What happened in the year since the launch of the Filecoin mainnet?

Source: Chain News

Author: Eric

From the release of the white paper and the ICO at the end of 2017 to the mainnet launch on the evening of October 15, 2020, Filecoin made everyone wait for about three years. The heated debates during the "space race" testing period before the mainnet launch seem just like yesterday, but in the blink of an eye, the Filecoin mainnet has been operational for over a year. This "giant" project, which raised over $200 million in half an hour, has it met market expectations? What are the memorable "stories" from the past year?

Strengthening Ecological Construction, Empowering Web3 Infrastructure Development

Essentially, Filecoin is just a bottom-layer broker market where demand-side users and storage providers establish trustless connections using blockchain. The real value of Filecoin will be reflected in the applications built on top of it in the future.

The first to realize this may have been the Ethereum developer ConsenSys. Shortly after the Filecoin mainnet launch, ConsenSys announced a collaboration with the Filecoin development team for project integration, allowing developers to integrate IPFS and Filecoin functionalities into their applications through Infura, Codefi, MetaMask, and other ConsenSys projects. In addition to ConsenSys, tech companies like Canada's ChainSafe have also supported Filecoin and launched storage applications based on IPFS and Filecoin.

In terms of storage applications, Filecoin has partnered with NFT projects or companies like Dapper Labs and OpenSea, enabling these projects to store metadata or content data of NFTs through the Filecoin network. Additionally, Textile has created cross-chain bridges using Filecoin storage for NEAR and Polygon, facilitating interaction for users and developers. Furthermore, the publicly listed company 9F has also engaged in ecosystem collaboration with the Polygon and Filecoin development team Protocol Labs regarding its NFT platform NFTSTAR and GameFi projects.

It is worth mentioning that two months ago, Filecoin and Chainlink jointly launched a funding program to support developers in creating "hybrid smart contracts" that integrate Chainlink oracles with Filecoin storage, opening the door for Filecoin to enter blockchain-native markets like DeFi that require extensive use of oracles.

In terms of investment in ecological construction, Filecoin has spared no expense, launching incubators and tens of millions of dollars in ecological investment funds in collaboration with institutions including Huobi, OKEx's Block Dream Fund, and Fenbushi Capital, providing rich funding and resources for the future construction of Filecoin's upper structure.

The layout of Filecoin in the Web3 field is numerous, and will not be elaborated here. Besides Web3, Filecoin has also launched a real data storage reward program to promote applications in traditional fields. This program designates several "witnesses," allowing miners to submit real data storage needs to the witnesses. If approved, the effective data capacity stored by the miner can be recognized by the network as ten times the effective storage capacity. For instance, if a miner stores 10G of real data, the Filecoin network will recognize it as 100G of effective storage capacity for mining rewards. Since the implementation of this program, companies including photography company Taopaipai have attempted to use Filecoin for data and file storage, as compared to traditional cloud storage, the payment for storage using Filecoin is made by the protocol rather than the customer, directly reducing the cost to zero.

Amazing Effective Storage Growth and Soaring Gas Fees

Compared to the steady progress of ecological construction, the Gas fees on the Filecoin network have been one of the most discussed topics in the market over the past year, and also one of the most dramatic "memory points."

The Gas fee model on the Filecoin chain essentially replicates Ethereum's EIP-1559, allowing Gas fees to be better predicted through a base fee + tip structure. As time goes on, the base fee adjusts according to the previous total Gas fees, and most importantly, the base fee is directly burned rather than paid to miners.

At the beginning of Filecoin's launch, due to concerns about its strict penalty measures and the fact that most miners were still experimenting with optimal and cost-effective configurations, while computing power steadily increased, Gas fees remained relatively low. Within the first month after launch, the Gas fee for packaging effective storage peaked at only 0.4 FIL/T, and since FIL prices were not high at that time, these fees were still bearable, with most miners' profits from selling machines sufficient to cover these costs.

In addition, the originally linear release rule for mining rewards was quickly modified after launch through FIP-004 to immediately release 25%, giving miners more confidence to face a series of on-chain costs.

However, the market soon underwent a dramatic change.

As miners' configurations were determined and their operations stabilized, everyone began the process of copying and pasting. With the continuous increase in computing power, various miners started competing by optimizing replication proofs, increasing Gas, and other methods to ensure the speed of their effective storage capacity packaging. However, amidst this competition, the effective storage space packaged across the network quickly reached 1EB by mid-November, marking the beginning of an era where profits were calculated based on the comparison of the network's effective storage space with the baseline. Subsequently, effective storage space skyrocketed and successfully surpassed the baseline by the end of March. Currently, the total effective storage space on the Filecoin network has reached an astonishing 12.37EB.

Google Vice President Kent Walker once stated that by the year 2000, the total amount of data stored in human history was about 12EB. This means that the effective storage space accumulated by the Filecoin network in less than a year is already sufficient to hold all of humanity's "civilization" from the year 2000. If calculated using a server with 36 slots filled with 12T hard drives, this capacity would require over 30,000 servers.

Alongside the effective storage space, Gas fees have also risen sharply. First, the base fee skyrocketed. Due to the need for quick submissions to the chain during the SubmitWindowedPoSt (a part of the replication proof) phase in the packaging storage process, miners hoping to quickly package capacity began a cost-agnostic competition, reminiscent of certain NFT launches on Ethereum.

The Gas consumption for packaging a 32G sector approached 7 FIL/T (about $161) in early February, and by early April, it absurdly reached 9 FIL/T (about $1600). Including the base fee, the Gas fees required to package 1T of capacity once exceeded 15 FIL, and all these Gas fees were effectively burned.

At these two points in time, the profit per T was only about 0.1 FIL/T, meaning that the storage capacity packaged during these two months, excluding costs like machines and electricity, would take 2 to 4 months just to break even on the Gas fees alone, not to mention the subsequent plummet in FIL prices after April and the further increase in the network's effective storage leading to a decrease in average profits.

This unusual change caused significant turbulence in the mining community. At that time, a considerable number of investors who had heavily invested in mining before or shortly after the Filecoin launch found themselves in a very awkward position: uncertain about the future trajectory of Gas fees, they hesitated to continue packaging storage capacity. "Clients are afraid that Gas fees will continue to rise, but they also worry that if the officials implement measures to lower Gas fees, it would be a loss," a miner told me. "Many large clients invested heavily in hardware early on due to concerns about rising hardware prices, and now they have no money left to pay Gas fees."

Faced with such a situation, the numerous miners came up with various solutions. Larger miners with scale and bargaining power chose to have clients pay the full Gas fees, while some smaller miners used their own funds to support clients, solely to retain them. "Clients completely ignore you," some smaller miners said. "These details were never written in the contract, and many clients don't even understand what Gas is; they just want to know why the hard drives they bought can't be put online now, so we can only use the company's money to cover it. It's already very hard to survive."

As the situation began to spiral out of control, Filecoin officials started looking for ways to reduce the high Gas fees. By moving some proofs off-chain and optimizing the aggregation proof process through code, they managed to bring Gas fees down to a reasonable range by the end of April, and through a series of subsequent upgrades, ultimately stabilized Gas fees below 0.5 FIL/T.

Rising Hardware Prices and "Never Been So Busy" Data Center Sales

The hardware components for Filecoin mining not only include hard drives and servers but also require AMD CPUs (which has also spurred demand for this architecture of servers) and high-performance GPUs. The supply of AMD CPUs and servers is relatively limited, and high-performance GPUs have long been monopolized by Ethereum miners, causing the hardware costs for Filecoin mining to rise sharply.

However, even with the total data volume of over 12EB on the Filecoin network, it still falls short compared to the data centers of internet giants like Alibaba and Tencent, which can accommodate over 100,000 servers. Therefore, the price fluctuations theoretically shouldn't be significant, but based on my observations, the reality is quite different.

Filecoin miners may have deep research in software technology, but most miners, especially small and medium-sized ones, have weak bargaining power with hardware brand agents, and many miners are completely unaware of hardware prices. Through an industry hardware procurement intermediary, I learned that some agents quote Filecoin manufacturers for servers at what can be described as "sky-high prices," and the industry rule is that when a company inquires about pricing, it "locks" that company's inquiry rights, meaning other agents cannot quote against it.

Under such circumstances, the final quotes given to clients by miners can create the illusion of a significant increase in market terminal prices. Although this situation may be an isolated phenomenon, it must be acknowledged that Filecoin mining has contributed to the AMD server market share returning to 16% and the continuous rise in prices.

Additionally, due to the high stability requirements for Filecoin mining, IDC data centers have become the preferred choice. I learned from a sales representative at an IDC data center that many Filecoin miners and miners have started inquiring about data center leasing this year. Their requirements are relatively less stringent, and unlike previous clients who needed to go through bidding and other procedures, this has made them exceptionally busy in the first half of this year, as they have to connect with clients while continuously searching for suitable data centers. "I've never been this busy before; it feels like if this continues, the inventory around first-tier cities will be insufficient."

Conclusion

Just like early Bitcoin, Filecoin has faced a lot of criticism along the way. Ethereum founder Vitalik Buterin has also stated that the scalability that Filecoin can bring to blockchain may be very limited. However, we cannot predict the future. Projects like Arweave, Crust, and Swarm, which are also storage infrastructures alongside Filecoin, have not yet seen explosive demand similar to DeFi on Ethereum. Whether blockchain-based storage can compensate for the disadvantages of cloud and local storage, and whether blockchain can quickly evolve to generate large-scale storage demand, remains to be seen in the next year or the next decade.