Reheated Cold Rice DePIN: Immature Traditional Technology + Immature Blockchain ≠ Low-Cost Myth

Interview: Beichen

Guest: Steven

The Web3 industry is currently in a very awkward phase—market downturn due to tightening liquidity, and at the same time, there are no new technological paradigms emerging, leading to the reheating of old ideas, such as DePIN (Decentralized Physical Infrastructure Network).

In fact, various blockchain-based hardware infrastructures have been emerging since 2017. Although none have succeeded yet (we can only say they haven't failed, like Filecoin and Arweave, which still have hope for the future), enthusiasm is high now, with a popular belief that DePIN can save over 75% in costs.

"Harsh Whistle" published “On Why Helium, Which Is Fundamentally Flawed, Doesn’t Work, and Discussing Web3 Research Methods” last week, analyzing why Helium, as a representative project of DePIN, fails from a business perspective.

The reasons can be summarized in two points: 1. Helium's blockchain only serves the purpose of issuing tokens and order settlement, it completely does not require a public chain; 2. The IoT services provided by Helium are contrary to the development trends of the IoT industry in reality (various manufacturers prefer to build their own IoT ecosystems).

In this episode, we invited communication technology expert Steven to interpret why blockchain is not suitable for hardware infrastructure from the perspective of communication engineering.

1. Beichen: DePIN is very popular now, believed to drive a new wave of physical infrastructure. This raises the question of whether the business logic holds up and if it is technically feasible. Please share your bold opinion on DePIN from the perspective of communication engineering.

Steven: From a technical standpoint, DePIN currently lacks feasibility and usability, and while there may be hope in the future, it is still far off.

2. Beichen: Why do you think the blockchain industry is talking about DePIN again? To me, this seems like the "chain reform" narrative that went bankrupt in 2018. My personal bias is that traditional industries turn to blockchain only when they can't survive.

Steven: DePIN and chain reform are not the same concept. The idea of DePIN was proposed even before blockchain emerged (for example, there were decentralized Wi-Fi networks in the communications industry, like CDN). Now, with blockchain, financing is easier, and its financial attributes are strong, leading to a natural combination with blockchain, which is today's DePIN. However, the current technology still cannot support the development of DePIN.

3. Beichen: Since the DePIN narrative involves many tracks, let's start with the DeWi (Decentralized Wireless Network) that you are most familiar with. Taking Helium as an example, does its LoRaWAN and 5G network actually have competitiveness?

Steven: In terms of communication technology, the IoT application scenarios that LoRa can meet are very limited, and Helium's combination with blockchain makes its costs very high, making it unsuitable for commercial use. Moreover, Helium's 5G is merely a supplement to telecom operators, essentially a 5G subcontractor, which is not very meaningful.

Although Helium officially boasts many application scenarios, the reality is that the IoT scenarios that LoRa can meet are very limited.

The wireless physical layer of IoT can be divided into long-range (hundreds of meters to tens of kilometers) coverage and short-range (within hundreds of meters) coverage. We are very familiar with Wi-Fi and Bluetooth, which are short-range coverage, while Helium's LoRa belongs to LPWAN (Low Power Wide Area Network). This is a so-called ultra-narrowband IoT protocol, which has very low requirements for network bandwidth, but the trade-off is that the data transmission volume is also very small. This determines that it is only suitable for light nodes and light transmission application scenarios, such as sensors like thermometers and water meters that only need to report data every few minutes, with only a few dozen bytes of data.

Furthermore, since LoRa operates on unlicensed frequency bands, the allocated frequencies vary by country. In Europe and America, the golden frequencies are 433, 868, and 915 MHz (these frequencies have certain penetration capabilities and decent rates, making them the best for voice services and medium-low speed data services), while in China, it is 470~510 MHz.

(Beichen: So do Helium's miners need to be adjusted? Steven: They must adjust the frequency and transmission power accordingly; otherwise, the signal will be interfered with.)

Although LoRa is an unlicensed band, any enterprise can establish its own LoRa network (just like setting up Wi-Fi at home), but the capacity of each frequency network is limited. For example, 5G can theoretically support 1 million devices per square kilometer, while our previous calculations for LoRa indicate that practical planning can only accommodate a few hundred to 2,000 devices.

Helium allows gateways (miners) to perform proof of work based on coverage, which means it does not care about the actual network needs in different locations; as long as it provides network coverage as much as possible, it can receive rewards (even in uninhabited areas). A normal IoT network should optimize node distribution based on actual demand.

IoT requires operation and maintenance, and any issues must be resolved quickly. A normal IoT system invests a lot of resources for dedicated maintenance, while Helium miners find it difficult to perform professional network maintenance, especially since these miners are so dispersed.

4. Beichen: After combining with blockchain, Helium claims to have reduced costs. How do you see this as actually increasing costs?

Steven: I absolutely disagree with the claim that Helium has cost advantages.

First, the cost of building the Helium network is very high. I once disassembled a Helium miner, which was quite expensive, nearly 10,000 yuan. When I opened it up, considering the cost of the chips, similar modules could be obtained for just over a hundred yuan on the market; they just use so-called custom chips (adding an ID in the chip for anti-counterfeiting).

Second, the structure of a distributed network inherently leads to higher data transmission costs. Each Helium miner must first have internet access devices, and secondly, monitoring devices and capabilities. While traditional IoT also requires internet interfaces and monitoring, IoT is strictly isolated from the internet, providing higher security, and does not require many devices and monitoring capabilities (for example, Alibaba Cloud can manage hundreds of thousands of sensors with just a few servers and platform services), which helps spread the costs.

The operation of a LoRa network does not require blockchain; Helium added blockchain verification for financing reasons, which further increased costs.

In fact, if you set aside the blockchain part and look directly at Helium's core business, you will find it is entirely a centralized project, merely placing transactions that can be automated on the blockchain. The blockchain is a completely pseudo-demand for Helium.

5. Beichen: If Helium's approach is unworkable, what kind of solution does IoT truly need?

Steven: Manufacturing enterprises (especially those in continuous production and large-scale discrete production) have very high reliability requirements for information transmission and prefer to build their own networks. They also implement electronic fences and other isolations to ensure that all data is transmitted only within private networks. Only those with lower security levels would consider using IoT cloud service providers like AWS or Alibaba Cloud, making it even less likely to casually adopt Helium's third-party LoRa network, not to mention using it for simple infrastructure like streetlight control.

5G is an important direction for IoT, as it can meet the needs of massive sensors or controllers in a very small area, while LoRa networks are unsuitable for manufacturing enterprises and infrastructure, especially since Helium has a more costly and less efficient distributed architecture.

6. Beichen: Additionally, there are concepts of distributed storage and distributed computing. Storage and computing indeed have market demand, but do they really have advantages over centralized services?

Steven: Whether distributed systems have advantages in storage and computing entirely depends on the problems they are trying to solve—distributed systems inevitably lead to reduced efficiency and increased costs, but if the business itself requires multi-node connections (for example, higher redundancy and reliability requirements, or the need to avoid certain legal restrictions), then distributed systems are suitable.

So there is undoubtedly a demand for distributed storage and distributed computing. However, how large this demand is and whether it can replace the tasks of centralized systems remains questionable.

Take Filecoin as an example. This is a great project that has solved many technical challenges, especially in storage proof and time-space proof. To be honest, developing it was not easy, but Filecoin's most basic economic model is unreasonable and has led to a series of issues.

First, in Filecoin's economic model, FIL, as a functional token, experiences price fluctuations that severely impact actual storage demand. Theoretically, the incentives for ecosystem participants should not be influenced by the token market.

The sunk costs of this economic model make it so that those mining FIL are afraid to use hard drives for mining because if something goes wrong (like power outages or network interruptions), they not only cannot make money but also risk losing their staked FIL, which contradicts Filecoin's original idea of "utilizing redundant storage space." Moreover, Filecoin's miners require GPUs to verify storage validity (replication proof), leading to actual storage costs that are higher than cloud storage, not to mention hard drives.

High costs further lead to insufficient usability, and no enterprise would be willing to store core data in the Filecoin network. This is not a problem that blockchain incentives can solve.

7. Beichen: In your view, what technical issues need to be addressed for a mature "decentralized network architecture based on blockchain technology"? Or is it fundamentally impossible to solve?

Steven: DePIN is a decentralized network architecture based on blockchain technology, and blockchain uses a triple-entry accounting method, requiring consensus among multiple nodes, which determines that its cost and efficiency are inferior to traditional digital systems.

For DePIN to develop in an ideal direction, three issues need to be resolved.

First, the reliability of infrastructure. The maintenance capabilities required for distributed networks are actually higher than those for centralized systems, necessitating miners to have strong professional maintenance capabilities, which current hardware and software maturity levels are far from meeting.

Second, decoupling between software and hardware. Applications dominated by centralized large companies bundle software and hardware together (like the Mac system). Only when there is sufficient decoupling between software and hardware, allowing for software-defined everything, can the ideal of decentralized distributed deployment in DePIN be realized.

Third, reliable mapping of real-world data. A reliable connection must be established between the data world and the physical world, which for blockchain means that oracles need to be very mature.

(Beichen: This is also the foundation for the realization of the RWA concept. Steven: It should be said that this is the foundation for whether blockchain has a future. Centralized systems are maintained by social contracts between people, while blockchain relies on code and probabilistic games. If the source data is unreliable, what future can we talk about?)

Moreover, in addition to overcoming technical bottlenecks, DePIN also needs to address community efficiency, which is a significant issue.

Take 5G as an example; the latest technical standard Release 18 has 35 million words in the main text (in Chinese), and different manufacturers have derived more product standards from it. From R&D to operation, this involves a very large division of labor and cooperation.

Given the current efficiency of various DePIN communities, it is impossible to complete this; they can only wait for centralized organizations to finish all these tasks and feed them to you. At that point, what significance does decentralization hold? It can only be said that the threshold for starting in the blockchain world is lower; many things cannot be initiated in the traditional world, so trying them in the blockchain world is a good thing.

8. Beichen: What you said indeed inspired me. DePIN is doing things in other fields off-chain, so the community can only introduce technologies from the external centralized world. Only in crypto-native areas like DeFi is there a possibility for new technologies to emerge within the community.

Steven: First, DeFi is related to finance, which naturally aligns with the technical characteristics of blockchain. Secondly, DeFi is about implementing engineering through code, while DePIN needs to interact with complex physical devices, which is not on the same level of difficulty.

9. Beichen: Finally, what do you think about the cycle of the DePIN narrative? Or how long do you think it will take for the market to validate its infeasibility?

Steven: From a technical perspective, both blockchain technology and the traditional technologies involved in DePIN are not mature enough yet. Using these technologies to carry such grand visions, to be honest, is almost hopeless. Only teams with real technical strength, like Filecoin, may see hope after continuing for another five to ten years.