Over the past year, AI has stepped into the spotlight of the Internet stage, setting off an unprecedented craze.

However, behind the hot concept of generative AI, the venture capital activities in the entire AI field are rapidly cooling down. According to a report from the Stanford University Artificial Intelligence Research Institute, the peak of AI-related merger and acquisition investments occurred in 2021. In the past three years, the transaction amount of related investments has actually been declining year by year.

The reasons are not difficult to understand. AI has become a game for giants. Whether it's Microsoft, Google, Meta, or Apple, they are all procuring chips and collecting data for training on their own, and only these companies can gain advantages in computing power and data. For startups, it is becoming increasingly difficult to launch an AI project from scratch.

In the eyes of most people, this is probably just a problem in the vertical field of AI. But is it possible that the "oligopolistic" trend presented in the AI venture capital field hides a larger underlying problem that has long existed since the invention of the Internet?

For example, when we access the Internet, why is it difficult to obtain a public IP address? Why are IPv4 address resources becoming increasingly scarce? Why are various data leaks and privacy infringements constantly occurring, and we are reminded to change various passwords every once in a while? Why are we having to buy more and more Internet services, from video websites and music memberships to cloud storage, and the burden is getting heavier?

Behind these specific but fragmented problems, it is actually the problem of the design of the Internet architecture. It can be said that the Internet, which is inseparable from our lives, does not fully adapt to today's network scale and application forms at its underlying design, and it needs a complete overhaul.

Underlying Challenges of Traditional Internet

To understand the underlying problems of the traditional Internet, the AI field is a good entry point.

AI has high demands for data collection, storage, and computation, and the cost is extremely high, which is the core factor that has led to the rapid oligopolization of the AI field in its early development.

The shortage of computing power is well known. AI training requires a large amount of GPU computing power, and this demand continues to grow with model iteration and parameter increase. The rush to purchase the most advanced GPUs from NVIDIA has continued for a year, driving NVIDIA's stock price to triple in the past 12 months. Many people have sighed that computing power is the most important resource in the AI era.

In addition to computing power, data is equally important. One of the biggest obstacles to the iteration of large AI models is data. In the past, large model manufacturers obtained data from the public Internet to build corpora. However, as many social media sites have begun to block AI companies from crawling data, it has become increasingly difficult to build a high-quality corpus. And the amount of data needed for training the next generation of models is increasing.

Taking GPT as an example, according to The Wall Street Journal, the total size of the corpus used in the training process of GPT-4 is about 120 trillion tokens, and the data requirement for GPT-5 is between 60-100 trillion, an increase of 5-8 times. It is actually very difficult for a single company to obtain data of this scale. Moreover, 99% of the data on the current Internet is produced by users, but the ownership of this data does not belong to the users, which is undoubtedly unfair.

Not only AI, with the improvement of Internet infrastructure, the volume of all data is rapidly expanding, from images, audio, and video to spatial models. The increase in data volume has led to high costs for servers and CDNs, and the cost of users using the network has also risen sharply.

The manifestation of this cost is ubiquitous, from the rising membership fees of video websites to the cost of 5G plans, and to the increasing size and cost of phone storage space and cloud storage services, as well as the extremely high cost and threshold of AI training today. The Internet is not only becoming more centralized and oligopolistic, but it is also becoming increasingly expensive, and users have lost their voice throughout the process.

With such high costs for data exchange and such low efficiency, many people will ask: Why can't we reuse the concept of P2P downloads from the past, allowing more people to participate in building the network and contribute their idle computing power, storage, and communication resources?

The core of the problem lies here.

The underlying design of the Internet we use today is a radial tree structure from the beginning. The upper layer is the server, the middle layer is the CDN node for auxiliary data distribution, and the bottom layer is the terminal. Although two terminals can establish a direct connection through the network, such connections are still few, and the vast majority of data still only flows between the terminal and the server.

So, the core logic of Web 1.0 is "distribution," and the flow of information on the Internet is one-way, more like television or radio, except that the content being distributed has become a blog post, a piece of music, or a game. The relationship between the server hosting information and the terminal requesting information is one-to-many.

By the time of Web 2.0, user-generated content (UGC) emerged, and the flow of data began to become bidirectional, but it was still only the flow between the terminal and the server, not truly networked. Because until the era of Web 2.0, the number of terminals connected to the Internet globally was not enough, and the terminals that could remain online for a long time were not the majority. Servers were needed as the cornerstone of the network.

Friends with a longer "Internet age" may easily understand where the problem lies. Just like when we used P2P download software in the early years, we needed someone to "seed" in order to achieve high download speeds. P2P downloads are a completely decentralized network structure, but at that time, we often did not want to keep the download software running in the background of the computer to "seed" for others because there was no benefit, and our computers were probably turned off most of the time.

In recent years, everyone has been talking about the Web3 revolution, which aims to break the "tree-like" centralized network structure of the past and promote the interconnection of terminals, ultimately forming a decentralized network. In the Web3 era, one of the biggest changes is that there are now more terminals that can provide resources online for a long time, and our routers at home, our phones are no longer "turned off." On the other hand, the various tokens under Web3 provide incentive mechanisms for all "shared resource" behaviors.

To some extent, for the problem of "low efficiency" of the Internet, Web3 has provided a complete theoretical foundation and has the background conditions for change.

In fact, "low efficiency" is just one aspect of the problem of the traditional Internet. The basic structural framework of the Internet we use today was established in the 1990s. At that time, engineers actually found it difficult to imagine that the Internet would become as popular as it is today. The series of network problems we face today, such as the shortage of IPv4 address resources and the easy hijacking and pollution of DNS servers, can all be traced back to the underlying design flaws of the traditional Internet.

For many years, there have been voices within the industry calling for the promotion of updates and iterations of the underlying framework of the Internet, but because the switching costs are too high, it requires the concerted efforts of various countries and vertical fields around the world to achieve fundamental solutions, so it has never been fundamentally resolved. Even IPv6 has never been fully popularized.

It can be said that the Internet "ship" we are sailing on today is already a "patched" ship, and it is becoming increasingly unable to meet the needs of the times. Ultimately, someone must shoulder the mission of promoting a bottom-up update of the Internet.

Remodeling the Internet with the Idea of "Decentralization"

More and more people are realizing that to remodel the Internet is to remodel its excessive centralization, which has led to a monopoly in the hands of a few giant companies and organizations. In the Web3 era, a large number of developers and companies have begun to focus on this problem.

From Bitcoin to NFT, the entire Web3 industry seems to have been focusing on applications, transactions, contracts, and authentication… These applications constitute the first wave of the Web3 ecosystem.

However, the idea of decentralization is not only able to solve specific application problems, but also has the opportunity to thoroughly remodel the underlying technical stack of the Internet, overturning the efficiency and security problems caused by the excessive centralization of the current Internet.

DePIN has emerged. Although it seems unfamiliar on the surface, it hides the biggest ambition in the Web3 field. These developers are trying to build a "decentralized physical network infrastructure."

Many DePIN projects, from the beginning, are designed to solve a series of legacy problems existing in the physical, network, and application layers of the traditional Internet. For example, a leading startup company in this field in China, SendingNetworks, has proposed a comprehensive solution from top to bottom.

At the physical layer, they have designed a decentralized router hardware to solve the problem of excessive centralization of computing power, data, and communication resources in the traditional Internet framework. Through a large number of decentralized routers in the network, data can be stored and distributed, utilizing the idle bandwidth of each user for transmission and access, thereby improving circulation speed and efficiency.

At the network layer, Sending Networks has addressed the most challenging issue of IP address allocation in the traditional Internet TCP/IP protocol. In the new network framework, users can directly access the Internet using their wallet address, which is the natural "ID" of the Web3 era. Based on the wallet address, communication protocols will also become more secure and reliable, avoiding the common attacks through IP vulnerabilities.

Based on the decentralized physical and network layers, a variety of decentralized application layers can be spawned.

For example, at the most basic level, deploying a website on a network based on wallet addresses can bring faster access and greater security due to its decentralized nature, making it difficult to attack. Developers can also build a CDN with a larger network of nodes based on a decentralized network, utilizing the idle bandwidth resources of each terminal to reduce content distribution costs and increase efficiency. Additionally, they can build applications such as streaming media, instant messaging software, and more, all leveraging the core advantages of DePIN: efficiency, privacy, and security.

Such a fundamental revolution ultimately has the potential to drive AI applications to break free from the control of a few giants and return to the hands of every participant on the Internet.

Today, major companies collect user corpus data from various UGC platforms such as Reddit for AI training, forming a corpus and then centrally conducting model training. However, in a decentralized basic network, data generation, storage, model construction, training, validation, and eventual application can all be carried out in a single network, with everyone participating, supervising, and sharing the results.

In this environment, we do not need to fear that AI technology will be dominated by a few giants, used to exploit users, or used for malicious purposes. All data and models will be accommodated in a more transparent and secure environment.

Currently, there are the most decentralized AI projects in the DePIN track, as the computing power and data gap in AI training has brought the problem of excessive centralization of the Internet into the spotlight. However, a group of developers, including Sending Networks, still believe that only by completely transforming the entire chain of the Internet from the physical, communication, to application layers from the ground up, can it be propelled into a new future.

The First Step Towards the Future

Of course, this is not the first time someone has proposed to "overturn the Internet." While painting a beautiful future for DePIN, the bigger problem still looms overhead: how to turn it into reality?

The development and popularization of the traditional Internet have been a one-way process from top to bottom. A very small number of computer scientists designed the underlying framework of the Internet in the early days, and then more and more companies, developers, and users participated in building the ecosystem, leading to the widespread adoption of the Internet.

It is precisely because of this that the entire operating mechanism of the traditional Internet, despite being outdated and obsolete, is still difficult to shake and increasingly difficult to change.

Therefore, the most urgent need to transition the entire Internet world to the future of DePIN is to create a new "motivation."

When the Internet has already covered the vast majority of people on Earth, the revolution of the Web3 era cannot be completed overnight from top to bottom as in the past, but must be from the bottom up, driven by the first batch of seed users to transform developers and businesses, gradually expanding into a transformation of the entire industry. Therefore, the core of all Web3 protocols and applications is about "incentive mechanisms."

Sending Networks has chosen router products as the cornerstone of the cold start of the entire decentralized network, building a decentralized physical network and constructing the future of DePIN through the financial attributes and incentive mechanisms of Web3.

The biggest reason for choosing router products as a breakthrough is that routers are products that truly penetrate all levels of the Internet. In this future Internet framework, routers are not only the information exchange terminals for user access to the Internet, but also edge nodes that can remain online for a long time in a decentralized network, and can integrate and utilize storage and computing resources within the local area network.

What binds all of this together is the "consensus layer" designed by Sending Networks. If users can ensure the availability of their routers as edge nodes for a long time and provide more communication, computing, and storage resources, they can be transformed into quantifiable "contribution proofs" on the blockchain and ultimately receive rewards.

The early DePIN network will be fully compatible with the traditional Internet, ensuring that users' basic experience will not be compromised, but the flame of the decentralized network will be ignited. Data flowing between decentralized terminals will increase, and its value will be increasingly reflected, until the revolutionary fire ignites the entire Internet infrastructure.

Many people may also be curious about how Sending Networks, as a startup company, can drive such a fundamental change. In fact, the core team of Sending Networks originated from the early Microsoft Windows development team, and during the era of mobile Internet, they also developed the Dolphin Browser, which had an active user base of 200 million.

For the past 20 years, the core team of Sending Networks has been dealing with the most fundamental technologies of computers and the Internet, solving the problem of "how information flows, is distributed, and presented." It's just that in the Web3 era, they gradually realized that to change the distribution and presentation of information, they must transform the Internet itself.

And now, they have taken the first step towards this ideal.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。