Powell’s chair term ends May 15, and the Senate votes on the crypto clarity bill May 14. This is the most consequential week for US crypto regulation in years. While traders watch the macro clock, the industry is moving fast. Circle just launched a new blockchain. Solana is testing its biggest upgrade ever. Strategy resumed buying. And AI is now doing the heavy lifting on both fraud prevention and trading strategy. A lot is happening below the price chart. Let’s dive in.
Circle Just Launched Its Own Institutional Blockchain
Circle raised $222 million for Arc, a new institutional layer-1 blockchain using USDC as its native gas token. Backers include BlackRock, a16z, Apollo, and Standard Chartered. Arc features sub-second finality, opt-in privacy, and quantum resistance at launch. CEO Jeremy Allaire says every company in the world will eventually be tokenized. What percentage of Arc’s token supply is allocated to network participants? Read more.
Strategy Just Resumed Bitcoin Buying After a Brief Pause
Strategy acquired 535 BTC for $43 million on May 11, resuming purchases after a one-week pause for its Q1 earnings call. Total holdings now stand at 818,869 BTC. Saylor also clarified his famous stance, saying the precise version is “never be a net seller of Bitcoin.” What is Strategy’s break-even issuance rate to remain a net BTC buyer? Read more.
Solana’s Biggest Upgrade Ever Just Entered Real Testing
Solana’s Alpenglow consensus overhaul is now live on a community test cluster for the first time, with external validators running the new mechanism. Internal tests showed finality times dropping roughly 100 times after the switch. A mainnet rollout could follow as early as late Q3 or early Q4 2026. What two existing systems does Alpenglow replace on Solana? Read more.
Binance’s AI Blocked Billions in Fraud Over Fifteen Months
Binance said its AI security infrastructure blocked more than $10.5 billion in potential user losses between January 2025 and the end of Q1 2026. AI-driven tools now account for 57% of the exchange’s fraud controls. Smart contract exploits now cost attackers as little as $1.22 per contract, down 22% month-over-month. How much did Binance’s AI systems protect in Q1 2026 alone? Read more.
MoonPay Just Built an AI That Trades for You MoonPay acquired Dawn Labs and launched Dawn CLI, a tool that converts plain-English instructions into automated trading strategies without any coding required. It launches initially on Polymarket. Users can review strategy code before deployment, with controls capping how much an agent can trade and which markets it can access. What guardrails does Dawn CLI have to prevent unintended trades? Read more.
How do AI trading agents decide when to buy and sell crypto?
TLDR: AI trading agents buy and sell crypto when their models predict a favorable risk adjusted return that is large enough to beat fees, slippage, and risk limits.
- They turn raw data (price, volume, order books, on chain, news, sentiment) into features, then use trained models to estimate short term future returns and probabilities.
- They only trade when signals clear thresholds after accounting for risk, liquidity, position limits, and transaction costs, often ignoring weak or noisy signals.
- Execution logic then decides how to trade (order type, venue, pacing), and risk management rules such as stops, sizing, and max exposure often matter as much as the prediction itself.
Disclaimer: This alpha is provided by CryptoMx. CryptoMx can make mistakes—please DYOR. Not financial advice.
