全球半导体与设备：如果存储不行半导体设备还能行吗？

Global Semis — Can semi cap work if memory doesn't? In the current volatile market, many investors agree with our long semi cap call, but worrythat they are all correlated with memory share prices anyway. Today’s note reviews theshare price history and debunks the myth. David Dai, CFA+852 2918 5704david.dai@bernsteinsg.com Historical correlation of memory and SPE share prices is not high. We use top 3companies and top 5 WFE companies as benchmark. During 2012-2018, memory and SPEcompanies correlation was only 0.4. Since 2019 the correlation has increased, but only to0.6. Conversely, SPE and SOX have had consistently high correlation of 0.8-0.9. Mark Li+852 2123 2645mark.li@bernsteinsg.com Stacy A. Rasgon, Ph.D.+1 213 559 5917stacy.rasgon@bernsteinsg.com SPE and memory share prices can go in different directions. There have been a fewnoticeable periods: From Jan 2015-Dec 2016 WFE was up 21.9% while memory wasdown 16.2%, gaining 38.2% outperformance in two years; In Jan 2021-Dec 2022, WFEwas up 15.3% while memory was down 34%, gaining 49% outperformance in two years.While we are positive on both WFE and memory from here, investors can still benefit fromdiversification between WFE and memory, rather than worrying that both must go up anddown together. This is especially the case considering that WFE growth is needed to addmemory capacity, which is the cause of memory price volatility. Juho Hwang+81 3 6777 6980juho.hwang@bernsteinsg.com Carmine Milano, CFA+44 20 7762 1857carmine.milano@bernsteinsg.com Edward Hou, CFA+852 2123 2623edward.hou@bernsteinsg.com Mean reversion?Memory and WFE performance might be the same in the long run… aswas the case during 2011-2019. While memory outperformed WFE by a whopping 661%since June 2025, it was just reversing the lengthy underperformance during 2019-June2025, and only managed to catch up in Feb 2026 when both were up 36x in15 years. Nowthat memory has significantly outperformed WFE, if mean reversion happens it could favorWFE from here. Yipin Cai, CFA+852 2123 2669yipin.cai@bernsteinsg.com Fundamentally, we like WFE long here. Memory capex is clearly accelerating; SK hynixhas recently announced their KRW 100tn (~$67bn) additional investment in their newCheongju fabs, and the Korean government seems to be considering measures to supportthe initiatives from Samsung and SK hynix in building fabs in the Southwestern Korea. Wesee consensus upward revision potential for WFE market and company EPS through 2028,and we expect further growth ahead. For memory, though we don’t think LTA can help much (replay of our recent webinar),the shortage & the imminent HBM price hike will still bring significant upward earningsrevisions. We still like Samsung, SK hynix, & Micron esp. after the recent pullback (report).We rate KIOXIA Underperform on valuation & long-term threat from China (report). BERNSTEIN TICKER TABLE INVESTMENT IMPLICATIONS We rate TEL (TP=¥59,200), Advantest (TP=¥39,200), Disco (TP=¥85,000), Lasertec (TP=¥50,000), Kokusai (TP=¥8,240.00), ASML (TP=€2,300.00) and Besi (TP=€280.00) Outperform. We rate Screen (TP=¥12,600) Market-Perform. Samsung Electronics:We rate Samsung Electronics Outperform with price target of KRW 440,000. SK hynix :We rate SK hynix Outperform with price target of KRW 3,300,000. Micron:We rate Micron Outperform with price target of US$1,300.00. KIOXIA: We rate KIOXIA Underperform with price target of JPY 40,000. AMAT (Outperform, $525.00):Exposure to key inflections is strong and valuation vs peers remains attractive. LRCX (Outperform, $340.00): The company is benefiting from key inflections (GAA, packaging, HBM, NAND upgrades) andCY26/27commentary seems supportive. KLAC (Outperform, $197.50 PT):Amid positive WFE trends KLAC posseses structural growth drivers, a strong and durablecompetitive position, lower China replacement risk, and disciplined capital allocation, warranting premium valuation, in our view. DETAILS Share price has been volatile recently, especially for memory stocks. A big question from investors today is whether semicapcan still work if memory stocks stop working. As such, we conducted a historical analysis and conclude that while the overallcorrelation between memory and semicap is strong, there has been multiple instances where semicap continued to work eventhough memory stocks were muted. Exhibit 1 shows the correlation between memory, semicap and the broader semiconductor sector. It can be observed thatthe correlation between semicap and the overall semis has been historically tight — averaging at 0.82, whilememory hashistorically shown a weaker correlation with other sectors, with both correlations (vs. WFE and vs. SOX) averaging at 0.54.The monthly returns show a similar pattern as in Exhibit 2. WFE vs. SOX move more or less in tandem, whereas it is somewhatnoticeable that memory moves idiosyncratically to the broader semiconductor peers, often for an. EXHIBIT 2:WFE vs. SOX correlation is usually more