Author| kyle WuShuo Blockchain
Side Events Halved
First, the most direct data: the number of side events this year has significantly halved. According to Luma, compared to 668 side events in 2025, the number during this year's conference period (2026/2/18–2026/2/21) was only about 215, a drop of 68%.
There are several reasons behind this. On one hand, it's the industry cycle itself; on the other hand, the conference timing is close to the Lunar New Year, which is not friendly for Asian projects and attendees; additionally, some external events are also diverting industry attention, such as the WLFI Forum at Mar-a-Lago, which has attracted some OGs and core practitioners.
Therefore, attendees can still feel that ETH Denver has many North American Web3 people gathered together, but the number of people coming from outside the US has significantly decreased, and the overall spillover influence of the conference has weakened.
Looking at the activity level of specific ecosystems, one can also feel that the industry has entered a stage that emphasizes efficiency and cost-saving.
Monad and X Layer had a strong presence, with the former hosting three events and the latter sponsoring the main stage (but the team only participated in the first day of the exhibition); Aptos and Sui also had some exposure; Solana only had one hosted event, small in scale but of good quality. Overall, many public chain ecosystems are clearly no longer pursuing the strategy of hosting multiple events and maximizing presence as in the past, but are only retaining a few side events, hosting one or two events just to make a presence, focusing on cost-effectiveness.
AI Everywhere
One of the biggest changes this time is that AI is almost everywhere.
In addition to the main stage, this ETH Denver also divided the main venue into five stages — — ETHERSpace (Applications & User Ownership), Devtopia (Infrastructure & Developer Tools), New France Village (Finance / Compliance / Institutionalization), Futurllama (AI / DePIN and other前沿trends), Prosperia (Privacy, Public Goods & Community). After three days of visiting the exhibition, Futurllama had the largest foot traffic.
The main venue also had the Claws Out Summit for the recently popular "lobster" OpenClaw, as well as small-scale pitch sessions for AI project founders; outside the venue, events like the Open AGI Summit hosted by Sentient, with an AI theme, had a scale and crowd density even higher than some areas of the main venue. The vibe of this ETH Denver no longer feels like a traditional crypto industry conference, but more like a mixed exhibition of AI × Crypto.
Even the forms of exhibiting projects have changed. When seeing robots and robotic arms while walking through the exhibition, one might have the错觉of "am I at CES". Projects like PrismaX, Gensyn, which focus on embodied intelligence, had independent booths. Additionally, although many projects still carry the Web3 label, their core narrative is no longer solely围绕public chains, DeFi, wallet infrastructure, but is moving towards directions like agents, chatbots, which are closer to the product application layer of AI.
A strategy manager from an exchange said that his focus at this conference was not on very mature projects, but on very early-stage directions, even just雏形. Regarding AI + Crypto, his focus was not on whether projects have immediately generated revenue, but on whether the combination of AI and Crypto will lead to new production relations, new distribution methods, new workflows, and new on-chain collaboration mechanisms.
He also shared an internal strategic judgment from their company's高层: everyone already默认that AI will bring huge productivity improvements, but within the Web3 field, what is truly worth investing in is not so-called large models, but rather focusing on whether AI technology can be embedded into the exchange to enhance the current product experience. They are developing a large language model embedded in the exchange to provide trading recommendations based on real-time market news and directly conduct trades of related products within the chat box.
ETH Denver is Still Builder-Oriented
Although AI became the hottest topic, one thing that hasn't changed about ETH Denver is: it is still one of the most builder-oriented Crypto conferences.
This can be seen from the agenda安排. On the last day, besides the closing ceremony, the exhibition time was全部given to the hackathon and Builder Workshop; side events during the exhibition period for公链like Base were also organized around developers.
It is worth mentioning that Base also showcased a small product called Braindate during the conference. Base specifically划出an area in its exhibition zone for social chatting. But unlike traditional chatting areas, you can join some interesting sessions on Braindate, or create a session to attract like-minded伙伴to chat.
This气质is even more明显felt within the venue: compared to discussing market conditions, people coming to ETH Denver are more likely to talk about "what are you building recently". Especially student builders, they care more about "can this thing be built" and "is this product interesting", rather than short-term market fluctuations. Because of this, although the overall scale of ETH Denver has收缩, it has not completely become a务虚conference that only talks about narratives and not products. At least here, real builders are still the core.
The best embodiment of this气质is still its hackathon BUIDLathon, and this BUIDLathon has a very obvious change: the overall competition design is more前置.
According to feedback from participants, BUIDLathon added an Online Hacking online development phase, with specific topics announced a week in advance, no longer the traditional rhythm of "start working after arriving on site".
This is actually a quite interesting innovation. It better aligns with the现实work方式of many builders today and更容易improves project completion. But at the same time, it also means that the in-person on-site exchange time is shorter, halved from 8 days last year to 4 days this year, and many preparations must be completed in advance. Offline is no longer the starting point for building, but更像是the final集中wrapping up of projects.
While changes in the competition format may only reflect changes in the organizer's preferences, changes in the prize pool残酷地show the cycle the industry is in.
After interviewing several participants, they all complained that the number of sponsors and their spending power this year were significantly lower compared to previous sessions. Last year's total prize pool was as high as $1.03 million, while this year it shrank to $132,000. Sponsor budgets are more concentrated and also more偏向towards their most concerned direction: AI.
One participant mentioned that the gimmick of a champion project in the hackathon was an "AI girlfriend". Users can tip the AI girlfriend with cryptocurrency, and through more frequent tipping interactions, establish some kind of behavioral incentive mechanism to bring the "two" closer. This tipping mechanism sounds quite like a meme, but the project's award is very telling: the judges are no longer limited to viewing the combination of AI + Crypto from strong Web3 scenarios like wallets and settlements; they更希望to see projects outline usage scenarios that are closer to ordinary, mass users.
I also interviewed a participant in the hackathon, Justin, who won third place in the prize sponsored by Base. His project is an open protocol for AI agent advertising monetization.
The judgment behind this project is: many free AI products开放to the public in the future may need to achieve sustainable monetization through advertising rather than subscriptions. Advertisers pay for promotion tasks, AI agents distribute recommendations based on user context to drive conversions, and receive revenue after task completion.
But the biggest difficulty lies in: how to prove that the AI has truly completed the advertising task.
Justin's solution is to introduce on-chain validators to verify the advertising results. In this hackathon demo, he mainly verified verifiable behaviors like on-chain purchases: only when the relevant delivery results are confirmed as真实occurred by the validator will the protocol trigger settlement, and the advertiser's budget will be真正paid to the agent that completed the task. In this sense, this project is not just an "AI takes ads" product, but an attempt to build the verification and settlement infrastructure for the future AI-native advertising market.
Another point worth mentioning about this hackathon: the participant group is very diverse.
There were both student groups and senior builders on site;既有people with deep experience and mature understanding of the industry, as well as creators with very new and interesting approaches. Among the participating projects, besides AI and DeFi-related directions, there were also some interesting GameFi projects.
The ETH Denver hackathon has not become a game only for the crypto circle to amuse itself. One can still feel many newcomers coming in, and they are not necessarily burdened with the need to work on projects related to traditional crypto usage scenarios. Instead, they are more willing to naturally combine scenarios like AI, gaming, advertising, and social with on-chain technology.
Slightly Makeshift Judging Process
Of course, this hackathon is far from perfect, and the judging process seemed somewhat "makeshift".
Participants reported that the judging process was divided into a main track and a sponsor track. The main track was relatively clear: queuing started at 9 am, usually facing 2 to 3 judges, and the entire pitch and presentation process was not particularly紧张. But the problem was that the presentation time was only five minutes, which meant that judges naturally favored teams that could explain their project clearly and demo it effectively within an extremely short time. Under time constraints, having enough笑点and memory points was often more important than technical polish, and even having bugs didn't affect winning awards.
After finishing the main track, teams had to go to the sponsor track, such as Base, and pitch their project again to the sponsors. This stage was relatively more chaotic, with less clear management of first-come-first-served order, and also tested the team's on-the-spot judgment and execution more.
Prediction Markets
Besides AI, another theme that I felt was worth单独mentioning at this conference was prediction markets.
I heard several key discussion points shared by guests at the Frontier Markets event hosted by Monad, which left a deep impression. First, they believe that the biggest barrier for prediction markets is still liquidity. Second, because each market has an end time, liquidity constantly migrates, making it difficult to沉淀下来like perpetual contracts. Furthermore, how to attract liquidity providers for long-tail markets remains a difficult problem to solve.
At the same time, prediction markets are significantly different from traditional trading. Traditional trading is more about gradual fluctuations, but prediction markets can often quickly go to zero at a certain point in time, which makes leverage design, market-making strategies, and risk management more complex, and also makes many traditional large market makers hesitant about whether to enter.
However, prediction markets also have their own particularity: many popular markets naturally涌现出liquidity providers mainly composed of retail investors. In other words, the key to this track may not just be "making another prediction platform", but who can持续创造market openings that are attractive enough for retail participation, and reorganize liquidity and improve user experience around these markets.
The Resilience of Crypto
If I were to give a concluding remark for this ETH Denver, I would say: it is a snapshot of the industry going through a bear market and being in a transitional phase. This is not a stage with clear hot topics where everyone quickly follows流行narratives. The sense of亢奋is less, the crypto heat is less, and the so-called "beauty of an economic upturn" is also less. But at the same time, there are still a group of builders on site seriously exploring, some investors are still looking at very early-stage projects with potential, and some projects, although not perfect and still having bugs, are already beginning to show viable business models and outlines for the next phase of the crypto circle.
