US Market Daily Analysis — 2026-05-18
US Stocks · Options · News · Views
The market's main thread is rotating from a pure semiconductor chase into a "semis cool, other sectors catch up" structure. The index was pressured intraday by Middle East risk, then rebounded as the expected strike was de-escalated, but the PHLX Semiconductor (SOX) still opened higher and closed lower. Memory names and the high-beta, extended leaders pulled back hardest — a sign that short-term chase appetite is already fading.
The broad market itself is not broken. Inside the S&P 500, advancers still outnumbered decliners, and consumer and software began absorbing capital rotating out of the high-flying semis. The 10-year Treasury yield consolidated around 4.595%, no longer pressing higher for now, but not yet offering a clear dovish signal for risk assets either.
The main risk to semis is short-term positioning and leverage structure, not a sudden fundamental reversal. Leveraged semiconductor ETFs mechanically buy into rallies and mechanically sell into declines; when volatility expands, both the upside and downside get amplified. Moves like Micron's — gapping up and then rolling over hard — are the direct manifestation of chase risk.
Risk awareness is starting to return to the options market. Call inflows have turned neutral, and put protection has clearly increased. SOXX needs close monitoring around 489; SMH around 533. A break (on an opening or closing basis) would most likely push semis into a sideways consolidation. CTA positioning is elevated — a decisive loss of the 7095 line near the Nasdaq could also trigger systematic de-risking.
NVDA remains the core asset most worth tracking. Market expectations for this print's profit and gross margin are very high; short-term reaction can't be judged on the beat/miss alone, but on a valuation-versus-growth basis NVDA still offers better risk-reward than AMD. If, post-earnings, price pulls back below 211 and completes a chip handoff, that is actually healthier for the slow-grind-higher structure to continue.
AMD has higher elasticity, but the current price has already discounted far-forward growth. 400–420 is intermediate support; below that, support thins materially. IGV is seeing consecutive volume expansion — if software continues to absorb rotational flows out of semis, a short squeeze remains possible. TSLA is back near its 200-day moving average and is neutral. BRK's valuation is on the low side of median but growth is weak. UNH has already worked through its easiest repair zone; digesting the 440–496 heavy-overhead supply will take time.
Short-term, do not chase semis higher — especially memory names and 3x leveraged ETFs that have already ripped. If semis hold SOXX 489 and SMH 533, the door to renewed strength stays open; if those break, prioritize cutting speculative size and wait for a new base and volume repair. NVDA — no rush to trim; re-evaluating above 300 is more reasonable. A post-earnings pullback below 211 is a better long-term chip-handoff zone.
AMD is best managed with support levels: as long as 400–420 holds, the trend has room to extend; once it fails, do not stubbornly hold for a valuation reversion. IGV can continue to be watched for volume expansion and short-squeeze potential toward the 200-day, but a squeeze is not the same as fundamental confirmation. TSLA is best treated as a range-bound name for now. BRK is more suitable for patiently waiting for a breakout signal near 510. UNH is only suitable for slow repair, and only if there is no new major negative catalyst on the health-policy front.