Open in interactive viewer → charts, metric popovers & call review

Hitachi Buries the Tariff Story and Rebrands Its Way Into the AI Era

A second upward revision, a brand-new Hitachi Digital unit, and a vocabulary that has quietly lost three of last year's most important words
6501.T · Earnings Call · 2026-10-28

The words that disappeared

Reading Hitachi's Q2 FY2026 call next to its own prior transcripts is an exercise in watching a narrative evaporate. Twelve months ago, the questions were about U.S. tariff exposure, indirect tariff effects on IT spending, and customers putting projects on hold. Kato was explicit: “At the time of April, we did not -- we only understood the direct impact, not the indirect” — Tomomi Kato, Representative Director and Executive Officer · 2025-08-01, and he described how “in the U.S. tariff indirect impact is leading to investment restraint” — Tomomi Kato, Representative Director and Executive Officer · 2025-08-01 at GlobalLogic. The keyword record bears this out — in FY2025 Q1, "indirect impact," "U.S. tariff" and "investment restraint" were literally the company's three highest-ranked themes, in that order.

On 28 October 2026, across the full transcript, tariffs are not mentioned once. Not by management, not by a single analyst. The concern has been replaced by "geopolitical risks" and — new — rising interest rates as a drag on customer demand. That is the single most informative thing about this quarter: the overhang that dominated four consecutive calls has simply been retired, and nobody on the call even bothered to ask whether it's truly gone or merely invisible.

What replaced it: a new company inside the old one

Where tariffs sat, Hitachi Digital now sits — the top-ranked keyword of the quarter by a wide margin. Hitachi is merging GlobalLogic and Hitachi Digital Services into a single operating entity, brand and all ("Engineer What the World Depends On"), with new CEO Anand Birje, who joined from digital-engineering competitor Encora after two decades at HCL. The pitch is unusually well-constructed: seven mission-critical verticals, four service lines, 200 top customers contributing ~80% of revenue, eleven ecosystem partners, and a claim to be one of the very few firms that can do physical AI — OT and IT, plant floor and data layer — inside regulated industries.

The financial reality is more sober. Kato conceded that on a standalone basis the digital engineering business delivered “revenue at the same level as the year before on a standalone basis” — Tomomi Kato, CFO · 2026-10-28, with full-year engineering service growth guided at just 3%. Birje's framing is that “we really look at this year as a year of reset and accelerate” — Anand Birje, CEO of Hitachi's Digital Business · 2026-10-28 — a deliberate culling of low-margin verticals now, in exchange for “higher double-digit growth opportunities in the coming years” — Anand Birje, CEO of Hitachi's Digital Business · 2026-10-28. That is a promise being made by a newly installed executive, on a business that has spent four quarters explaining why it couldn't grow. Watch it closely; the goodwill is real but it is unearned.

It also goes some way to explaining why enterprise platforms and "agentic platforms" now clutter Hitachi's keyword set — that's the new vocabulary of the merged entity, not of the conglomerate's industrial core.

The hardware is where the money actually is

Strip out the rebrand and the quarter was genuinely strong. Q2 revenue rose 15% year-on-year, all four sectors posted higher sales and profit, and Kato noted that “both revenues and adjusted EBITA reached record highs for a second quarter” — Tomomi Kato, CFO · 2026-10-28. Guidance was lifted again — revenue +JPY 300bn, adjusted EBITA +JPY 40bn — following a Q1 raise, with full-year adjusted EBITA margin now guided to 13.6% excluding special factors, up 120bp.

The order book tells the sharper story: power grid orders +47% in Q2 on large HVDC projects in Europe and a U.S. data center project; mobility +69%; and Connective Industries' measurement and analysis orders +77%, driven by semiconductor manufacturing equipment and metrology/inspection. That last number lands squarely in a global theme — KLA, reporting the same day, made process control a top-five keyword with its semiconductor process-control revenue line, and Advantest, SK Hynix, Amkor and ASM all reported into the same capex cycle this week. Hitachi's semi-adjacent slice is small but it is compounding fast.

Capital returns moved with it: the interim dividend is being raised 36% year-on-year and JPY 290bn of buyback was executed in H1 — half the full-year plan. The portfolio is being swept out in parallel — Hitachi Construction Machinery shares transferred, the Astemo stake being sold, an ATM joint venture, and the electric appliances business gone by the end of next fiscal year. This is a conglomerate actively narrowing itself.

The quiet deceleration nobody priced

The most important line of the day was buried in an answer about power grids, and it should be read carefully against last quarter, when the same CFO told an analyst growth would run until 2035:

Up until last year, we said that it will continue to grow until 2030. In the last few months, we are saying that it will continue to grow into 2034. Since the volume itself will be larger, I think the growth will probably slow down.

Tomomi Kato, CFO · 2026-10-28

Note what has happened: the endpoint moved, the growth rate was explicitly guided lower, and the word that defined this business for two years — "super cycle," a top-15 keyword back in FY2024 Q4 — has fallen out of the vocabulary entirely. The CapEx Hitachi sank in FY2024–26 is now arriving as capacity, which mechanically compresses growth rates even as absolute revenue keeps climbing. That is a far better business than the one investors bought in 2024, but it is no longer a scarcity story, and the multiple that goes with scarcity will not survive the transition.

The AI demand behind it is also being stress-tested. Asked directly whether data center projects were showing credit stress or cancellations, Tamai was categorical: “for the projects that we are running today, there have been no delays” — Shinichiro Tamai, Executive · 2026-10-28 — with Hitachi Energy's AI data center exposure running a bit over 10% of that unit's revenue. The question itself is the tell: a year ago no one asked, now the buy-side is asking about counterparty risk on gigawatt-scale orders. Global tape agrees that the trade has cooled — AI data center names ran up 53% over 360 days but are down mid-single digits over the last 90, with 50 decliners against 6 advancers. Hitachi's order book is still accelerating into a tape that has stopped.

Two hinges to watch

The first is rates. Buried in the capital allocation remarks is a genuinely odd juxtaposition: the dividend was raised partly in view of rising interest rates. Simultaneously, Kato flagged that the company will monitor “the impact of rising interest rate on customer demand” — Tomomi Kato, CFO · 2026-10-28. For a Japanese industrial with a fortress balance sheet, higher rates are a tailwind to returns and a headwind to customers at the same time — and the broader market's keyword board this quarter is dominated by exactly this: rising rate environment, Fed hikes, hiking cycles. Every regional bank reporting in the last five days has made the same noise. Hitachi is one of the few industrials naming it as a demand risk rather than a margin one.

The second is China's rare-earth and magnet export controls. Kato confirmed procurement of magnets and motors for CI and mobility is being re-routed, with strategic inventory and alternative suppliers in place, and the impact "limited" so far — but he declined to say which products or how much inventory covers. That is an unpriced option on the downside, and it received less than two minutes of the call.

Net: a record quarter, a second upgrade, a confident new digital story, and three quiet retirements — tariffs, super cycle, and the tariff-era caution that once made this management team sound defensive. The rebrand is the loud thing. The decelerating power grid growth rate is the thing that matters.