The first half of 2026 ended on Wednesday, and the technology sector’s H1 scorecard produces a result that would have seemed implausible a year ago: MSCI’s emerging markets technology index, which covers large and mid-cap tech stocks outside the United States and Europe, gained more than 90% in the six months to June. South Korea’s KOSPI surged 101.1%, powered almost entirely by the memory chip wave that Micron, SK Hynix, and Samsung have been riding through AI infrastructure demand. Japan’s Nikkei 225 gained roughly 39%. The pan-European Stoxx 600 Technology index added 23.4%. The Nasdaq 100, home of Nvidia, Apple, Microsoft, and Alphabet, added 19.9%. The U.S. version of the MSCI tech index, the one investors think of as the AI trade, gained 19.4%. Six percent less than Japan. Ninety percent less than emerging market tech. The specific divergence and what it says about where AI investment is generating returns is what the H1 scorecard is actually telling, and it is the story that begins with Wednesday’s final H1 data.
The Magnificent Seven – Microsoft, Nvidia, Apple, Meta, Alphabet, Amazon, and Tesla – lost approximately $2.3 trillion in combined market capitalization in June alone, as investors applied renewed scrutiny to whether the hundreds of billions being spent on AI data center infrastructure will generate returns on any timeline that justifies current equity valuations. The CNBC Magnificent Seven Index fell 10% in June. Microsoft shed roughly 22.9% of its value in the first half of the year. Nvidia, despite its position as the primary supplier of AI chips, gained only 7.3% for the six months. The contrast with the companies that supply components to the infrastructure these same hyperscalers are building was stark: the Philadelphia Semiconductor Index, which includes TSMC, Micron, and ASML, gained more than 90% year-to-date. The Roundhill Memory ETF, tracking SK Hynix and Samsung specifically, rose 166%. That supplier-versus-buyer divergence is what NewsTrackerToday picks up as the H1 story in its cleanest single-line form.
Liam Anderson reads the divergence without ceremony: “Mag 7 down 10% in June. Philadelphia Semiconductor Index up 90% for the year. Memory ETF up 166%. The market is rewarding chip suppliers and punishing chip customers. The buyers are spending more than they can earn on it yet. The sellers have pricing power that their buyers cannot refuse. That’s the trade that worked in H1.” Deutsche Bank strategist Jim Reid identified four factors behind the Mag 7’s June underperformance: unwinding of extreme positioning, concerns over AI hyperscaler capital expenditure ROI, a more hawkish Federal Reserve stance under Chair Kevin Warsh, and rising chip costs. The chip cost factor is the one that most directly connects the Mag 7’s equity decline to the memory index’s extraordinary gains: Apple raised MacBook prices $300 because memory costs have quadrupled; Micron posted 84.9% gross margins in Q3; the same pricing event that benefits Micron’s shareholders pressures Apple’s.
Ethan Cole reads the macro context: “KOSPI up 101% in six months. South Korea is an economy that runs on semiconductor exports. When AI infrastructure spending accelerates globally, South Korea’s stock market outperforms every major Western index by a factor of 5 to 10. That’s not a coincidence. It’s a supply chain consequence.” Tom Lee, head of research at Fundstrat Global Advisors, offered the framework that may define H2: the market is trying to understand the new narrative around the Mag 7, which moved from asset-light companies producing free cash flow to balance-sheet-intensive capital allocators. His argument is that investors will eventually reframe the AI capital expenditure as a productive asset rather than a cost, at which point the Mag 7’s repricing reverses. The KOSPI-Nasdaq divergence is what NewsTrackerToday draws as the test of that thesis: if AI capex validates in Q2 earnings commentary, the buyer-versus-seller trade reverses and U.S. tech closes the gap.
The second half of 2026 will be determined by a question H1’s data cannot answer: do Q2 earnings from Amazon, Microsoft, Alphabet, and Meta show AI infrastructure investment converting into measurable revenue, or do the numbers confirm that the spending cycle is still in its pre-revenue phase? Tom Hulick, CEO of Strategy Asset Managers, told reporters on the final trading day of H1 that markets are “very fluid” but nowhere near catastrophic failure. Dan Ives of Wedbush Securities described H1 as a series of “gut check moments” in a trade he still believes points toward structural AI adoption. The Q2 earnings season that opens in mid-July is the event that NewsTrackerToday stays with as the fulcrum: either the revenue story arrives and U.S. tech re-rates, or it doesn’t and the KOSPI-Nasdaq divergence represents a durable reallocation of AI investment returns from buyers to suppliers.
The uncomfortable implication of H1 2026’s technology scorecard is that the AI story’s financial beneficiaries were not the companies that captured the most media attention. The top performers were not the frontier model makers, the super app builders, or the robotaxi operators. They were the companies that manufactured the raw material – the memory chips and logic chips and semiconductor equipment – that all of those companies needed and could not get enough of. KOSPI at 101%, Memory ETF at 166%, Philadelphia Semiconductor at 90%: three expressions of the same basic dynamic. The AI revolution has, so far in 2026, been most financially rewarding to the people who sell shovels into a gold rush. And what News Tracker Today names as the open question for H2 is whether the gold seekers – the hyperscalers, the frontier labs, the enterprise AI deployers – eventually start showing returns that justify what they paid.
