AI × Web3: Who is building the infrastructure of the new era?

The true transformation of technological paradigms often begins with a surge, rather than a system. The current wave of AI is exactly that.

As a primary investor, I firmly believe that betting on the deep transformative forces of the industry is more valuable than chasing superficial narratives.

In the past year, I have been involved in numerous projects exploring the integration of the real world with on-chain systems. However, an increasingly obvious trend is that regardless of the project roadmap, it ultimately needs to incorporate AI collaboration logic to enhance competitiveness and efficiency.

For example, RWA projects need to consider how to leverage AI for risk control optimization, off-chain data validation, and dynamic pricing; Consumer or DeFi projects require AI to achieve user behavior prediction, strategy generation, and incentive distribution.

Therefore, whether it is asset digitization or experience optimization, these seemingly independent narratives will ultimately converge on the same technological logic: the infrastructure must have AI integration and carrying capacity to support 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 widely"; the real paradigm shift lies in the reconstruction of collaborative logic.

Just like the early transformation of the internet, it was not only due to the invention of DNS or browsers, but because it allowed everyone to participate in content creation and turn ideas into products, thus giving rise to an entire open ecosystem.

AI is also on this path: Agents will become intelligent co-creators for everyone, helping to transform expertise, creativity, and tasks into automated productivity tools, and even realize monetization.

This is a question that is difficult for the current Web2 world to answer, and it is also the underlying logic behind my focus on 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 a chain-level structure: Sahara.

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

My investment logic is not simply to combine the narrative of public chains with AI and then choose well-established teams to bet on.

Investment is essentially a choice of worldview, and I have always been asking a core question: Can the future of AI be jointly owned by more people?

Can it reconstruct the value attribution and distribution logic of AI through blockchain, allowing ordinary users, developers, and other roles to participate, contribute, and continuously benefit? Only when this logic emerges do I believe that such projects have the potential to become disruptors, rather than "just another public chain."

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

This sentence seems simple, yet 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 fuel fees or governance, making it difficult to truly support a positive cycle of the ecosystem, and even harder to sustain the development of emerging tracks.

I am well aware that this path is full of challenges, but precisely because of this, it is a revolution not to be missed—this is also the reason for my firm investment.

As I emphasized before in my discussion on the "evolution from Web2 to Web3": the real paradigm shift is not about creating a single product, but about building a supportive system.

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

From Investment to 8x Valuation Follow-up Heavy Investment

If I initially invested in Sahara because it is realizing the true mission of AI as a leader in my heart – to build an AI economy and infrastructure system, then what made me eagerly chase after a follow-up investment at an 8x valuation in just six months is the extremely rare strength I felt from this team.

The two co-founders each have their own characteristics: one is the youngest tenured professor at the University of Southern California, specializing in the AI field. The value of a tenured professor from a top American university born in the 90s is not only reflected in the academic field but also in the fact that at this age, they still have dreams, energy, and the determination to achieve those dreams. 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, maintain emotional stability, and remain humble.

Another co-founder, Tyler, previously served as the North America Investment Director at a well-known institution, responsible for investment and incubator operations, and his understanding of Web3 goes without saying. His level of self-discipline is astonishing: he only sleeps in multiples of 1.5 hours, insists on working out to stay in shape regardless of how busy he is, and even abstains from alcohol to keep his mind clear, working more than 13 hours a day. I jokingly said he is a robot, and he simply responded: "I am lucky to have this busyness today." He derives dopamine from advancing project progress; dreaming is his passion, requiring no other fuel.

After getting to know them, I also started to establish a regular routine as much as possible, my emotions gradually stabilized, and I began to work out...

So when someone says that Sahara received capital favor because of luck, I always unreservedly add, "the capital's pursuit is an inevitable result." I deeply remember the difficulty of primary financing in this round of the market, yet Sahara was being chased for investment.

Well-known investors include a certain famous fund, a certain trading platform, and a certain venture capital institution. Sahara has opened the investment era for a certain tech giant entering the Web3 AI field, and its receipt of the company's AI award is an important reason for the investment. In addition, some funds heavily invested in AI, national banks, and others are also honored guests of Sahara. You can see a group of institutions that are more focused on traditional technology and industrial resources starting to quietly lay out AI × Web3 because of Sahara.

Capital will only pay for a certain direction and execution ability—this is a positive feedback on the depth of Sahara technology, team background, system design, and execution capability.

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

More than 3.2 million accounts have been activated on the test network, with over 200,000 data platform annotators (millions queued), serving clients that include several tech giants and well-known enterprises, and achieving revenue in the tens of millions of dollars.

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

The Ultimate Challenge of Public Chains: Allowing All Contributors to Continuously Benefit and Drive Positive Economic Circulation

Returning to our initial judgment logic: in a system that combines AI and blockchain, 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 labeling 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 annotations 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 enhances the quality of the model but also allows more people to understand and participate in this decentralized AI ecosystem, linking data contributions with earnings to form a true positive cycle.

A typical example is a voice project on a public blockchain that 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 stars on GitHub and over 2 million downloads on a certain AI model platform.

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

Sahara's "permissionless copyright" mechanism ensures the rights of all participants while guaranteeing the open circulation and reuse of AI assets—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 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 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; these contributors are 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 can realize a long-term, traceable, and sustainable incentive system.

No matter 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, where each link after the model is launched—including invocation, combination, and cross-chain reuse—also has a built-in profit-sharing mechanism, allowing value to be captured over a longer period.

Model developers, optimizers, validators, and computing power contribution nodes 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, with each call generating new revenue for the original contributor.

For this reason, I agree with Sahara's fundamental belief: a truly healthy AI economic system cannot simply be the plundering of data or the buying out of models, 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 rewards can continue to be received 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 cover up the challenges that 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 a single chain or ecosystem.

Its system was designed from the beginning to be open, full-chain, and standardized: it supports 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, it also carries 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 ecological participants assess its technological maturity, stability, and future predictability. Although the system itself has been built, whether it can genuinely attract a large number of projects to land based on its standards remains uncertain.

It is undeniable that Sahara has achieved key validation: serving several tech giants and well-known enterprises by providing them with relevant data services and addressing some of the industry's most challenging data demand issues, becoming an early signal of the feasibility of this system.

However, it is important to note that these collaborations mainly come from the Web2 world. The long-term development of Sahara is still determined by the maturity and penetration of the entire Web3 AI sector. Sahara benefits from the overarching trend of Web3 AI, but to truly unlock the value of its infrastructure, it still requires more native Web3 AI products and technological solutions to be implemented and perfected.

But no