Author:kyle, Wu talks about blockchain

Bianhui cut in half

Let’s talk about the most direct data first: the number of side events this year has been significantly reduced by half. According to statistics from Luma, compared with the 668 side events in 2025, the number of side events during this year’s conference (2026/2/18–2026/2/21) is only about 215, a drop of 68%.

There are several reasons behind this. On the one hand, it is the industry cycle itself; on the other hand, the conference time is close to the Lunar New Year, which is not friendly to Asian project parties and participants; in addition, some external activities are also diverting industry attention, such as Mar-a-Lago WLFI Forum, which also sucked away some OGs and core practitioners.

So, participants can still feel that there are many North American Web3 people gathering together at ETH Denver, but there are obviously far fewer people coming from outside the United States, and the spillover influence of the entire conference is weak.

Judging from the specific ecological activity, we can also feel that the industry has entered a stage that is more efficient and cost-saving.

Monad and X Layer have a relatively strong presence. The former hosted three events, and the latter sponsored the main stage (but the team only participated in the first day of the exhibition); Aptos and Sui also had some exposure; Solana only hosted one event, which was small in scale but of good quality. On the whole, it is obvious that many public chain ecology no longer pursues the past multi-store activities and full presence, but only retains a small number of Side Events, holding one or two activities for fun, and pays attention to cost-effectiveness.

AI is everywhere

One of the biggest changes this time is that AI is almost everywhere.

In the main venue, there is also the Claws Out Summit for the recent "lobster" OpenClaw, and there are also small roadshows for AI project founders; outside the venue, like the AI-themed Open AGI Summit hosted by Sentient, the scale and crowd density are even higher than some main venue areas. The temperament of this ETH Denver is no longer like a traditional encryption industry conference, but more like an AI × Crypto mixed exhibition.

Even the form of the exhibitors has changed. When you see robots and robotic arms while visiting the exhibition, you will have the illusion of "I have just walked to CES." Some projects focusing on embodied intelligence, such as PrismaX and Gensyn, have independent booths. In addition, although many projects still carry the Web3 label, their core narratives are no longer simply centered around public chains, DeFi, and wallet infrastructure in the past, but are moving in the direction of agents, chatbots, etc., that are closer to the application layer of AI products.

An exchange strategy manager said that when he attended this conference, he was not focused on very mature projects, but on very early or even embryonic directions. Regarding AI + Crypto, what he focuses on is not whether the project will generate income immediately, but whether new production relationships, new distribution methods, new workflows and new on-chain collaboration mechanisms will emerge after the combination of AI and Crypto.

He also shared with me an internal strategic judgment made by their company's senior management: Everyone has already agreed that AI will bring huge productivity improvements, but the real investment in the Web3 field is not the so-called large model, but the focus is on whether AI technology can be embedded within the exchange to improve the current product experience. They are currently developing a large language model embedded in the exchange, giving trading recommendations based on real-time market news, and conducting transactions on related products directly in the chat box.

ETH Denver is still Builder-oriented

Although AI has become the hottest topic, one thing that has not changed at ETH Denver is that it is still one of the most builder-oriented Crypto conferences.

This can be seen from the agenda. In addition to the closing ceremony on the last day, all the exhibition time was given to the Hackathon and Builder Workshop. Side events during public chain exhibitions such as Base were also organized around developers.

It is worth mentioning that Base also demonstrated a small product called Braindate during the conference. Base has dedicated an area in its exhibition area for everyone to socialize and chat. But unlike traditional chat areas, you can add some games you find interesting on Braindate, or create a game to attract like-minded friends to chat.

This temperament can be more clearly felt in the venue: instead of discussing the market, more people coming to ETH Denver are still chatting about "what are you building recently?" Especially student builders, they care more about "whether this thing can be made" and "whether this product is interesting" rather than short-term market fluctuations. Because of this, although the overall scale of ETH Denver has shrunk, it has not completely turned into a retreat conference that only talks about narratives and not products. At least here, the real builder is still at the core.

The hackathon BUIDLathon that best embodies this temperament is still its hackathon BUIDLathon, and this BUIDLathon has an obvious change: the overall competition system design is more advanced.

Based on the feedback from the contestants, BUIDLathon has joined the Online Hacking online development stage. The specific topic selection will be informed to everyone one week in advance. It is no longer the traditional rhythm of "people arrive at the scene and then start working".

This is actually a very interesting innovation. It is more in line with the current working methods of many builders, and it is also easier to improve project completion. But at the same time, it also means that the time for in-person communication is shorter, which has been reduced from 8 days last year to 4 days. A lot of preparation work must be completed in advance. Offline is no longer the starting point of the build, but more like the final centralized finishing of the project.

Changes in the competition system may only reflect changes in the organizer's organizational preferences, but changes in the prize pool cruelly demonstrate the cycle of the industry.

After interviewing many contestants, they all complained that the number and intensity of sponsors this year were far behind compared to previous years. The total prize pool last year was a whopping $1.03 million, but this year it's down to $132,000. Sponsor’s budget is more focused and geared towards what they care about most: AI.

Some contestants mentioned that the gimmick of a certain championship project in the hackathon was an "AI girlfriend." Users can use cryptocurrency to reward their AI girlfriends, and through more frequent reward interactions, a certain behavioral incentive mechanism can be established to bring the "two people" closer together. This reward mechanism sounds like a joke, but the award of this project also speaks volumes: the competition judges are no longer limited to looking at the combination of AI + Crypto from strong Web3 scenarios such as wallets and settlements. They also hope to see the project outline usage scenarios that are closer to ordinary and mass users.

I also interviewed Justin, a contestant who participated in the hackathon and won the third place in the Base sponsorship bonus. The project he is working on is an open protocol for AI agent advertising monetization.

The judgment behind this project is that many AI products that will be free and open to the public in the future may need to achieve sustainable monetization through advertising rather than subscriptions. Advertisers pay for promotion tasks, and the AI ​​agent distributes recommendations based on user context, drives conversions, and earns revenue after the tasks are completed.

But the biggest difficulty is: how to prove that AI has really completed the advertising task.

Justin’s plan is to introduce an on-chain validator to verify the advertising results. In this hackathon demo, he mainly verified verifiable behaviors such as on-chain purchases: only when the relevant delivery results are confirmed as real by the validator, the agreement will trigger settlement, and the advertiser's budget will actually be paid to the agent that completed the task. In this sense, this project is not just an "AI-received advertising" product, but an attempt to build a verification and settlement infrastructure for the future AI-native advertising market.

There is another point worth mentioning about this hackathon: the participating groups are very diverse.

There were both college students and senior builders on site; there were people with deep experience and mature understanding of the industry, as well as creators with new and interesting practices. In addition to AI and DeFi related directions, there are also some interesting GameFi projects among the participating projects.

ETH Denver’s hackathon did not turn into a game only for crypto circles. It can still be felt that many new people are coming in, and they do not necessarily carry the burden of doing projects related to traditional crypto usage scenarios. Instead, they are more willing to put AI, games, advertising, social networking and other scenarios together naturally with on-chain technology.

Slightly revealing the review process of Caotai

Of course, this hackathon is far from perfect, and the review process seems a bit "rough and shabby".

Contestant feedback, the review process is divided into main track and sponsor track. The main track is relatively clear here: you start queuing up at 9 a.m., usually facing 2 to 3 judges, and the whole pitch and presentation process is not particularly stressful. But the problem is that the presentation time is only five minutes, which means that the review is naturally more biased towards teams that can clearly explain and demonstrate the project in a very short time. Under time constraints, having strong enough laughs and memorable points is often more important than being technically polished, and even bugs will not affect winning the award.

After finishing the main track, the team will continue to the sponsor track and find sponsors such as Base to talk about the project again. This stage is relatively chaotic. The first-come, first-served order management is not clear enough, and it also tests the team's on-the-spot judgment and execution.

Prediction Market

In addition to AI, another topic in this conference that I think is worth mentioning separately is the prediction market.

I heard several discussion points shared by the guests at Frontier Markets hosted by Monad, and I was deeply impressed. First, they believe that the biggest barrier to the prediction market is still liquidity. Secondly, since each market has an end time, liquidity will continue to migrate and it is difficult to settle in the long term like a perpetual contract. Furthermore, how to attract liquidity providers in the long-tail market is still a difficult problem to solve.

At the same time, there are obvious differences between prediction markets and traditional trading. Traditional trading is more about gradual fluctuations, but the prediction market may often quickly return to zero at a certain point in time, which makes leverage design, market-making strategies and risk management more complicated, and also makes many traditional large-scale market makers hesitant to enter the market.

However, the prediction market also has its own particularities: many popular markets will naturally have liquidity providers with retail investors as the main body. In other words, the key to this track is not necessarily just “making another prediction platform”, but who can continue to create markets that are attractive enough for retail investors to participate, and reorganize liquidity around these markets and improve user experience.

The resilience of encryption

If I were to give a conclusion for this ETH Denver, I would say: it is the appearance of all beings when the industry is experiencing a bear market and is in a transitional stage. This is not a stage where hot topics are clear and everyone quickly follows the trend around popular narratives. There is less excitement, less enthusiasm in the currency circle, and less so-called "beauty during economic upswings." But at the same time, there are still a group of builders on the scene who are seriously exploring, and some investors are still looking at projects with great potential in very early stages. Although some projects are not perfect and still have bugs, they have begun to reveal feasible business models and outlines of the next stage of the currency circle.