Siemens Keeps Buying Software With Money It Earns Selling Hardware. That's the Whole Plan.

In FY2024 Siemens posted record net income of €9.0 billion⁷ — and in the same year its flagship Digital Industries unit watched its revenue fall, dragged down by the very factory-automation hardware that built the company. Two facts that look contradictory are actually the same story. The hardware that makes Siemens famous is the thing it is quietly trying to depend on less. And it has spent the better part of two decades, and roughly $18 billion — the sum of three deal values measured on differing bases (debt-inclusive price, enterprise value, and equity value) — buying its way out.

The official story is that Siemens is a 170-year-old industrial giant — turbines, trains, the gear inside factories. The truer story is that Siemens has been running a single, patient capital-allocation play for eighteen years: take the lumpy cash thrown off by cyclical hardware and convert it into recurring industrial software that doesn't swing with the inventory cycle. The hardware isn't the destination. It's the funding source.

Three checks, eighteen years, one idea

The pivot did not begin with Mentor Graphics, the deal most people remember. It began in January 2007, when Siemens agreed to buy UGS — a product lifecycle management company — for $3.5 billion including assumed debt.¹ That was the first leg. The second arrived in November 2016, when Siemens announced it would acquire Mentor Graphics, the chip-design (EDA) software house, at $37.25 a share — an enterprise value of $4.5 billion, a 21% premium to the prior day's close.² The deal cleared CFIUS and closed in March 2017.³ The third, and largest, came in late 2024: Altair Engineering, a simulation and data-science company, for $113.00 a share — an equity value of about $10.6 billion, an enterprise value near $10 billion.⁴ It completed in 2025.⁵

Notice the shape. Each piece slots into the engineering workflow the one before it left open: design the part (PLM), design the chip inside it (EDA), then simulate the whole thing before it ever exists (simulation and AI). This isn't three opportunistic deals. It's one product being assembled across two decades and three checks that grow as the conviction does.

Why the hardware can't be the answer

To see why the software matters, watch the automation business breathe. Through the upcycle, Digital Industries was a money machine: in Q2 FY2023 it reported all-time-high profit, with full-year margin guidance raised to 22.5%–23.5%.⁸ Then the cycle turned. Customers had over-ordered through the supply-chain panic and spent the following years working that inventory back down — destocking — and automation revenue fell, dragging DI's reported numbers down with it in both FY2023 and FY2024.⁶⁷ Same business. Opposite math. A year apart.

Here is the move that makes the strategy coherent. In the very quarters automation was sliding, the software inside DI was going the other way — significant growth in FY2023 from large PLM and EDA contract wins, and higher software revenue again in FY2024.⁶⁷ Software didn't follow the inventory cycle, because nobody destocks a software license the way they destock a stack of motion controllers. The pivot is therefore not a fashion. It's structural insurance: each acquisition buys a slice of revenue that keeps growing on the exact days the hardware is shrinking.

The objection that could be right: SaaS

The honest counter is that buying software is the easy part; earning on it is the hard part — and there's a live reason to doubt the math. As Siemens shifts PLM from upfront license sales to subscription pricing, it trades a big check today for a thinner stream stretched over years. In FY2023 that transition visibly held back reported software revenue growth and margin even as the underlying business was winning contracts.⁶ If that drag is temporary — the familiar J-curve every software company crosses on the way to recurring revenue — the thesis holds beautifully. If it's permanent, then Siemens has paid hardware-cycle prices, three times over and rising, for a software base that never reaches software margins. The whole bet rests on which of those two it turns out to be, and the FY2023 filings can't yet tell you. That's the genuine open question, not a rhetorical one.

Siemens has done the unglamorous thing the way it does most things: slowly, expensively, and with a straight face. Eighteen years, three companies, and a budget that climbed from $3.5 billion to nearly $10 billion as the conviction hardened. The factories still hum and still swing with the inventory cycle. But every few years Siemens takes the cash those swings throw off and buys another piece of a business that doesn't swing at all. The genius isn't the software. It's deciding, before the market forced the choice, that the thing paying the bills shouldn't be the thing you're betting on.