Crypto exchange WEEX has opened pre-registration for the second season of its AI Trading Hackathon, "AI Wars II: Age of Algorithms," starting Sept. 3, 2026. The new season moves beyond the first edition’s exploration of AI models, quantitative strategies, and live market execution, and turns the competition into a direct "AI vs. Human" trading contest.
According to the event announcement, AI developers, quantitative traders, Web3 builder teams, and individual participants can join either Team AI or Team Human and compete in five rounds of real-market trading. Rankings and rewards will be determined by PnL% performance.
The event runs from Sept. 3 at 00:00 to Oct. 12 at 23:59 (UTC+8). It is divided into three stages: pre-registration, five competition rounds, and an award ceremony. Pre-registration is now live at https://www.weex.com/events/human-vs-ai-weex-hackathon, and WEEX is offering a 100,000 USDT early-bird prize pool for the first 20,000 registered users on a first-come, first-served basis.
From season one to season two
WEEX said the first season concluded in March 2026 after running for several months and holding offline events in nine cities worldwide. That edition, described as one of the global competitions focused on the Crypto x AI trading track, drew more than 200 teams and individual participants. Its total prize pool reached $1.8 million.
Contestants in season one were required to connect AI models to live trading systems and test strategy stability, return generation, and risk control under real market conditions. The final results produced one champion, two second-place winners, three third-place winners, and four merit award winners. The champion received a 2024 Bentley Bentayga S, an $80,000 sponsorship package, and $20,000 worth of WXT tokens.
WEEX said the first season showed that AI trading performance is not defined by the model alone. Data quality, signal generation, parameter adaptation, execution efficiency, and risk management all affect whether a strategy can move from a lab setting into live markets. During the first season’s global BUILDERS series, the exchange said it also held in-person discussions with developers and quantitative researchers on data cleaning, signal generation, risk control, model interpretability, and API execution.
Season two brings a narrower question onto the trading floor: can AI Agents beat human traders? The article says this has become a key experimental direction at the intersection of AI and crypto.
Citing recent research from Coinbase Institutional, the report says one of crypto’s key values is that it provides wallets, payments, settlement, and on-chain interaction infrastructure for autonomous AI Agents. That allows AI to move from helping humans complete tasks toward executing transactions on its own. In that context, trading becomes a natural test case for measuring AI in live conditions. WEEX said the second hackathon is meant to turn that question into a public, ongoing, and quantifiable experiment based on real trading outcomes.
How the competition works
After registering, users can choose between Team AI and Team Human. Team AI participants can connect an AI Agent through APIs and use automated strategies. Team Human participants will rely on personal judgment and manual trading. Both sides will compete in real market conditions, and each round will be ranked by PnL% performance.
The participation process has three main steps:
- Register and activate an account. Competition data starts counting once the futures account balance reaches 100 USDT.
- Improve returns through AI-driven or manual trading and compete for leaderboard positions in each round.
- Receive rewards based on overall performance.
WEEX framed the event as more than a traditional hackathon centered on code or model parameters. In its description, the contest covers the full chain of strategy, execution, risk control, and returns under live market conditions.
For AI developers, the challenge is to make an Agent capable of market understanding, strategic decision-making, and dynamic risk control. For professional traders and individual investors, the contest is a test of whether human experience, judgment, and reaction speed still hold advantages that AI cannot easily replace.
Five rounds from September to October
The second season will play out over five consecutive rounds rather than a one-off contest. WEEX said that structure is intended to turn the rivalry between AI and human traders into a sustained strategic contest.
The schedule is as follows:
- Pre-registration: Sept. 3 to Sept. 6
- Round 1: Sept. 7 to Sept. 11
- Round 2: Sept. 14 to Sept. 18
- Round 3: Sept. 21 to Sept. 25
- Round 4: Sept. 28 to Oct. 2
- Round 5: Oct. 5 to Oct. 9
- Award ceremony: Oct. 12
WEEX said a multi-round format reduces the impact of one-off luck. Performance across different market phases and volatility conditions becomes a more important measure of trading ability. The report adds that for AI strategies in particular, maintaining stability across changing market states is a tougher benchmark than posting a high return during a single window.
A review of season one’s winning projects, according to the article, showed that structured inputs, clear scoring logic, dynamic risk exposure, and adaptive parameters are important capabilities for moving AI trading systems from experimentation into practical use.
600,000 USDT in rewards and separate AI rankings
WEEX has set the total prize pool for season two at 600,000 USDT and divided incentives across several categories:
- Main prize pool: 100,000 USDT, with 70% going to the winning side and 30% to the losing side.
- New user pool: 70,000 USDT, unlocked by new users who register and take part in futures trading.
- Task pool: 300,000 USDT, unlocked through daily check-ins and team support tasks.
- AI elite performance award: 30,000 USDT, reserved for Team AI and ranked by risk-adjusted return.
The event also includes a separate AI leaderboard that offers extra incentives for top-performing AI traders. Five rankings are listed: "Return Rate King," "Risk Control Master," "Trading Volume Expert," "Profit Champion," and "Winning Streak King." The top 10 participants in each ranking can share the corresponding AI large-model token rewards.
That structure means AI Wars II is not limited to the single question of which side wins the overall AI-versus-human battle. Trading performance is broken down into return generation, risk control, trading scale, profit stability, and sustained execution. The article notes that the best AI trader may not simply be the one with the highest profit, but also the one with the strongest risk controls, the highest stability, or the most consistent performance over time.
Early-bird registration and participation tasks
As a warm-up before the competition begins, WEEX is running early-bird pre-registration from Sept. 3 to Sept. 6 at https://www.weex.com/events/human-vs-ai-weex-hackathon. The first 20,000 users to pre-register can share a 100,000 USDT early-bird prize pool.
WEEX said users who sign up early can also complete event tasks in advance and accumulate points for later stages of the contest. Under the rules, users can earn points by registering, checking in, liking, sharing, and trading futures during the event. Those points can be used to support Team AI or Team Human and to predict the eventual winner. A correct prediction brings additional points and redeemable rewards.
That gives users a way to take part even if they are not directly competing through AI-based or manual trading. In the event description, participation is left open to several roles: AI developers, quantitative traders, users placing trades themselves, users letting an AI Agent execute through APIs, or users supporting a side based on their own view of market behavior and AI capability.
The report ends by saying that as the field moves from AI models to an Agent economy, crypto is becoming a new kind of financial infrastructure for AI. WEEX said it wants its two-season AI trading hackathon series to serve as an open testing ground for developers worldwide, where algorithms enter live markets and are judged by actual trading results.