Goldman Sachs: Japan's Semiconductor Industry May Be Entering Its Largest-Ever Cycle
0xBroomberg
Goldman Sachs overhauled ratings and price targets for 22 Japanese electronics/semiconductor companies, calling this AI-driven cycle potentially the largest and longest in history — and still in its early stages.
Why does Goldman call this cycle "the largest ever"?
Goldman sees two waves: the first from AI server upgrades and rising component intensity; the second from mass adoption of edge-AI and physical-AI devices.
This means → it is not a single spike. Two waves stacking could push the cycle's duration and scale past every prior semiconductor upcycle.
Goldman places the industry in the early stage of this cycle, forecasting that earnings, CROCI, and ROE will surpass historical peaks.
How is Goldman picking winners?
Goldman scores the 22 companies on three dimensions: ① how much AI lifts earnings short- and medium-term; ② how badly rising material costs hit near-term margins; ③ whether management can execute in a fast-changing environment.
In plain terms = an AI label alone is not enough. A company must also absorb cost pressure and have competent leadership — all three boxes checked.
This reflects Goldman's core thesis: the earnings-growth gap between companies will keep widening. Picking the right stock matters more than picking the right sector.
Group one: core AI beneficiaries — who tops the list?
Ibiden (ABF substrate maker, Buy): price target raised sharply. The stock has been a market favourite since 2025.
This means → ABF substrates — a key material in AI chip packaging — sit at one of the most certain links in the AI hardware demand chain.
Murata (MLCC maker, Buy, Conviction List), Taiyo Yuden (Buy) both see target hikes. Renesas Electronics (power-management chip leader, Buy) has been Goldman's top pick since early 2026.
Group two: who is being pulled into the AI orbit?
Goldman keeps Buy ratings on TDK, MinebeaMitsumi, and Nidec, and upgrades Rohm from Neutral to Buy — citing growing earnings contribution from SiC and other AI-linked semiconductors.
Among Neutral-rated names, Hirose Electric, Nichicon, NGK, and Niterra get increased attention as likely beneficiaries of AI and semiconductor expansion.
Kyocera stays at Buy, positioned as a laggard catch-up play. In plain terms = the stock price has not yet reflected improving fundamentals.
Group three: who risks falling behind?
Nitto Denko is downgraded from Buy to Neutral. Earnings growth is steady but Goldman sees no visibility on non-linear upside — put simply, growth is linear with no clear catalyst for acceleration.
Japan Aviation Electronics is cut from Neutral to Sell: earnings growth in autos and smartphones is decelerating.
Mabuchi Motor stays at Sell with elevated downside risk to first-half earnings. Alps Alpine stays Neutral — material-cost headwinds are significant, and an operating loss is possible in Q1.
What does this mean for investors?
Goldman's framework is clear: in the early stage of an AI cycle, stock selection beats sector bets. Within the same Japanese semiconductor universe, some companies get sharp target hikes while others are downgraded to Sell.
This means → buying "a basket of Japanese semis" could deliver cycle gains and laggard drag in the same portfolio.
This reflects a broader signal: AI hardware demand is already diverging — companies close to the AI infrastructure core are accelerating; those further away risk having margins consumed by cost pressure.
Content is for reference only, not financial advice.