Author: Yolo Shen; Source: Foresight Ventures
Traditional computers consist of five components: computer, memory, controller, bus, and I/O. Looking at the development of blockchain, the advancement of computers and memory is relatively complete. If we compare the entire distributed system to a human, then the human brain and memory system are already complete, but the perception and sensory systems are still in a very primitive state. At the current stage, Depin is undoubtedly the most popular buzzword. How can it be achieved? All of this undoubtedly needs to start with “trusted touch,” as we all know that “sensing” requires the spine and nervous system to process.
If we say that the blockchain system is like building consciousness on an iceberg, then the sensor network represented by the underlying layer of Depin is the subconscious under the iceberg. Now the challenge arises – who is the spine and nerves of the distributed system? How do we build the spine and nerves? In this article, we will start with the lessons learned from the development of the Internet of Things, construct the development ideas of Depin, and help builders land better.
I. A Review of the History of the Internet of Things
Looking back at the history of the Internet of Things since 2015, there were two main difficulties at that time: one was the single input and output of hardware devices. The second was that after the devices were connected to the network, the product features were not enhanced, and the devices did not have the characteristics of scalability.
During this period, the core issue was: what changes would occur when the microcontrollers of hardware devices were connected to the network? From the perspective of networking, it made it possible for hardware devices to upload and download. The next question was, why do hardware devices need to upload and download? Can uploading and downloading increase product competitiveness? At that time, we saw a batch of products such as smart curtains, smart air conditioners, etc., due to the relatively determined I/O wiring of hardware in the design phase, the space for software development was relatively limited. Therefore, after connecting to the network, the product features only added a feature of mobile control, such as “remote air conditioning, remote curtain pulling,” with functions mostly being remote + traditional controllers. For C-end users, this design was indeed somewhat redundant. Another core issue was whether IoT devices have the ability to scale after being connected to the network. As we mentioned earlier, with more uploading and downloading after hardware is connected to the network, if downloading is equivalent to upgrading and expanding functions, then uploading is data aggregation and integration. The value of the latter data lake was very cumbersome in the early days of the IoT era, with a huge contradiction between the exponentially rising storage costs and the difficult-to-exploit data sales.
In summary, IoT devices, in both downloading and uploading modes, fail to enhance product power and service dimensions. So, can the Depin era achieve this?
How Does AI Bring Change?
From the characteristics of AI, we see many possibilities:
Human-like, independent upload and download requirements. If the edge side cannot reason large models, then the end-side needs independent networking. This will transform the radial structure from mobile end as the center and devices as satellites to devices independently networking communication structure.
Device sovereignty. Moving from simple product sales to user purchases + data sales dual-drive. Devices are responsible to users as a whole and to data merchants as a sensor collection.
“Data trustworthy, privacy reliable” is a prerequisite for ordinary devices to transform into mining machines. If the data is not trustworthy, then logically, opening multiple virtual machines can hack the entire incentive system; if privacy is not reliable, then in the long run, the user’s interaction willingness will be suppressed.
Combining the development of AI, we see that Depin may have some different possibilities:
The emergence of AI has increased the necessity for AI hardware to independently network. The cost of device networking may rapidly decrease in the next three years, combined with the decrease in storage and computing costs, the cost of edge computing/sensor deployment is also likely to decrease significantly. So, when many devices have already been deployed, converting them into mining machines to collect sensor data will reach a tipping point.
After solving the problem of devices independently linking to the cloud, there will be more scenarios for device-to-device interconnection. How to use NFC and various low-cost hardware for interaction play will also become a potential point of innovation.
Commercializing various collected sensory data is the core bottleneck of device mining. Establishing standards for abstract information goods is the main challenge.
II. Investment Themes and Perspectives of Depin:
Based on the development experience of the Internet of Things over the past 5 years and the changes in AI characteristics, we believe there are three major investment themes:
Hardware infrastructure centered around cellular modules.
Abstract communication layer services with communication information goods as a bulk commodity.
Broad miners as a form of distributor services.
Investment Theme One: Depin Infrastructure Centered Around Address Bus Modules
What is a module?
A module is a functional module that integrates baseband chips, memory, power amplifiers, etc., on a circuit board and provides standard interfaces. Various terminals can achieve communication functions through wireless modules. With the development of the entire computing network, the definition of modules continues to evolve, forming a cellular networking + computing power + edge application ecosystem:
Traditional cellular IoT modules: Basic connection modules, mainly for cellular communication, these modules only contain chipsets that support this type of connection, without additional functions.
Smart cellular IoT modules: In addition to providing connection functions like traditional modules, they also integrate additional computing hardware in the form of central processors and graphics processors (CPU and GPU).
AI cellular IoT modules: In addition to providing the same functions as smart cellular IoT modules, they also include specialized chipsets for AI acceleration, such as neural, tensor, or parallel processing units (NPU, TPU, or PPU) for AI acceleration.
From the perspective of the entire industry chain, upstream chipmakers and downstream device manufacturers occupy most of the value chain. The module layer in the middle has high market concentration and low margins as features. Traditional service devices mainly include PCs, phones, and POS machines. Due to its huge concentration, once a module layer with widespread consensus is deployed, it naturally migrates various existing devices into mining machines. If traditional Web3 users are individuals, then the middle layer represented by modules will enable a large number of smart devices to enter Web3, and the transactions between these devices will generate a significant amount of on-chain demand.
Looking back at the competition between NVIDIA and Intel in the early days, we gained a lot of historical experience: In the early years, the computer chip market was dominated by the Intel CPU X86 system. In some edge markets, such as the graphics acceleration field, there was competition between the acceleration card ecosystem led by Intel and NVIDIA’s GPU. In the broader market (areas with uncertain demand), Intel CPUs collaborated with NVIDIA GPUs. The two companies thrived together for a period. The turning point came with Crypto and AI, where a large number of computing tasks were characterized by small tasks being highly parallel, which matched the computational characteristics of GPUs. As the tide approached, NVIDIA made preparations on several dimensions:
Cuda’s parallel computing instruction set helped developers better utilize GPU hardware.
Fast iteration capabilities. Surpassing Moore’s Law in iteration speed earned them survival space.
Cooperative competition with CPUs. Effectively leveraging and utilizing Intel’s existing resources, they quickly seized market opportunities in some decision-sensitive areas.
In the module market competition, the Crypto Stack is undoubtedly the best technology stack for building protocols and ecosystems. The migration of existing devices to cash flow mining machines will create beta-level opportunities. Dephy is an important player in this, achieving distribution management responsibilities for the entire Depin network by integrating modules, ledgers, and identity layers.
Investment Theme Two: Data Bus – Sensor-Represented Data Collection Mining Machines
What exactly is a mining machine? We believe that hardware/software that can produce specific information resources and intend to obtain token resources can be called a mining machine. In this understanding, a mining machine has several criteria:
Does it generate specific information resources?
Is it able to settle tokens?
Therefore, in the entire process, the production of specific information resources PoPW (proof of physical work) by devices becomes very important. Therefore, we believe that every sensor producing PoPW data must have a trustworthy (TEE/SE) to ensure the credibility of edge-side data collection. In the field of sensors, they can create various horizontally scalable networks, for example, video resources collected by different devices’ cameras will be standardized and measured in a network. Compared to independently collecting from different devices, horizontally scalable sensors + reliable modules can build a larger PoPW resource market, and collected video materials can be better priced according to a unified measure, which is more conducive to forming a bulk market of information resources. This is something that Device-Focus does not possess.
Investment Theme Three: Control Bus – Broad Bus Communication Infrastructure
Since some Depin devices exist in the physical world and are related to traditional commercial society, and the Crypto world is characterized by PermissionLess, it is crucial to manage various participating entities in a real-time manner without KYC. We believe that the entire Web3 world needs a communication abstraction layer that integrates cellular networks and public IP networks, where users/devices only need to pay Crypto currency to access corresponding network services. Specific directional race tracks include:
Integrating traffic. Bridging global operator traffic resources, considering traffic as a bulk information commodity traded with tokens.
Integrating number segments. Bridging global number segment resources, considering numbers as an identity layer traded with tokens, with Blockchain as the governance system.
Integrating IP resources. Bridging public IP resources, integrating the public IP pool as a resource, using public IPs as a resource pool, supporting any access redirection, trading with tokens for pricing, with Blockchain as the governance system.
III. Conclusion
Depin should not be based on devices as units, as devices do not have horizontal scaling capabilities. The core of Depin lies in Pin, and the core of Pin is the authorization code. We view devices as collections of sensor modules, and the pin code of each sensor module is both the permission for data networking and the authentication permission for PoPW. Only devices with networking permissions, contributing recognized devices, can be called mining machines. Therefore, the core of the entire Depin track lies in how to make edge devices contribute in a measurable way, so that the contributions of different devices with the same sensors can have a consistent measure.
Different from the transmission of information in traditional computers, it can be divided into three categories: transmitting various data information (Data Bus), transmitting various address information (Address Bus), and transmitting various control signals (Control Bus). Similarly, the DePin bus will also exist as: an identity credential for device networking (Address Bus), a PoPW credential for data verification (Data Bus), and a means of device management (Control Bus).
Due to the presence of Real World Assets (RWA) attributes, Depin projects exist in the physical world and are related to real economic life. Therefore, more real-time management methods are needed to achieve autonomous risk control. The main implementation channels are twofold: first, by governing the flow of cellular operator traffic, once a device violates the rules, it can revoke the mining rights of the device from the flow end, which is a more real-time management method compared to Slash. Second, through the method of buying out upstream resources through miners + resource pools. For example, if a dealer has 100 number resources and when 30 are at risk, a warning may be issued to suspend or revoke the license. Today, we will mix these 30 resources with resources from other dealers and buy out real resources (RWR) through miners. By mixing the number segments, the resources can be managed for risk control, while ensuring as many resources as possible are obtained while safeguarding the risks of upstream dealers. Replicating the Liquity model in various types of RW resources.
Disclaimer: All articles from Foresight Ventures are not investment advice. Investment involves risks, so please assess your personal risk tolerance and make investment decisions carefully.
