The Infrastructure Revolution of the AI Era: Who Will Build the Future On-Chain Systems?

When the technological paradigm truly shifts, we often first see the hype rather than the system. The AI wave we are experiencing is no different.

As a first-level investor, I have always believed that betting on the transformative power deep within the industry is far more valuable than chasing superficial narratives. In the past year, I have come across numerous projects in RWA, consumer goods, and information finance, all of which are exploring the intersection of the real world and on-chain systems. However, the increasingly evident trend is that regardless of the path a project takes, it ultimately needs to enter the collaborative logic of AI, leveraging AI to enhance competitiveness and efficiency.

For example, RWA projects need to consider how to use AI for risk control optimization, off-chain data verification, and dynamic pricing; consumer or DeFi projects require AI to complete user behavior prediction, strategy generation, and incentive distribution, among others. Whether it's asset digitization or experience optimization, these seemingly independent narratives will ultimately converge on the same technological logic: if the infrastructure does not possess the ability to integrate and support AI, it will not be able to sustain the complex collaboration of the next generation of applications.

In my opinion, the future of AI is not just about becoming "stronger and stronger" and being "used more and more"; the real paradigm shift lies in the reconstruction of collaborative logic. Just like the early transformations 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 also on this path: intelligent agents will become a co-creation body for everyone, helping you turn expertise, creativity, and tasks into automated productivity tools, even realizing monetization. This is a question that is difficult to answer in today's Web2 world, and it is also some of the underlying logic I focus on in the AI + Web3 track: making AI collaborative, transferable, and profit-sharing is the system that is truly worth building.

Today I want to discuss the only project so far that attempts to systematically build the underlying operation of AI from an on-chain structural perspective: Sahara.

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

My investment logic is not simply to combine public chain narratives with AI and then look for teams with seemingly good backgrounds to bet on. Investment is essentially a choice of worldview, and I always ask a core question: Can the future of AI be jointly owned by more people?

Can it leverage blockchain to reconstruct the value attribution and distribution logic of AI, allowing different roles such as ordinary users and developers to participate, contribute, and continuously benefit? Only with the emergence of this logic do I believe that such projects have the potential to be disruptors, rather than "just another public chain."

To find the answer, I basically scanned through all the accessible AI projects until I encountered Sahara. The response from Tyler, the co-founder of Sahara, was: To build an open, participatory ecosystem that everyone can own and benefit from.

This sentence is simple, but it 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 mostly limited to Gas Fee or governance, rarely able to genuinely support a positive cycle of the ecosystem, let alone sustain the development of an emerging track.

I am well aware that this road is full of challenges, but precisely because of this, it is an irresistible revolution - which is also the reason for my firm investment. As I emphasized when discussing the "evolution from Web2 to Web3": the true paradigm shift is not about creating a single product, but about building a supportive system. And Sahara is one of the most anticipated cases I predicted at that time.

From investment to 8 times valuation follow-up investment heavy position

If I initially invested in Sahara because it was fulfilling what I believe to be the true leading mission of AI—building an AI economy and infrastructure system, then what has driven me to rush in and invest at 8 times the pre-round valuation in just half a year is the rare strength I have felt 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 in an American university born in the 90s is reflected not only in the academic field but also in the fact that at this age, they still have dreams, energy, and the determination to realize those dreams. Knowing Professor Ren for more than a year has shown me what it's like to be a genius who can work for more than ten hours a day, with stable emotions and humility.

Tyler, the former investment director of a well-known investment institution, is responsible for North American investments and incubators, and his understanding of Web3 goes without saying. He is astonishingly self-disciplined: he only sleeps in multiples of 1.5 hours, insists on working out no matter how busy he is to maintain his condition, and doesn’t touch a single piece of candy for a clear mind, working over 13 hours a day. I once joked that he is a robot, to which he simply replied: "I am lucky to have such a busy life today." His source of dopamine comes from pushing project progress every day; dreaming is his passion and doesn’t require any other fuel.

I am very glad to have met them, which changed me. I have also finally started to sleep regularly as much as possible, my emotions have gradually stabilized, and I have begun working out...

So when someone says that Sahara gained the favor of capital because of luck, I always candidly add, "The pursuit of capital is an inevitable result." I vividly remember how difficult primary financing was in this round of the market, but Sahara was being chased for investment by the primary market.

What everyone remembers is that some well-known investment institutions have invested in Sahara. Sahara has opened the investment era for a major technology company entering the Web3 AI field, and its receipt of the AI Grand Prize from that company is a significant reason for the investment. In addition, some AI-heavy funds and national banks are also guests of Sahara. You can see that a group of institutions that are more focused on traditional technology and industrial resources are quietly placing bets on AI × Web3 because of Sahara.

Capital will only pay for a direction and execution that have certainty - this is positive feedback on the depth of Sahara technology, team background, system design, and execution capability.

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

More than 3.2 million accounts have been activated on the testnet, with over 200,000 data platform annotators (millions in queue). Their clients include some leading technology companies, and they have already achieved revenue in the tens of millions of dollars.

On this infrastructure on-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 narrative projects".

The ultimate challenge of public chains: to ensure that all contributors benefit continuously and drive a positive economic cycle.

Returning to our initial judgment logic: In a system where AI and blockchain are combined, is there really a mechanism that allows every contributor to be seen, recorded, and continuously rewarded?

Model training and data optimization rely heavily on a large amount of labeled data and interactive support; conversely, if there is a lack of user contributions, the project itself has to invest more funds to procure data and outsource labeling, which not only increases costs but also diminishes the value-driven nature of community co-construction.

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

Through the mechanism of Sahara, it not only improves the quality of the model but also allows more people to understand and participate in this decentralized AI ecosystem, linking data contribution with benefits, thus forming a true virtuous cycle.

A typical example is a voice project on a public chain, which quickly built a high-quality dataset covering multiple languages and accents through 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 labeling 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 it said that this is a scenario supported by long-term value?

Imagine if you want to build an AI application, you naturally hope that 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 models 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 likely to become 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 is possible to achieve a long-term, traceable, and sustainable incentive system.

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

But this is not just a revenue model for the data annotation and model training stages. Sahara builds an economic system that covers the entire lifecycle of AI models, with built-in profit-sharing mechanisms at every stage after the model goes live, including calls, combinations, and on-chain reuse, 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 one-time transaction or buyout.

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

Because of this, I agree with Sahara's underlying belief: a truly healthy AI economic system cannot simply be the plunder of data and the buyout of models; it cannot just allow a few people to reap all the benefits. It must be open, collaborative, and win-win—where everyone can participate, every valuable contribution can be recorded, and rewards can continue to be gained in the future.

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

Although I am optimistic about Sahara, I will not obscure the challenges the project will face due to my investment position.

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

Its system was designed from the beginning to be open, on-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 call Sahara's model services and complete on-chain settlements.

However, despite the extreme scarcity of such 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 composable AI protocol layer, the key for Sahara lies in how ecosystem participants assess its technological maturity, stability, and future predictability. Although the system itself 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 Sah