DRAM/NAND bullishness is being reinforced by a sharp reset in pricing expectations. The key read-through is that the memory cycle is stronger than many legacy frameworks assume, with Q1 2026 PC DRAM Blended ASP revised from +50~55% to +110~115%, Server DRAM from +60~65% to +93~98%, and Mobile DRAM from +45~50% to +58~63%. That kind of upward revision supports staying constructive on memory rather than fading the move too early. 1
China-exposure skepticism is getting more entrenched. Even if Japan shifts back toward a more pro-China stance, the harder problem is demand recovery: Chinese consumers may not return to Japanese brands after replacing them with domestic alternatives. The takeaway is blunt—making money in China is structurally harder now, not just politically noisy. 2
GOOGL backing a Texas data center leased to Anthropic points to AI capex staying massive and long-duration. The setup is not incremental: the Nexus construction loan is expected to exceed $5 billion, the site spans 2800 acres, is projected to deliver about 500 megawatts by the end of 2026, and ultimately scale to about 7.7 GW. This reads like hyperscalers and model labs continuing to lock in power and capacity ahead of demand, not tapping the brakes. 3
NVDA still screens as a core AI infra asset even after the drawdown. The valuation angle is the key punchline: it is being framed as the cheapest it has been in a decade while still serving as the compute backbone of the AI buildout. That combination argues the market may be derisking the multiple faster than the long-run infrastructure thesis is deteriorating. 4
Forward estimates imply large-cap AI leaders are getting rerated hard enough to look optically cheap. $NVDA and $META approaching sub 15 on 2027 Forward P/E is notable because it suggests a lot of skepticism is already in the tape if earnings power holds. 5
Microsoft screens as a quality growth compounder trading below its usual premium. The numbers matter: current PE is 22, below its historical average; 2027 forward PE has fallen below 20; revenue is growing 17%; annual EPS growth is close to 20%; gross margin is above 50%; net margin is close to 40%. For a company of this size, that mix still looks like elite-quality cash generation without an excessive multiple. 6
In panic tape, technicals can get steamrolled, so factor selection matters more than signals. The cleaner setup is to lean into companies with low PE, high growth, and strong revenue, because fundamental resilience carries more weight when war headlines keep risk assets under pressure. Averaging in from the left side can maximize rebound upside, but the trade-off is obvious: cash can run out before the market finds footing. 7
Sharp bear-market rallies after an initial washout are common, and history says not to confuse them with a durable turn too quickly. The analogs cited are precise: 2008 fell 19% first, then rallied 14% before more downside; 2018 fell 11%, then rallied 8%; 2020 fell 15%, then rallied 10% before the next leg. The message is to respect reflex bounces but avoid declaring all-clear too early. 8
A second consecutive year of extreme equity damage would be historically unusual. Since the end of the GFC in 2009, there have been no back-to-back years with a 20%+ drawdown in the $SPX. That does not rule out another ugly year, but it raises the bar for assuming a repeat of a 20%+ hit in 2026. 9
Seasonality is offering a tactical tailwind into April. The data points are specific: the last 13 Aprils closed green in 10 of 13 cases, a 77% win rate; since 2005, when the $SPX was red in March, April closed green in 6 of 7 cases, an 86% win rate. Not a guarantee, but enough to justify staying open to an April relief rally. 10
Relative strength is highly concentrated, and that concentration is itself the signal. High-beta winners with strong momentum are making new cycle highs versus high-beta losers, which argues for buying stocks getting dragged down by market-wide pressure rather than names whose own weakness is breaking the tape. In this regime, leadership matters more than broad bargain hunting. 11
Drawdowns are not automatically thesis-breakers. The edge is less about avoiding 50% paper losses and more about staying anchored when volatility is trying to shake out conviction. That mindset matters most when the underlying thesis is intact and the tape is pushing emotional, not analytical, decisions. 12
Biotech can become investable once the universe is tightened with the right filters. Adding a market-cap threshold and focusing on high short ratios turned a historically avoided group into a productive hunting ground, especially over the past year. The broader takeaway is that a bad sector can become a good strategy when the basket construction improves. 13
The metals supply problem is still being badly underestimated. Higher energy prices, higher interest rates, and ongoing wars are all headwinds to bringing on new mine supply, not solutions. With institutional capital shortages and technical constraints still in the system, the path to producing the metals the world urgently needs looks tighter than consensus is pricing. 14