AI × Web3: Who will build the infrastructure of this era?

When a technological paradigm shift occurs, we often see the craze first, rather than the system. The current wave of AI is no exception.

As an early investor, I have always believed that betting on the transformative power at the core of the industry is far more valuable than chasing superficial narratives.

In the past year, I have examined a large number of projects such as RWA, Consumer, and infoFi. They are all exploring the intersection between the real world and on-chain systems. However, an increasingly obvious trend is that regardless of the project direction, there is a need to integrate the collaborative logic of AI to enhance competitiveness and efficiency.

For example, in the RWA field, we need to consider how to use AI for risk control optimization, off-chain data validation, and dynamic pricing in the future; similarly, in the Consumer and DeFi sectors, where user experience needs urgent optimization, AI is also required for user behavior prediction, strategy generation, incentive distribution, and so on. Other sectors also have similar demands.

Therefore, whether it is asset digitization or experience optimization, these seemingly independent narratives will ultimately converge on the same technological logic: if the infrastructure lacks the integration and carrying capacity of AI, it will not be able to support the complex collaboration of the next generation of applications.

In my opinion, the future of AI is not just about becoming "increasingly powerful" and "more widely used." The real paradigm shift lies in the reconstruction of collaborative logic.

Just like the early transformation of the internet, it was not because we invented DNS or browsers, but because it allowed everyone to participate in content creation for the first time, turning ideas into products, thereby giving rise to an entire open ecosystem.

AI is following a similar path: intelligent assistants will become co-creation partners for everyone, helping you transform expertise, creativity, and tasks into automated productivity tools, and even achieve monetization.

This is a problem that is difficult to solve in the current Web2 world, and it is also the underlying logic of my focus on the AI + Web3 track: making AI collaborative, fluid, and profit-sharing is the system that is truly worth building.

Today I want to discuss the only project that is currently attempting to systematically build the underlying operations of AI from a chain-level structure: Sahara.

The essence of investment is a worldview, recognizing the value system of choices.

My investment logic is not simply to combine the narrative of public chains with AI and then look for well-backed teams to invest in.

Investment is essentially a choice of worldview. I have always been questioning a core issue: Can the future of AI be jointly owned by more people?

Can it use blockchain to restructure the value attribution and distribution logic of AI, allowing ordinary users, developers, and other roles to participate, contribute, and continuously benefit? In short, I only believe that relevant projects have the potential to be disruptors if this logic emerges, rather than just being "another abandoned public chain."

To find the answer, I examined almost all the AI projects I could access until I encountered Sahara. The answer given to me by Sahara's co-founder Tyler was: to build an open, participatory ecosystem that everyone can own and benefit from.

This sentence, although simple, precisely hits the soft spot of traditional public chains: they often serve developers in a one-sided manner, and the design of token economics is largely limited to Gas fees or governance, rarely able to truly support a positive cycle of the ecosystem, making it even more difficult to sustain the development of an emerging track.

I am well aware that this road is filled with challenges, but precisely because of this, it is a revolution that cannot be refused - and that is the reason for my firm investment.

As I emphasized before in my discussion on "The Evolution from Web2 to Web3": the true paradigm shift lies not in creating a single product, but in building a supportive system.

And Sahara was one of the cases I was most looking forward to at that time.

From investment to 8 times valuation follow-on heavy investment

If I say that my initial investment in Sahara was because it is pursuing what I truly believe to be the leading mission of AI – building an AI economy and infrastructure system. Then let me tell you that in just six months, I invested again at an eightfold pre-round valuation because I felt an extremely rare strength in this team.

Two co-founders, one of whom is the youngest tenured professor at the University of Southern California, specializing in AI. The value of a tenured professor from a top American university born in the 90s is reflected not only in the academic field but also in their dreams, energy, and the determination to achieve those dreams at such a young age. Having known Professor Ren for over a year, I have witnessed what it means to be a genius who can work for more than ten hours a day, with stable emotions and humility.

Another Tyler, who served as the Director of North American Investment and Incubation at a well-known investment institution, is undoubtedly knowledgeable about Web3. His level of self-discipline is astonishing: he only sleeps in multiples of 1.5 hours, insists on working out to maintain his condition no matter how busy he is, and even avoids sugar to keep his mind clear, working over 13 hours a day. I once joked that he was like a robot, and he simply responded calmly: "I am lucky to have the busy schedule I have today." His dopamine comes from advancing project progress every day; dreaming is his passion, and he doesn't need any other fuel.

Meeting them also changed me. I started to maintain a regular routine as much as possible, my emotions gradually stabilized, and I began to work out...

So when someone says that Sahara gained the favor of capital due to luck, I always unreservedly add: "The pursuit of capital is an inevitable result." I vividly remember the difficulty of primary financing in this market cycle, yet Sahara was being chased for investment by the primary market.

What everyone remembers is that certain well-known investment institutions have invested in Sahara. Sahara has opened the investment era for a large technology company to enter the Web3 AI field, and its receipt of the company's AI award is a significant reason for the investment. In addition, some AI-heavy funds and national banks are also guests of Sahara. What you can see is a group of institutions that are more focused on traditional technology and industrial resources, beginning to quietly bet on AI × Web3 because of Sahara.

Capital will only pay for certain directions and execution capabilities - this is positive feedback on the depth of Sahara technology, the team's background, system design, and execution ability.

This is also why it can produce some real and solid structural indicators:

More than 3.2 million accounts have been activated on the test network, and there are over 200,000 data platform annotators. ( millions are queued, and their clients include several leading technology companies, with revenue reaching tens of millions of dollars.

On this infrastructure chain, at least from "who will do it" to "can it be done", Sahara has already gone deeper and more steadily than 99% of "AI concept projects".

![AI × Web3: Who Will Build the Chain for This Era?])https://img-cdn.gateio.im/webp-social/moments-cf477a069888f7e6e1a56aa0d7248930.webp(

The ultimate challenge of public chains: to ensure that all contributors continue to benefit and drive positive economic cycles.

Returning to our initial judgment logic: In a system that combines AI and blockchain, is there really a mechanism that can allow every contributor to be seen, recorded, and continuously rewarded?

Model training and data optimization rely heavily on extensive support from annotations and interactions; conversely, if there is a lack of user contributions, the project itself must invest more funds to procure data and outsource annotations, which not only increases costs but also weakens the value-driven community collaboration.

Sahara is one of the few Web3 AI projects that allows ordinary users to "participate in data construction from day one." Its data labeling task system operates daily, with a large number of community users actively involved in labeling and prompt creation. This not only helps improve the system but also invests in the future with data.

Through the mechanism of Sahara, not only is the quality of the model improved, but it also allows more people to understand and participate in this decentralized AI ecosystem, linking data contributions to rewards and forming a true positive cycle.

A typical example is a voice project on a public blockchain, which quickly built a high-quality dataset covering multiple languages and accents by leveraging Sahara's decentralized data collection and human-machine collaborative annotation, significantly improving the training efficiency of its TTS and voice cloning models. This also propelled its open-source project to gain thousands of GitHub stars and over 2 million downloads.

At the same time, users participating in data annotation also received token rewards issued by the project, forming a two-way incentive loop between developers and data contributors.

Sahara's "permissionless copyright" mechanism ensures the open circulation and reuse of AI assets while safeguarding the rights of all participants—this is the underlying logic driving the explosive growth of the entire ecosystem.

Why is this considered a scenario with long-term value support?

Imagine if you want to build an AI application, you naturally hope your model is more accurate and closer to real users than others.

The key advantage of Sahara is that it connects you to a vast and active data network—hundreds of thousands, and in the future millions of annotators. They can continuously provide you with customized, high-quality data services, allowing your model to iterate faster.

More importantly, this is by no means a one-time transaction. Through Sahara, you are connecting to a potential early user community; and these contributors are very likely to become the real users of your product in the future.

This connection is not a one-time buyout. Through Sahara's smart contract system and rights confirmation mechanism, it enables a long-term, traceable, and sustainable incentive system.

Regardless of how many times the data is called, contributors will receive continuous profit-sharing, with earnings dynamically linked to usage behavior.

But this is not just a revenue model for data labeling and model training stages. Sahara builds an economic system that covers the entire lifecycle of AI models. In every aspect after the model goes live, including calls, combinations, and cross-chain reuse, a profit-sharing mechanism is also built-in, allowing value to be captured over a longer period.

Model developers, optimizers, validators, computing power contribution nodes, etc. can now continuously benefit at different stages, rather than just relying on a single transaction or buyout.

Such a system brings a compound effect for model combination calls and cross-chain reuse. A trained model, like building blocks, can be repeatedly called and combined by different applications, and each call creates new revenue for the original contributor.

Because of this, I agree with Sahara's underlying belief: a truly healthy AI economic system cannot be just about data plunder and model buyouts, nor can it be just about a few people reaping all the benefits. It must be open, collaborative, and mutually beneficial—where everyone can participate, every valuable contribution can be recorded, and continuous rewards can be obtained in the future.

![AI × Web3: Who will build the chain for this era?])https://img-cdn.gateio.im/webp-social/moments-f100e78f0ea1234e20c00aebaa95457a.webp(

But the closer we get to the true structure, the more challenges there are.

While I am optimistic about Sahara, I will not cover up the challenges the project will face because of my investment position.

One of the major advantages of the Sahara architecture is that it is not limited to any specific chain or single ecosystem.

Its system was designed from the beginning to be open, full-chain, and standardized: supporting deployment on any EVM-compatible chain, while also providing standard API interfaces that allow Web2 systems—whether e-commerce backends, enterprise SaaS, or mobile apps—to directly invoke Sahara's model services and complete on-chain settlement.

However, despite the extreme scarcity of this architectural design, there is a core risk: the value of the infrastructure lies not in "what it can do," but in "who is willing to do what based on it."

To become a trusted, adopted, and integrated AI protocol layer, the key for Sahara lies in how ecosystem participants assess its technological maturity, stability, and future predictability. Although the system has been built, whether it can truly attract a large number of projects to land based on its standards remains uncertain.

It is undeniable that Sahara has achieved key validation: providing relevant data services to several leading technology companies and addressing some of the industry's most challenging data demand issues, becoming an early signal of the feasibility of this system.

But what needs to be seen is that these collaborations mainly come from the Web2 world, which truly determines Sahara.