The question Tiger SG is asking, what did you miss in H1, is more uncomfortable than it looks. Because the answer for most investors is not a single stock. It is a structural misread of how the entire market was rotating underneath the headline numbers.
The S&P 500 rose 9.5% in H1 2026, slightly behind its historical annual average of 12.8%. That sounds orderly. It was anything but. Beneath the index, the old winners became the laggards. The hyperscalers, Microsoft, Alphabet, Amazon, Meta, and Oracle, solidly underperformed the market. Microsoft was on track for its worst monthly loss since 2008, down 20% in June alone. Oracle fell 30%.
Meanwhile, investors piled into memory chip companies whose products help power AI. Samsung, Micron, and SK Hynix are now the 10th-, 13th-, and 14th-most-valuable public companies in the world, surpassing Berkshire Hathaway and JP Morgan Chase.
If you spent H1 holding Mag 7 names and wondering why your portfolio was flat, that rotation is the answer.
What H1 Actually Was: The Three-Act Play
Act One was Iran. The conflict that erupted in late February sent oil from the $70s to above $116 at the peak. After rising more than 64% in the early stages of the conflict, Brent crude declined 19.3% and 20.8% in April and May respectively, leaving it virtually unchanged from pre-war levels by June. The energy spike fed directly into inflation, resetting rate expectations and putting every high-multiple tech name under immediate pressure.
Act Two was the memory supercycle. The best stock in the S&P 500, SanDisk, was up around 800%. SanDisk surged 257.88% in Q2 alone, leaving it up over 4,900% for the past year. Micron was up roughly 300% year to date. The VanEck Semiconductor ETF gained 82% in the first six months of the year, its strongest first-half performance since its inception in May 2000.
Act Three was the pivot that most retail investors missed entirely. The AI trade did not die in H1 2026. It bifurcated. The hyperscalers will spend $805 billion on capex this year, up from $449 billion in 2025, consuming 93% of their cash flows from operations compared with 33% in 2023. That number is the market's entire concern about Microsoft, Alphabet, and Amazon. The spend is real. The return is delayed. So the market sold the spenders and bought the suppliers.
The stock most investors wish they had bought in H1: SanDisk at its April lows following the memory selloff. The stock most investors bought in H1 that hurt them: any Mag 7 name held without trimming into the AI capex narrative.
The Macro Setup Entering H2
This is where the context for the next six months sits, and it is more complex than the Tiger SG prompt suggests.
In his first meeting as Fed Chair, Kevin Warsh struck a surprisingly hawkish tone at the June FOMC, expressing greater inflation concerns than markets expected. The most recent reading of core PCE rose for the third consecutive month to 3.4%. Markets are now pricing in 53% odds of a rate hike before year-end, with Kalshi pricing 68% probability of a hike before July 2027.
That is the single most important number for H2. If Warsh hikes, every high-multiple name in your portfolio re-rates downward. Memory stocks at 8 to 10x forward earnings are relatively insulated. Software names at 30 to 40x forward earnings are not.
June's jobs report added just 57,000 jobs, roughly half of expectations, with the prior two months revised down by 74,000 combined. That weakened the hike case marginally. But Cleveland Fed Governor Beth Hammack explicitly stated that AI infrastructure demand is fueling inflation and that rate increases may be needed if elevated prices persist, framing AI capex as a demand-pull inflation driver rather than a cyclical or tariff-driven phenomenon.
Read that carefully. A voting FOMC member just said AI spending is itself inflationary. That is a structurally new thesis that did not exist in any prior rate cycle. If it gains traction inside the Fed, it means the very engine driving the market's best performers is simultaneously the argument for raising the rates that would compress their multiples. That is the most dangerous macro feedback loop of H2 2026.
The Shiller CAPE ratio is currently at 41. Its historical average is 17, and it has crossed 40 only twice before: during the dot-com era, when it peaked at 44 before the S&P 500 fell 49%, and briefly in January 2022 before a nine-month bear market wiped out 25% of the index. That is not a prediction. It is context for how much forgiveness the market has for disappointment.
The H2 Watchlist: What Actually Matters
NVDA is the most important single name to watch in H2. NVDA's fiscal year 2026 revenue was $215.94 billion, up 65.47%, with earnings of $120.07 billion. Vera Rubin is now in full production with early adopters including OpenAI, Anthropic, xAI, Dell, Oracle, CoreWeave, and all three major HBM4 memory suppliers qualified and shipping. The Rubin supercycle launching into H2 is the single biggest positive catalyst in the market. The stock is down 17% from its all-time high despite record earnings. An RSI of 28 with positive divergence at channel support creates a tactically oversold setup. Fifty-nine analysts have a Strong Buy consensus with a mean target of $298 implying 40%+ upside from current levels.
Memory stocks, specifically MU and SNDK, remain structurally sound through 2027 per multiple CEO commentaries. But after 800% and 300% year-to-date gains respectively, the easy money is behind you. On July 1, SanDisk shed nearly 9% and Micron dropped 7% as investors used the calendar turn to harvest gains. The thesis is intact. The entry point is no longer obvious.
The Mag 7 laggards, specifically MSFT and META, are the contrarian play for H2 if the AI monetisation narrative starts to confirm. YTD, NVDA, Apple, Alphabet, and Amazon are merely marginally higher. Microsoft, Meta, and Tesla are down big. The bearish consensus on hyperscaler capex is now fully priced in. Any confirmation that Azure is accelerating above 40%, or that Meta's AI infrastructure is generating ad revenue at scale, turns these names into the biggest H2 surprises. Microsoft's July earnings and the Copilot adoption numbers are the catalyst to watch.
Energy is the sector nobody is talking about on watchlists but everybody should have a view on. If the Iran ceasefire holds and oil stays near pre-war levels, energy unwinds and inflation fears ease, loosening the rate hike probability and re-rating growth stocks simultaneously. If ceasefire talks collapse, oil spikes again and the entire macro picture resets negative.
The SPCX IPO is a H2 wildcard. SpaceX listed at $135 on June 12. With a 4% float and index inclusion beginning within days of listing, the structural buying mechanics alone could drive meaningful moves regardless of fundamentals. First earnings in November 2026 is when the reality check arrives.
The Uncomfortable H2 Reality
Most investors framing their H2 watchlist are making the same mistake they made in H1: anchoring to last half's winners.
SanDisk at $2,300 is not the same trade as SanDisk at $40. The structural shortage thesis is real. But a name up 4,900% in 12 months needs to deliver exponential earnings growth just to hold its multiple, let alone re-rate higher. The margin for error is zero.
NVDA at $200 with Vera Rubin in full production and a 40% upside to consensus target is a more interesting risk-reward than chasing memory names at their peaks. That is the trade most people are not making because NVDA underperformed memory YTD and recency bias is doing its work.
The most honest H2 watchlist: NVDA for the Vera Rubin cycle, MSFT for the Copilot monetisation confirmation, one energy name as an Iran hedge, and cash equivalents as genuine protection against the 53% rate-hike probability that the market is not fully pricing into growth stock multiples.
The regret trade of H1 was not buying SanDisk at the April lows after the selloff. The regret trade of H2 will most likely be not owning NVDA at $200 when Vera Rubin ships at full scale into Q3.
I am not a financial advisor. Trade wisely, Comrades.
Comments