Mitsubishi Heavy Booked a Trillion Yen More and Guided Profit Exactly the Same
A ten-year-old South African steam project ate the risk buffer, the GTCC windfall is still customer cash, and the Logisnext exit is the first real proof of portfolio management.
7011.T · Earnings Call · 2026-11-07
The trillion-yen order book that didn't reach the profit line
Mitsubishi Heavy's second quarter is a study in asymmetry. Order intake for the half came in at ¥3.3 trillion, the order backlog swelled to ¥11.5 trillion — up ¥1.2 trillion from the prior year-end — and the full-year intake outlook was lifted to ¥6.1 trillion. Almost all of that lift is one segment: Energy Systems order intake went from ¥2.2 trillion to ¥3.2 trillion, with Eisaku Ito noting that “the Energy business is going to increase by JPY 1 trillion compared to the initial guidance” — Eisaku Ito, Executive (likely CEO or CFO) · 2026-11-07 on the back of GTCC demand.
And then business profit guidance did not move: ¥390 billion, unchanged. That is the whole story in one line. A company that just added a trillion yen of forward revenue kept its earnings promise exactly where it was, and the reason is buried in a single line of the profit bridge — a ¥30 billion one-off charge, ¥20 billion of it for a South African thermal project.
The ghost of the Hitachi deal
CFO Hiroshi Nishio was unusually candid about the provenance of that charge: “This is a project that we have inherited when the thermal business was integrated with Hitachi. Construction has continued for more than 10 years and operation has started by phases over the years.” — Hiroshi Nishio, Executive (likely COO or similar) · 2026-11-07 The twelfth and final unit started operating in September 2025; MHI is still negotiating with the customer over who eats the construction-phase overrun, and booked the provision for accounting purposes anyway.
The more telling detail is what it consumed. On the prior call management had explicitly built a ¥20 billion risk buffer for onetime expenses into the Energy plan. This quarter that buffer was spent, and then some:
Last time, we have provided for JPY 20 billion risk buffer for onetime expenses that was included. But this time, in Q2, the one-off expense has already been recognized in JPY 30 billion, so that no buffer is incorporated for in the second half.
So H2 now runs without a shock absorber in the segment that carries the entire growth story. First-half business profit landed at ¥168 billion on a like-for-like basis (excluding Logisnext) versus ¥188.4 billion a year earlier — a decline, masking a genuinely strong ¥76 billion of revenue and margin improvement that the provision wiped out.
Thirty percent more capacity, again
Ito framed the GTCC ramp as a fresh act of focused resource deployment: “to securely answer strong demand, we are going forward to increase production volume by 30%... We are not only talking about CapEx or human resource recruitment.” — Eisaku Ito, Executive (likely CEO or CFO) · 2026-11-07 It is a good line. It is also not new. Eighteen months ago, pressed on exactly this, the same management team said: “So in terms of capacity, 30% increase. We've been speaking about that in the past.” — Hisato Kozawa, Executive (likely CFO or similar senior financial officer) · 2025-02-08
The keyword record confirms the theme is aging rather than breaking out. Gas turbines, data center power demand and advanced payment were the dominant company keywords in early 2025; this quarter's ranked list is instead headed by order intake, Mitsubishi Logisnext and wide optimization. The conversation has migrated from is demand real to how do we allocate capital around it. That is a maturer, less-promotional framing — and it also means the easy narrative juice has been extracted.
Customer money, not earnings
Free cash flow of ¥151 billion in the half looks like a triumph until you read the balance sheet: contract liabilities rose ¥260 billion, driven by advance payments on GTCC orders. Management conceded this is borrowed-in-time, not earned. The prior call said it plainly — “it's still an just an advanced payment. At some point, we may need to kind of use that” — Hisato Kozawa, Executive (likely CFO or similar senior financial officer) · 2025-02-08 — and the same logic was repeated in February when an analyst asked whether the cash pile unlocks dividends.
The portfolio response is the genuinely new item. Selling down the capital relationship with Mitsubishi Logisnext, announced September 30, removes a cyclical forklift business from the order, revenue and profit lines and gives management a demonstration of what it calls comprehensive portfolio management. It also lowers the bar: the outlook now excludes ML, and Ito expects the Energy uplift to more than cover the hole.
Defense is the other soft spot worth naming. FY25 defense orders of ¥1.2 trillion are down against the boom years of FY23 and FY24, and management pre-empted the question by pointing to the Australian general-purpose frigate selection announced August 1 — a win that still requires Japanese government coordination and international contract negotiation before it becomes a number. On domain expansion, Ito simply declined: “we will refrain from providing the detailed explanations today.” — Eisaku Ito, Executive (likely CEO or CFO) · 2026-11-07
The awkward market backdrop
Here is the contrast that makes this report more than an accounting note. MHI's order book is levered to the same force — AI-driven electricity demand — that equity markets have spent the last quarter aggressively de-rating. Across the global tape, AI data center exposure has fallen roughly 29% over ninety days with 57 names trading down, and data center capacity is off more than 40%. Meanwhile the customer base MHI sells turbines to keeps signing. One of these two things is wrong.
MHI is not positioned as an AI trade, and it does not want to be. When analysts pushed in February on whether GTCC orders were data-center-direct, the answer was that demand arrives through utilities, indirectly. That conservatism is now a virtue: the company's backlog was built on power scarcity and grid replacement, not on hyperscaler capex ticks. Its US exposure — ¥1.1 trillion of sales last year, with manufacturing and service footprints onshore — also quietly insulates it from the tariff question that dominated so many other industrial calls this quarter.
The honest reading: this was a strong quarter dressed in flat guidance. Underlying margin improvement of ¥76 billion, a record second-quarter net income of ¥114.9 billion, and a backlog that grew faster than revenue all argue the operating machine is working. What holds the story back is that the incremental trillion yen of demand arrived with a ten-year-old construction dispute attached, a spent buffer, and free cash flow that belongs to customers. Investors get the growth; they just don't get the earnings yet — and management has now removed its own safety net for the half where the GTCC conversion was supposed to show up.