. The service share, paid to registered Service Orchestrators, starts at 5% and reaches 10% over the first, ungated quarter. After that it can rise by five points a quarter, but only if payment volume through Filecoin Pay meets a dollar target: $3,500 for the first step, then 2.7 times the previous target at each step. Whatever neither share takes is burned.
Only one Orchestrator is registered at launch, and the FIP leaves open registration to a “Phase 2” with no date. Which entity runs that first Orchestrator was not public on 19 October. Counting from the activation epoch, the first quarter ends around 18 January 2027 and the consensus share would reach its 50% floor around 18 January 2029, if the schedule is not changed.
Network data: capacity is falling and fees are small
Raw storage power has been falling since 2022 and quality-adjusted power since the end of 2023. In Filfox’s quarterly samples, raw byte power peaked at about 17,286 PiB on 12 September 2022 and quality-adjusted power at about 25,527 PiB on 24 December 2023. Raw power has lost about 18% a quarter for the past year, and the quality-adjusted decline has steepened:
| Date (Filfox sample) | Raw byte power | Quality-adjusted power |
|---|---|---|
| 12 Sep 2022 | 17,286 PiB | 18,888 PiB |
| 24 Dec 2023 | 9,060 PiB | 25,527 PiB |
| 10 Oct 2025 | 2,934 PiB | 22,625 PiB |
| 11 Jan 2026 | 2,406 PiB | 19,808 PiB |
| 15 Apr 2026 | 1,974 PiB | 17,004 PiB |
| 18 Jul 2026 | 1,609 PiB | 14,318 PiB |
| 19 Oct 2026 | 1,309 PiB | 11,871 PiB |
By our arithmetic, raw power fell about 55% and quality-adjusted power about 48% in the year to 19 October 2026, and raw power is about 92% below its 2022 peak. FIP-0118 itself cites a net loss of 20 to 37 PiB of quality-adjusted power a day in July 2026. Filfox counted 467 active storage providers with power on 19 October. Protocol Labs reported more than 3,300 providers in October 2021, but that count used another method, so the pair points to consolidation without giving a clean rate of decline.
How much of the remaining space holds client data is not published as a current primary figure. The latest independent measure we found is Messari’s State of Filecoin report for the third quarter of 2025, as reported by RootData and other outlets: about 1,110 PiB of active data, 35.2 million active deals and utilisation of 36%, up from 32% in the second quarter. We found no comparable figure for 2026.
Money flowing through the chain is modest. DefiLlama’s Filecoin page counts transaction fees only, which on this chain are burned: $1,778 on 18 October 2026, $69,037 over 30 days and $1,964,924 over the past year. Storage deal payments are not part of that metric. For paid storage, FilecoinTLDR reported on 4 September 2026 that Filecoin Pay’s annualised run rate grew from $663 in January to $59,327 at the end of August, with active payers up from 73 to 119. That run rate equals about 3% of a year of gas fees (our arithmetic).

Governance: FIPs, the Foundation and the end of Fil+
Protocol rules change through Filecoin Improvement Proposals (FIPs), drafted and debated in public on GitHub, then shipped by the client teams in network upgrades. The repository held 97 FIP files on 19 October 2026, the highest number being FIP-0118. Earlier proposals introduced the “free fault” for honest providers (FIP-0002), the 75/25 reward vesting (FIP-0004) and proof aggregation with SnarkPack (FIP-0009). Mainnet has had three upgrades in about 13 months: Golden Week on 24 September 2025, Fire Horse on 27 May 2026 and Solstice.
Several teams write node software, so the network does not hang on one codebase: Lotus, the reference client maintained by Protocol Labs, Forest by ChainSafe, Venus by IPFSForce, and Curio for storage providers. Protocol Labs says it completed its move to a decentralised network of independent teams in July 2024. The Filecoin Foundation, which received 5% of the genesis supply, and the PL Filecoin Impact Fund signed the 2026 network strategy together.
That strategy, published on 19 February, set the direction in one line: “This year marks a shift in ecosystem focus, away from growing supply and towards scaling demand.” Solstice follows from it. Filecoin Plus relied on allocators who granted DataCap to clients whose data they judged useful. FIP-0118 argues that useful data cannot be detected by a program, so the system needed a human review pipeline open to gaming, and the Solstice blog post of 17 July 2026 said Fil+ verification had become a weak signal of useful data. The replacement pays services according to what clients settle through Filecoin Pay.
The 2017 token sale and the SEC
Protocol Labs sold FIL in 2017 through Simple Agreements for Future Tokens (SAFTs), under the US Rule 506(c) exemption that limits buyers to accredited investors. Its Form D filings with the SEC, signed by Juan Batiz-Benet as chief executive, show two offerings. The pre-sale, first sold on 21 July 2017, raised $51,658,489 from 154 investors. The second offering, first sold on 10 August 2017, had raised $148,238,169 of a $160 million target from 1,241 investors by the amended filing of 25 August.
That second filing came before the sale closed. CoinDesk reported on 7 September 2017 that the public sale, run through CoinList, ended at about $205.8 million, which with the $52 million pre-sale made more than $257 million. Protocol Labs’ own timeline gives $205.8 million as the total, the figure CoinDesk gives for the public sale alone. The specification puts the tokens sold at 7.5% of the base supply, or 150 million FIL. CoinDesk, whose parent company disclosed a stake in Protocol Labs, named Sequoia Capital, Andreessen Horowitz and Union Square Ventures among pre-sale buyers.
The token’s legal status came up in 2023. On 17 May, Grayscale disclosed that SEC staff considered FIL to meet the definition of a security and had asked it to withdraw the registration statement of its Grayscale Filecoin Trust, Decrypt reported. Grayscale said it did not believe FIL was a security, and a Protocol Labs spokesperson said “Filecoin is not a security.” Grayscale filed its withdrawal request on 9 June 2023.
The SEC’s complaints against Binance (5 June 2023) and Coinbase (6 June 2023) both listed FIL among the crypto assets the agency alleged were securities. Both cases were later dropped: the SEC announced the Coinbase dismissal on 27 February 2025, saying it did not reflect a view on the merits, and the SEC and Binance filed to dismiss on 29 May 2025, The Block reported. No court has ruled on FIL’s status. We found no SEC enforcement action against Protocol Labs itself.
Audits of the Filecoin protocol
The core was audited before and around mainnet, and the specification keeps an index of the reports. Consensys Diligence reviewed the built-in actors (the protocol’s on-chain programs), in a report the spec dates 19 October 2020. Sigma Prime reviewed the proving subsystem in July 2020 and the Lotus daemon and miner in October 2020. NCC Group reviewed the Bellman and BLS signature code in a report dated 20 October 2020. Least Authority audited the Lotus implementation (final report 19 December 2020) and the Venus client (June 2021). Oak Security audited the FVM’s Ethereum runtime in a report dated 9 March 2023.
We found no public audit of the new actors added by Solstice, the ones that weight and pay the reward streams, as of 19 October 2026. The FVM audits announced in April 2022 for the Rust built-in actors did not lead to a published report we could find either.
Filecoin against Arweave, Storj and the cloud
Filecoin sells storage for a fixed term, 180 to 540 days per deal, which the client must renew. Arweave describes itself as permanent information storage. Storj sells S3-compatible cloud storage spread over a global network of resources, without Filecoin’s on-chain proofs. We found no neutral price comparison across these networks, and each project’s own comparisons favour itself.
The price that matters most may be the one set by centralised cloud providers. Filecoin’s newer products are built to look like them: Onchain Cloud bills by the TiB per month and ships with a developer SDK, and FilecoinTLDR reports that Fil One, an S3-compatible service launched in June 2026, charges $4.99 per TB a month with no egress fees. Whether those products bring paying clients will show in Filecoin Pay volumes, quarter by quarter.
Filecoin risks: thin demand and an untested reward model
We rate Filecoin 3 out of 5 for risk. The core protocol has run for six years and was audited, and both SEC cases that named FIL were dropped in 2025. The score stays at 3 because of the following:
- Paid demand is small next to the size of the network: about $2 million of fees in a year (DefiLlama) and a Filecoin Pay run rate near $59,000 a year at the end of August 2026, according to an ecosystem source.
- Capacity is shrinking, and smaller rewards for block winners under Solstice may push more providers out unless paid services replace that income.
- The new reward system starts with a single Orchestrator whose identity was not public on 19 October 2026, and we found no audit of its new actors.
- The end of vesting cuts gross issuance, but net supply still depends on burns and on pledge locked or released, which nobody can fix in advance.
- Many ecosystem figures come from Filecoin-affiliated sources, and independent utilisation data stops at Messari’s third quarter of 2025.
Nothing on this page is investment advice. The risk score describes the project’s structure, not a view on the token’s price.
Solstice dates to watch in 2027
The first Solstice quarter ends around 18 January 2027 (our count from the activation epoch). The second ends around 20 April 2027, and its Filecoin Pay volume decides whether the service share takes its first gated step, about 17 days later at the earliest, so around 7 May 2027. The FIP gives no date for open Orchestrator registration. Before any of that, the daily power figures on Filfox will show whether giving every new sector the 10x weight slows the outflow of providers.
Strengths and points of vigilance
Strengths
- Mainnet has run since October 2020 with daily, on-chain proofs that providers still hold the data.
- The core protocol, proofs and EVM runtime were audited by Consensys Diligence, Sigma Prime, NCC Group, Least Authority and Oak Security.
- Supply has a 2 billion cap, 43 million FIL have been burned and the founders' six-year vesting ended in October 2026.
- Solstice ties part of future block rewards to storage that clients actually pay for.
- The SEC cases against Binance and Coinbase that named FIL were dropped in 2025.
Points of vigilance
- Transaction fees came to about $2 million over the past year according to DefiLlama, a small sum for a network of this size.
- Raw storage power fell about 55% in the year to 19 October 2026, according to Filfox data.
- The new reward model starts with a single Service Orchestrator whose identity was not public at launch, and no audit of its actors was found.
- Independent utilisation data stops at Messari's third quarter of 2025.
- Net supply still depends on burns and pledge flows, so the end of vesting does not by itself shrink supply.
Team
- Juan Benet Founder of Protocol Labs
- Molly Mackinlay Co-author of FIP-0118 (Solstice)
- Irene Giacomelli Co-author of FIP-0118 (Solstice)
Funding and audits
| Round | Amount | Date | Investors |
|---|---|---|---|
| SAFT pre-sale (SEC Form D, Rule 506(c)) | $51,658,489 sold to 154 investors | Accredited investors; CoinDesk named Sequoia Capital, Andreessen Horowitz and Union Square Ventures | |
| SAFT public sale (SEC Form D, Rule 506(c)) | $148,238,169 sold of a $160,000,000 offering to 1,241 investors at the filing of 25 August 2017; about $205.8 million at close per CoinDesk | Accredited investors through CoinList (per CoinDesk) |
Roadmap
- 2020-10-15Partono i premi di blocco sulla mainnet, al blocco 148.888Done
- 2023-03-14Filecoin Virtual Machine attiva con la versione di rete 18Done
- 2026-02-19La strategia di rete 2026 sposta l’attenzione dall’offerta alla domanda a pagamentoDone
- 2026-03-26Filecoin Onchain Cloud attivo sulla mainnetDone
- 2026-05-27Versione di rete 28, Fire HorseDone
- 2026-10-15Finisce il vesting di sei anni dei token di Protocol Labs e Filecoin Foundation (circa)Done
- 2026-10-19Versione di rete 29, Solstice: finisce Filecoin Plus, i premi di blocco si dividono con un fondo serviziDone
- 2027-01-18Fine del primo trimestre di Solstice, la quota per i servizi sale al 10% (circa, calcolata dall’epoca di attivazione)
- 2027-04-20Fine del secondo trimestre; il suo volume su Filecoin Pay decide il primo scalino condizionato (circa)
- 2029-01-18La quota di consenso tocca il minimo del 50% se il calendario non cambia (circa)
Crypto-assets are volatile and you can lose all the money you invest. Nothing published on Coindaro is investment advice.






