🔥 AI capex outlook loosens, crypto market must watch for capital diversion

TSMC fell 4.6% before the bell. Even with second-quarter net profit hitting a historic high, it didn’t help. The market is pricing a steep capex curve—spending over the next three years is set significantly higher than the past three years. Full-year guidance was raised to $60.0-$64.0 billion, with Arizona also to add another $100 billion.
AI demand is seen through 2030, but the marginal returns on capex are being questioned. The cautionary tale: IBM plunged 26%. When a giant burns money for AI infrastructure, money flowing out of risk assets is a natural reaction. As a high-beta asset class, crypto is often the first to feel pressure during macro capital rebalancing.
A chain reaction in semiconductors is already visible: memory and optical communications both fell. SK Hynix ADR dropped more than 6%, while Western Digital fell more than 7%. On-chain leveraged capital is facing a double squeeze—falling underlying asset prices paired with funding rates turning negative. SK Hynix’s on-chain contract funding rate briefly hit -55%, as traders betting on a reversal accumulate risk.
Crypto’s current structural fragility lies in this: the AI narrative was one of the core stories that attracted incremental capital. When AI capex expectations shift from good news to concern, crypto doesn’t lose just capital flows—it loses its narrative anchor as well. Bitcoin’s bullish options strike price slipped by $10,000, and traders’ expectations are cooling.
Downside risk: if AI capex continues to run hotter than expected, traditional markets will siphon away liquidity, making crypto’s existing-position tug-of-war even more brutal. The liquidation spiral for on-chain leverage hasn’t ended, and mirrored risks in the Korean market still hang overhead.
$btc #sk #ai #defi #On-chain data
#sk #btc #区块链 #Crypto market #coinsphere