AI × Web3: Who will build the chain for this era?

When a true shift in the technological paradigm occurs, we often see a surge first, rather than the system. The wave of AI that we are experiencing is no different.

As a primary investor, I have always believed that betting on the transformative forces deep within the industry is far more worthwhile than chasing superficial narratives.

Over the past year, I have seen a large number of projects such as RWA, Consumer, infoFi, etc. - they are undoubtedly exploring the intersection of the real world and on-chain systems.

But an increasingly obvious trend is that regardless of which path a project takes, it ultimately has to enter the collaborative logic of AI, utilizing AI to enhance competitiveness and efficiency.

For example, RWA, thinking about how to optimize risk control with AI, off-chain data verification, and dynamic pricing is the future;

Whether it is consumers urgently needing excellent user experience or DeFi, AI is also needed to accomplish user behavior prediction, strategy generation, incentive distribution, and there are other directions in this type of track that I won’t elaborate on.

Therefore, whether it is asset digitization or experience optimization, these seemingly independent narratives will ultimately converge into the same technological logic: if the infrastructure does not have the ability to integrate and carry 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 "stronger" and being "used more and more"; the real paradigm shift lies in the reconstruction of collaboration logic.

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

AI is also following this path: Agents will become everyone's intelligent co-creation entity, helping you turn expertise, creativity, and tasks into automated productivity tools, and even monetization.

This is a question that is difficult to answer in today's Web2 world, and it reflects some underlying logic I see in the AI + Web3 space: making AI collaborative, transferable, and revenue-sharing is the truly worthwhile system to build.

What I want to talk about today is the only project so far that attempts to systematically build the underlying operation of AI from the chain-level structure: Sahara.

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

My investment logic is not just the narrative of public chains combined with AI, and then seeing which team seems to have a better background before placing a bet.

Investment, in essence, is 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 leverage blockchain to reconstruct the value attribution and distribution logic of AI, allowing ordinary users, developers, and other roles to have the opportunity to participate, contribute, and continuously benefit? It's simple, only with the emergence of this logic do I believe such projects could become disruptors, rather than just "abandoned public chain +1".

To find the answer, I basically scanned all the AI projects I could access until I came across 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 their token economies is mostly limited to Gas Fees or governance, rarely able to truly support the positive cycle of the ecosystem, and even more difficult to 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 - and the reason why I am firmly investing.

As I emphasized in my previous article discussing the "Evolution from Web2 to Web3": the real paradigm shift lies not in creating a single product, but in building a supporting system.

And Sahara is one of the most anticipated cases in my assessment at that time.

From Investment to 8x Valuation Follow-up Heavy Investment

If I initially invested in Sahara, it was because it was doing exactly what I believe to be the true mission of AI - to build the AI economy and infrastructure system. What drove me to invest again at an 8x pre-round valuation in just six months is the rare strength I felt in this team.

Two co-founders, one is the youngest tenured professor at USC, specializing in AI. I believe that the value of a tenured professor in their 90s is not only in the academic field but also in the fact that at this age, they still have dreams, energy, and the courage to realize those dreams. Having known Professor Ren for over a year, I have witnessed what it means to work for more than ten hours a day, with stable emotions and humility.

Tyler, the former director of investments at a trading platform's lab, is responsible for North American investments and incubators, and his understanding of web3 goes without saying. He is astonishingly disciplined: he only sleeps in whole multiples of 1.5 hours, maintains his fitness regardless of how busy he is to keep in shape, and doesn't touch a single piece of candy to keep his mind clear, working over 13 hours a day. I used to joke that he was a robot, and he simply replied, "I’m lucky to be busy like this today." His source of dopamine comes from advancing projects every day; building dreams is his passion, and he doesn't need any other fuel.

I am very glad to have met them, which changed me. I have finally started to sleep regularly as much as possible, my emotions are gradually stabilizing, and I am working 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 deeply remember that this round of primary market financing was difficult, but Sahara was being chased by investments from the primary market.

What everyone remembers is that a certain investment institution and a certain trading platform invested in Sahara. Sahara has ushered in an era of investment for Samsung's entry into the Web3 AI field, and its winning of the Samsung AI Award is an important reason for Samsung's investment. In addition, some heavily invested AI funds, national banks, and so on are also special 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 place bets on AI × Web3 because of Sahara.

Capital will only pay for certainty in direction and execution - this is the 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 test network, with over 200,000 data platform annotators (millions more in queue). Their clients include several leading enterprises, and they have already achieved 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 steadier than 99% of the "AI Narrative projects".

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 allows every contributor to be seen, recorded, and continuously rewarded?

Model training and data optimization rely heavily on a large amount of annotated and interactive support; conversely, if there is a lack of user contribution, the project itself must invest more funds to procure data and outsource annotation, which not only increases costs but also diminishes the value-driven aspect 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 labeling task system operates daily, with a large number of community users actively participating in labeling and prompt creation. Not only does it help improve the system, but it is also investing in the future with data.

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

A typical example is the Myshell 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 has also propelled its open source projects like VoiceClone and MeloTTS to receive thousands of GitHub stars and over 2 million downloads on Hugging Face.

At the same time, users participating in data annotation also received token rewards issued by Myshell, creating 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 protecting 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 soon 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; these contributors are likely to become the actual 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 the 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 invocation, combination, and cross-chain reuse, a profit-sharing mechanism is built in, allowing value to be captured over a longer cycle.

Model developers, optimizers, validators, and computing power contribution nodes can continuously benefit at different stages, rather than just relying on one-time transactions or buyouts.

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 underlying belief: a truly healthy AI economic system cannot simply be about the plunder of data or the acquisition of models, nor can it be about allowing a few individuals to monopolize the benefits. It must be open, collaborative, and win-win - where everyone can participate, and every valuable contribution can be recorded and continue to receive returns in the future.

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

While I am optimistic about Sahara, I will not hide the challenges the project will face due to my investment stance.

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

Its system has been 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 call Sahara's model services and complete on-chain settlements.

However, despite the extreme scarcity of this architectural design, it also carries a core risk: the value of the infrastructure does not lie 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 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: serving multiple leading enterprises, providing them with relevant data services, and addressing some of the industry's most challenging data demand issues, thus becoming a verification of the feasibility of this system.