If I said that the core background of io.net is "grassroots", what would you think?
has raised US$30 million in financing and has been favored by top capitals such as Hack VC, Multicoin Capital, Delphi Digital, and Solana Lab. It is not so "down-to-earth" no matter how you look at it, especially the label of superimposed GPU computing power/AI revolution, which is synonymous with high-end.
However, under the hustle and bustle of community discussions, key clues are often overlooked, especially the profound changes that io.net may bring to the global computing network—— Different from the "elite" positioning of AWS, Azure, and GCP, io.net is essentially taking the civilian route：
Misplaced and supplemented the neglected "waist + long tail" computing power needs, gathered idle GPU resources, and created an enterprise-level, decentralized distributed computing network, empowering more incremental/stock computing power resources for a wider range of small and medium-sized users' AI innovation, and realizing the "re-liberation of productivity" for global AI innovation with low cost and high flexibility.
Behind the AI wave, the neglected undercurrent of computing power production relations
What are the core productivity resources of this AI wave and the future digital economy era?
Without a doubt, computing power.
The global artificial intelligence hardware market is expected to grow at a compound annual growth rate (CAGR) of 24.3% to exceed $473.53 billion by 2033, according to Precedence Research.
Even if we put aside the forecast data and look forward from the logical perspective of increment and stock, we can clearly find that in the future development of the computing power market, there will be two main contradictions that will exist for a long time:
- In the incremental dimension, the exponential trend of computing power demand-side growth is destined to be far greater than the linear trend of computing power supply-side growth;
- Stock dimension, Due to the head effect, computing power has been "pinched", leaving low-end and long-tail players with no money to eat. However, a large number of distributed GPU resources have been idle, resulting in a serious mismatch between supply and demand；
Incremental dimension: computing power demand is far greater than supply
First of all, there is the incremental dimension. In addition to the rapid expansion of AIGC large models, countless AI scenarios in the early stages of the outbreak, such as medical care, education, and intelligent driving, are rapidly spreading, all of which require massive computing resources. Therefore, the gap in the current market for GPU computing resources will not only continue to exist, but even continue to expand.
In other words, from the perspective of supply and demand, In the foreseeable future, the market demand for computing power must be far greater than the supply, and the demand curve will still show an exponential upward trend in the short term.
On the supply side, due to limitations of physical laws and real production factors, whether it is process improvement or large-scale factory construction to expand production capacity, at most it can only achieve linear growth, which is destined to have a long-term computing power bottleneck in AI development. 
If you want to stay together for warmth, or if you have any questions, follow Gong Zonghao: Ten Yuan Knowledge Bureau
Stock dimension: There is a serious mismatch between supply and demand for waist and long-tail players
At the same time, with limited computing resources and facing serious growth bottlenecks, Amazon Cloud Technology (AWS), Microsoft Azure, and Google Cloud Platform (GCP) together account for more than 60% of the cloud computing share, and are in a clear seller's market.
They hoard high-performance GPU chips and monopolize a large amount of computing resources, while The small and medium-sized computing power demand parties in the waist and long tail not only have no bargaining power, but also have to face many problems such as high capital costs, KYC entry barriers, and leasing terms In addition, due to revenue ratio considerations, traditional cloud service giants will inevitably ignore the differentiated business demands of "waist + long tail" users (such as shorter, more immediate, smaller leasing needs, etc.).
But in fact, outside the computing power network of cloud service giants, a large amount of GPU computing power is idle and useless. For example, there are hundreds of thousands of third-party independent Internet data centers (IDCs) around the world that have a small waste of resources in training tasks, and even include massive computing power in crypto mining farms and encryption projects such as Filecoin, Render, and Aethir.
According to io.net official estimates, the current idle rate of IDC graphics cards in the United States alone is as high as over 60%, which creates an ironic paradox of supply and demand mismatch: More than half of the computing power resources of tens of thousands of small and medium-sized data centers, encryption mines and other operators are wasted on a daily basis and cannot bring in effective revenue. However, waist and long-tail AI entrepreneurs are enduring high-cost and high-threshold computing power services from cloud giants, and even more diverse innovation needs cannot be met.
Drought kills drought and waterlogging kills, these two basic premises are clarified. In fact, we can see at a glance the core contradiction between the current global AI development and the global computing power market - on the one hand, AI innovation is everywhere, and the demand for computing power continues to expand. On the other hand, the "waist + long tail" computing power needs and idle GPU resources cannot be effectively met, and they are outside the current computing power market.
This problem is not only the contradiction between the growing computing power demand of AI entrepreneurs and the backward growth of computing power, but also the contradiction between the majority of "waist + long tail" AI entrepreneurs, computing power operators and the unbalanced and inadequate mismatch of supply and demand, so it is far beyond the scope of the solution capabilities of centralized cloud service providers.
Because of this, market needs are also calling for new solutions. Just imagine, if these operators with computing power can flexibly choose to rent out computing power when they are free, can they obtain a computing cluster similar to AWS at a low cost?
It is important to know that it is extremely expensive to build a data network with such a large computing power. This has given rise to a computing power matching platform specifically targeted at idle computing resources in the middle and tail and small and medium-sized AI entrepreneurs, in order to mobilize these scattered idle computing resources and match them with small and medium model training and large models in medical, legal, financial and other segmented scenarios.
can not only meet the diversified computing power needs of the middle and tail end, but will also be A misplaced complement to the existing computing power service pattern dominated by centralized cloud giants：
- The cloud service giant with massive computing resources is responsible for "urgent and dangerous needs" such as large model training and high-performance computing;
- Decentralized cloud computing power markets such as io.net are responsible for more diversified "flexible and low-cost needs" such as small and medium-sized model calculations, large model fine-tuning, inference deployment, etc.;
is actually to provide a more inclusive dynamic balance supply and demand curve between cost-effectiveness and computing power quality, which is also more in line with the economic logic of market optimization of resource allocation.
Therefore, distributed computing power networks such as io.net are essentially a solution that integrates "AI + Crypto", that is, using a distributed collaboration framework combined with basic economic means of token incentives to meet the needs of the mid-tail AI market with huge potential but in exile. It allows small and medium-sized AI teams to customize and purchase GPU computing services that large clouds cannot provide, realizing the "re-liberation of productivity" in the global computing power market and AI development.
So to put it bluntly, io.net is not a direct competitor of AWS, Azure, and GCP. Instead, it is a "misplaced supplementary comrade" that works with them to optimize the allocation of global computing resources and jointly expand the market pie. It is just a front in charge of different levels of "cost-effectiveness & computing power quality" needs.
It is not even ruled out that io.net can create a market share that is no less than the size of the existing three cloud giants by aggregating the supply and demand players in the "waist + long tail".
io.net: Global GPU computing power matching platform
Precisely because io.net is based on Web3 distributed collaboration + token incentives to reshape the production relations of the middle and tail computing power markets, we can actually see the shadow of the sharing economy such as Uber and Didi in it, that is A matching trading platform for GPU computing power similar to Uber and Didi.
As we all know, before Uber and Didi did not exist, in a broad sense the user's "call and get" taxi-hailing experience did not exist, because several private cars formed a huge and disorderly network of idle vehicles. If you wanted to take a taxi, you could only wave and wait on the roadside, or apply to the corresponding taxi center company in each city for dispatch. This is time-consuming, highly uncertain, and it is also a strong seller's market, which is not friendly to most ordinary people.
And this is actually a true portrayal of the supply and demand sides of the entire computing power market. As mentioned above, The small and medium-sized computing power demanders in the waist and long tail not only have no bargaining power, but also have to face many problems such as high capital costs, KYC entry barriers, and leasing terms.
Looking at it specifically, how does io.net realize its positioning as a "global GPU computing power distribution center + matching market", or what kind of system architecture and functional services are needed to help mid- and long-tail users obtain computing power resources?
Flexible and low-cost matching platform
io.net’s biggest attribute is its asset-light computing power matching platform.
That is, like Uber and Didi, it does not involve the actual operation of heavy assets such as GPU hardware, which is extremely risky, but connects the supply of mid- and long-tail retail computing power (many of which are considered second-class computing power in large clouds such as AWS), and through matching, revitalizes the computing power resources that were originally idle (private cars) and the mid-tail AI demand (taxi-hailing people) that are in urgent need of computing power.
One end of io.net is connected to thousands of idle GPUs (private cars) such as small and medium-sized IDCs, mining farms, and encryption projects, and the other end is connected to the computing power needs of hundreds of millions of small and medium-sized companies (taxi hailers). Then io.net serves as a matching platform for intermediate scheduling, just like a broker matching countless buy and sell orders one by one.
This helps entrepreneurs train more personalized small and medium-sized AI models by gathering idle computing power, using low costs and more flexible deployment configurations, and greatly improves the utilization of resources. The advantages are obvious. No matter whether the market is too cold or too hot, as long as there is resource mismatch, the demand for matching platforms will be the strongest:
- On the supply side, suppliers of idle computing power resources such as small and medium-sized IDCs, mines, and encryption projects only need to connect with io.net. There is no need to set up a BD department, nor do they need to be forced to sell it to AWS at a discount due to the small scale of computing power. Instead, idle computing power can be directly matched with suitable small and medium-sized computing power customers at market prices or even higher prices at extremely low friction costs, thereby obtaining benefits;
- On the demand side, small and medium-sized computing power demanders who originally had no bargaining power in front of big clouds such as AWS can also use the io.net resource pipeline to connect to smaller-scale computing power that requires no permission, no waiting, no KYC, and more flexible deployment time. They can freely select and combine the chips they need to form a "cluster" to complete personalized computing tasks;
The computing power supply side and demand side in the middle and tail end have similar pain points such as weak bargaining power and low autonomy in front of big clouds such as AWS. io.net Revitalize the supply and demand of the waist and mid-to-long tail, and provide such a matching platform that allows both supply and demand parties to complete transactions at better prices and more flexible configurations than big clouds such as AWS.
From this perspective, similar to platforms such as Taobao, the emergence of inferior computing power in the early stage is also a development law that cannot be eliminated by the platform economy. io.net has also set up a reputation system for the supply side or the demand side, accumulating points based on computing performance and network participation, and obtaining rewards or discounts.
Decentralized GPU cluster
Secondly, although io.net is a matching platform between retail supply and demand parties, current computing scenarios such as large models require several graphics cards to perform calculations together—— It not only depends on how much idle GPU resources your matching platform can aggregate, but also how closely the distributed computing power on the platform is connected.
In other words, this distributed network covering small and medium-sized computing power in different regions and different sizes needs to implement a "decentralized but centralized" computing power architecture: several distributed GPUs can be put into the same framework for training according to the flexible computing needs of different scenarios, and ensure that different GPUs communicate and collaborate very quickly, at least to achieve features such as low latency that is sufficient for use.
This is completely different from the dilemma of some decentralized cloud computing projects that can only be limited to the use of GPUs in the same computer room. The technical implementation behind it involves the "troika" of the io.net product portfolio: IO Cloud, IO Worker, and IO Explorer.
- The basic business module of IO Cloud is Clusters, which is a GPU group that can self-coordinate to complete computing tasks. Artificial intelligence engineers can customize the desired cluster according to their own needs. It is also seamlessly integrated with IO-SDK to provide a comprehensive solution for expanding AI and Python applications;
- IO Worker provides a set of user-friendly UI interfaces that allow suppliers and buyers to effectively manage their supply operations on a web application, ranging from functions related to user account management, computing activity monitoring, real-time data display, temperature and power consumption tracking, installation assistance, wallet management, security measures and profitability calculations;
- IO Explorer primarily provides users with comprehensive statistics and visualizations of all aspects of the GPU Cloud. It enables users to easily monitor, analyze and understand every detail of the io.net network by providing complete visibility into network activity, vital statistics, data points and reward transactions;
Because of the above functional architecture, io.net allows computing power suppliers to easily share idle computing resources, greatly lowering the entry barrier. Demanders do not need to sign long-term contracts and endure the long waiting times common in traditional cloud services. They can quickly set up clusters with the required GPUs and obtain services such as super computing power and optimized server response.
Lightweight elastic demand scenario
To be more specific, when talking about the misaligned service scenarios of big clouds such as io.net and AWS, they mainly focus on the lightweight elastic needs of some big clouds that are not cost-effective. Among them, diversified scenarios that everyone can think of, such as model training in subdivisions and fine-tuning of large models for small and medium-sized AI entrepreneurial projects, are all included here.
In addition, there is a universally applicable scenario that is easily overlooked: model inference.
As we all know, the early training of large models such as GPT requires the use of tens of thousands of high-performance GPUs, super computing power, and massive data for long-term high-quality calculations. This is also the absolute advantage of big clouds such as AWS and GCP.
But after training, the main computing power requirement becomes the long-term model inference. The computing power requirement at this stage is much higher than the training stage—— Inference based on already trained models, that is, the daily dialogue and interaction scenarios between our ordinary users and models such as GPT, accounts for 80%-90% of the AI computing share.
What’s interesting is that the overall computing power of the inference process is smoother. It may only take dozens of GPUs and a few minutes to get the answer, and the requirements for network latency and concurrency are lower. At the same time, most AI companies may not train their own large models separately, but only choose to optimize and fine-tune around a few large head models such as GPT. These scenarios are naturally suitable for the distributed idle computing resources of io.net.
In addition to the high-intensity and high-standard application scenarios of the minority, the broader, daily lightweight scenarios are also a virgin land that urgently needs to be developed. It seems very fragmented, but the market share is even larger—— According to the latest report from Bank of America, high-performance computing accounts for only a small part of the data center total available market (TAM), only about 5% share.
In short, it’s not that AWS, GCP, etc. are unaffordable, but that io.net is more cost-effective.
The winner and loser of Web2 BD
Of course, in the end, the core competitiveness of io.net and other platforms for distributed computing resources lies in BD capabilities, which is the key to victory.
In addition to the miracle of NVIDIA’s high-performance chips giving birth to graphics card brokers, the biggest problem that plagues many small and medium-sized IDC and other computing power operators is that “the wine is good but the alley is dark.”
So from this perspective, io.net actually has an exclusive competitive advantage that is difficult to replicate for projects in the same track - there is a Web2 BD team directly based in Silicon Valley. They are all veterans who have been immersed in the business field of the computing power market for many years. They not only understand the diversified scenarios of small and medium-sized customers, but also understand the terminal needs of many Web2 customers.
According to io.net official disclosure, there are currently twenty or thirty Web2 companies that have expressed their willingness to purchase/lease computing power, and are willing to try or make mistakes because of lower-cost and more flexible computing power services (some may not be able to wait for computing power in AWS). Trial and error, and a single customer needs at least hundreds or thousands of graphics cards (equivalent to hundreds of thousands of dollars in computing power orders per month).
This kind of real terminal willingness to pay on the demand side, In essence, it will also attract more idle computing resources to actively pour in on the supply side, making it easier to take the lead in breaking the circle of Web2&Web3 and forming a first-mover network scale effect.
Tokenized computing power ecological settlement layer is possible
As mentioned above, io.net is based on Web3 distributed collaboration + token incentives to reshape the middle and tail computing power market, of which the latter mainly uses a dual-token model of IO and IOSD:
- Token IO, its utility includes paying computing power rental fees, giving IO Workers allocation incentives, rewarding AI and ML deployment teams for continued use of the network, balancing partial demand and supply, pricing IO Worker computing units, and community governance, etc.;
- The stable currency IOSD, pegged to the US dollar, can only be obtained by burning IO, aiming to provide a stable value storage and transaction medium for the io.net platform;
In addition, io.net is also considering supporting the supply side to increase the probability of being rented by mortgaging IO, and the demand side can also mortgage IO to prioritize the use of high-performance GPUs, thereby allowing the development of a complete ecosystem around the mortgage function to capture the incremental value generated by the entire computing power ecosystem.
This actually raises another question. Since io.net has gathered huge idle computing power resources, can it go further and directly combine Crypto’s computing power tokenization gameplay to give GPU greater on-chain financial possibilities?
For example, in the future, it is entirely possible for io.net to build a dedicated chain for computing power based on a huge computing power network. By providing tokenized infrastructure services without permission and entry barriers, anyone and any device can directly access Tokenize computing power (such as converting A100 and H100 into standardized tokens or NFT), thereby allowing trading, staking, lending, lending and leverage.
This is equivalent to creating a vast on-chain market for GPU computing power for users. Users and funds from all over the world can enter freely and efficiently. We can simply imagine two scenarios to get a glimpse of what kind of imagination space the on-chain financial market around computing power can have in the future.
1.Security computing power Token
For example, if a computing power operator on io.net owns several A100 or H100 graphics cards, but he has financial needs at this time or wants to secure them in advance, he can package the corresponding computing power value of this part of the graphics card into an NFT or FT on io.net—— The Token represents the discounted cash flow of the computing power of the corresponding graphics card in the next year (or a certain period of time), and can be priced in IO.
Since most ordinary investors do not have the opportunity to directly purchase, hold, and run AI computing power, this type of Token provides market traders with an opportunity to gamble on the rise and fall of computing power prices in the future, Operators with large amounts of computing power but tight cash flow have also gained financial leverage and can flexibly realize liquidity according to actual needs anytime and anywhere。
During , the graphics card behind the Token was operated by io.net, and the subsequent cash flow earned from the corresponding computing power was divided proportionally (Token holders receive 0.9, and operating nodes receive 0.1).
And because it is a standardized token, it can be freely circulated and traded in CEX or DEX like other Tokens. This will further form real-time computing power pricing with free entry and exit, and truly transform GPU computing power into a resource with global liquidity.
2. Bond computing power Token
In addition, you can also raise funds by issuing bond tokens to purchase high-performance graphics cards and increase network computing power. The bond principal is equivalent to the value of the graphics card equipment itself, and the bond interest is the cash flow income from leasing the graphics card computing power in the future, which means The potential computing power rental value and future income of the graphics card is the market value of the Token. By holding the Token, you can obtain real RWA income.
This is equivalent to creating a vast GPU computing power market for global users. Users and funds from all over the world can freely and efficiently enter the GPU computing power market without worrying about high thresholds and high funds. It further fully connects real graphics cards with various decentralized financial products, laying the foundation for more and richer supporting services for users in the future.
More importantly, the entire process uses IO as the main transaction/circulation currency, making io.net/IO expected to become the settlement layer/settlement currency of the entire global computing power ecosystem, and this prospect of an on-chain financial market around computing power tokenization can almost recreate a valuation space similar to the narrative of io.net's decentralized computing power network.
Summary
Generally speaking, Web3, as a new type of production relationship, is naturally adapted to AI that represents a new type of productivity. This is also the simultaneous advancement of technology and production relationship capabilities. From this perspective, the core logic of io.net is to change the production relationship between traditional cloud service giants, medium and long-tail computing power users, and global idle network computing resources by adopting the economic infrastructure of the "Web3+ token economy":
Provide solutions to the real pain points of AI computing power supply and demand, build a two-sided market that covers and serves the "waist + long tail" GPU computing resources & user needs, optimize the supply and allocation of computing resources, and bring global AI development, especially the liberation of productivity in small and medium-sized AI innovations.
’s vision is undoubtedly ambitious. If it succeeds, it will most likely become the core matching infrastructure and value settlement layer of the global GPU computing ecosystem. It is expected to obtain the best valuation premium. It is highly imaginative, but it is also undoubtedly full of challenges.
If you want to stay together for warmth, or if you have any questions, follow Gong Zonghao: Ten Yuan Knowledge Bureau