Bitcoin started the year at $87,500 and later slipped to $82,000, while Ether fell from $3,000 to $2,400. Over the same stretch, Micron was up more than 100%, SK Hynix had gained 88%, and MediaTek jumped 78% in just 15 days. That contrast captures how capital has been allocated in 2026.
A single trading day in early May made the gap even clearer. SK Hynix rose 12.5% and Micron added 10%, while Bitcoin climbed only 1.2% on the same day. Memory and AI-linked equities were posting sharp moves in hours; major crypto assets were struggling to match that over months.
Semiconductor names keep drawing the market’s attention
In U.S. equities, the source notes that Micron’s year-to-date gain passed 100% and its market capitalization moved above $700 billion. AMD surged 15% in one session, while NVIDIA kept pushing to fresh record highs. On the index side, the Nasdaq was up 22% for the year and the S&P 500 had advanced 16%.
The rally was not limited to the U.S. SK Hynix climbed 88% on the year, helping South Korea’s stock market break above 7,300 for a record high, and Samsung’s valuation returned to $1 trillion. In Taiwan’s market, MediaTek rose from 1,470 to 3,155, good for a gain of more than 115%. Nanya Technology doubled, Winbond gained 70%, and the Taiex rose 34% in four months to top 40,000.
Crypto fails to keep pace with the AI trade
Against that backdrop, crypto’s annual performance looks weak. Bitcoin was in negative territory for the year based on the figures cited, and Ether was down about 20%. The piece also points to weaker showings from SOL, ADA, and DOGE, though it does not provide exact numbers for those tokens.
The divergence is not only about price action. The article argues that semiconductor and AI shares are being backed by visible orders, capacity constraints, price increases, and clearer revenue channels. Crypto, by comparison, is still leaning more heavily on narrative and asset positioning than on near-term earnings visibility that equity investors can model.
Compute spending dominates as crypto narratives cool
The source frames the current market cycle around AI compute infrastructure. Cloud companies are said to be committing $2 trillion in contracts to secure computing capacity. Anthropic is described as paying Google $200 billion over five years, while OpenAI is paying AWS $138 billion. That spending ultimately feeds into chips and memory, where HBM remains undersupplied, DRAM prices are rising 40% per quarter, and advanced packaging capacity is booked out to 2028.
Crypto is missing that same line of monetization, according to the article. Over the past six months, ETF inflows have slowed, on-chain activity has stayed muted, the Iran war has weighed on risk appetite, and Federal Reserve rates remain high. Bitcoin’s identity keeps shifting between digital gold and a risk asset, but in this market, the author argues that investors have been choosing AI shares, gold, silver, and oil instead.