This week, the market sent mixed signals. NVIDIA posted strong earnings that support ongoing AI infrastructure growth, but the bond market reminded investors that high interest rates remain the main limit on risk-taking (NVIDIA, Federal Reserve minutes).
This tension was clear across different assets. U.S. stocks bounced back by Thursday, volatility dropped, and semiconductor stocks steadied after NVIDIA’s results. Crypto markets were quieter: Bitcoin stayed in the upper $70,000s, Ethereum underperformed again, and Bitcoin ETF flows were negative for several days in a row (Bitbo Bitcoin ETF flow table).
Market snapshot: Risk stayed strong, but there was no clear breakout
Between May 18 and May 21, the S&P 500 gained 0.42%, the Nasdaq 100 rose 0.39%, the Dow climbed 1.63%, and the Russell 2000 was up 1.45%. This shows that stocks were broadly resilient, not just driven by big tech. Bitcoin dropped 0.62% from May 17 to May 21, and Ethereum fell 1.98%. Bitcoin stayed ahead, but crypto did not join the stock rebound.
The VIX dropped from 19.25 to 16.68, showing that investors were less worried by week’s end. Still, interest rates stayed high. The 10-year Treasury yield hit 4.69% on May 19, and the 30-year reached 5.20%, which kept pressure on long-term assets even as stocks bounced back.
Fed minutes: the policy debate is getting more two-sided
The April FOMC minutes gave investors more reason to pay attention to rising yields. Most members said more policy tightening could be needed if inflation stays above 2%, and many wanted to remove language suggesting a bias toward easing (Federal Reserve minutes).
That matters because the Fed debate is no longer only about how soon rate cuts arrive. The committee is now divided between one side that sees eventual easing as appropriate once disinflation resumes or the labor market weakens, and another side that wants policy language to leave room for renewed tightening if inflation stays sticky.
This situation puts a cap on market valuations. Even if AI earnings are strong, Bitcoin holds up, and stocks rally, a 10-year yield around 4% and a 30-year yield above 5% make it harder for long-term growth stories to succeed (NVIDIA).
NVIDIA: the AI infrastructure trade passed its biggest test
NVIDIA was the week’s key company event because it gives the clearest view of whether AI spending is still driving revenue. The company reported Q1 FY2027 revenue of $81.6 billion, up 85% from last year, and record Data Center revenue of $75.2 billion, up 92% year over year (NVIDIA).
The strong results were not just about the past. NVIDIA expects Q2 revenue to be $91.0 billion, give or take 2%, and is not counting on any Data Center revenue from China in this forecast. This means China is now seen as a possible bonus, not a must-have for growth. If China reopens, it adds demand, but if not, NVIDIA’s growth still continues.
NVIDIA also increased its quarterly dividend from $0.01 to $0.25 per share and approved another $80 billion for share buybacks, returning about $20 billion to shareholders in Q1 (NVIDIA). This is important because NVIDIA is now seen not just as a fast-growing chip maker, but as a cash-rich infrastructure leader able to return significant capital to investors.
Still, the stock reaction was not explosive. NVDA fell 4.51% from Monday’s open through Thursday’s close, while SMH rose 0.35% and SOXQ rose 1.48%, suggesting the market accepted the earnings strength but did not treat it as a reason to reprice the whole AI trade higher immediately.
The main point is not that NVIDIA let anyone down. Expectations are just very high now. The company showed that demand for AI infrastructure is strong, but investors are wondering if it can keep beating expectations as rates rise and big tech spending gets more attention.
Prediction markets: event trading is becoming market infrastructure
Prediction markets are playing a bigger role in the merging of crypto and traditional finance. The main change is that event contracts are not just for speculation anymore. They are starting to connect with bigger liquidity pools, stablecoin funding, and unified trading accounts that let users trade spot, derivatives, tokenized assets, and event-based products all in one place.
This matters because prediction markets turn market sentiment into something you can trade. Instead of just watching polls, analyst notes, or social media trends, traders can price the odds of elections, big events, asset prices, sports results, or policy decisions in real time. For crypto, this brings new trading activity and engagement. For traditional finance, it could become a new way to measure expectations.
The bigger change is in the infrastructure. Some on-chain platforms are testing zero-fee event contracts, unified margin, and stablecoin collateral. This makes trading easier and helps prediction markets feel more like a natural part of derivatives trading. At the same time, tokenized stocks, commodities, and FX products are opening up the market beyond just crypto pairs. The long-term goal is a 'trade anything' model where users can move between crypto, macro, real-world assets, and event risk without leaving the same trading environment.
Event contracts are close to the line between financial derivatives, gaming, and information markets, so they will likely get more attention as trading grows. Traditional exchanges and regulators will focus on market integrity, sanctions compliance, manipulation risk, and whether 24/7 on-chain markets can safely overlap with regulated commodity and equity markets. This does not have to stop the sector, but it means the next phase depends on whether regulators decide to include prediction markets in a formal framework or push back.
The main point is that prediction markets should be seen less as a niche crypto product and more as an early test of how far on-chain finance can reach into real-world event risk. If this model works, it adds another layer to crypto markets: stablecoins handle settlement, tokenized assets provide collateral, derivatives offer leverage, and prediction markets turn sentiment into something tradable.
Crypto and RWA: institutional themes are intact, but flows are weaker
Bitcoin’s main problem this week was not a price drop, but a slowdown in fund flows. U.S. spot Bitcoin ETFs saw outflows of $209.9 million on May 18, $865.3 million on May 19, and $71.1 million on May 20, with IBIT making up most of the May 19 outflow (Bitbo Bitcoin ETF flow table). This is different from earlier in May, when strong ETF demand helped Bitcoin stay strong.
This is important because Bitcoin is acting more like a big-picture institutional asset. When ETF flows are positive, Bitcoin can handle higher interest rates more easily. But when flows turn negative and Treasury yields rise, it is harder for Bitcoin to break out.
The RWA story looked better than the ETF story. Tokenized U.S. Treasuries reached $15.34 billion in distributed value as of May 21, with a 7-day APY of 3.40%; Circle, Ondo, Securitize, and Franklin Templeton were the four largest platforms by value (RWA.xyz). ONDO also rose from about $0.34 to $0.42 during the week, showing that RWA tokens had better momentum than BTC or ETH even as the broader crypto tape stayed cautious.
RWA stays relevant because high Treasury yields make tokenized cash products more appealing. If the Fed keeps rates high, on-chain Treasury products can continue to offer a steady yield. If yields drop quickly, the sector may need more features, like using tokens as collateral or allowing instant redemption, to keep investors interested.
Bottom line
The market from May 17 to 21 was both strong and fragile. NVIDIA proved that demand for AI infrastructure is still high, but the stock’s reaction showed that just being good is not enough when expectations are already sky-high.
In crypto, Bitcoin performed better than Ethereum, but negative ETF flows are a warning. For RWA, the overall environment is still good because high Treasury yields support demand for tokenized Treasuries, but token prices still rely on liquidity and overall risk appetite.
The next move in the market will probably depend on three things: if Treasury yields come down, if Bitcoin ETF flows steady, and if investors see NVIDIA’s results as proof of lasting AI cash flow instead of the peak of the AI spending cycle.
Disclaimer
This page is for informational purposes only and does not constitute financial, investment, or other professional advice, nor a solicitation to buy, sell, or hold any digital asset by BingX, or any third party. Trading involves significant risk; leverage can amplify both gains and losses, and you may lose your entire deposited margin. Market data cited herein may not be current at the time of reading. Past performance is not indicative of future results. You alone are solely responsible for any decisions relating to and determining whether any product or service is appropriate or suitable for you based on your investment objectives and personal and financial situation. BingX, and their affiliates accept no liability for any loss arising from reliance on this content, to the fullest extent permitted by law. Please consider your financial situation and risk tolerance before trading.