Europe's electrical champions are selling the equipment of a generation. Their share prices assume the cash keeps compounding long after the boom, and the evidence that it will, the service that follows the equipment, has not yet reached the filings.
At their October 2 closes, the three European leaders of electrification are priced for three different things. Schneider Electric, at €303.00, needs its trailing free cash flow of €5.79 billion to grow 14.6% a year for nine years before a buyer earns 10% a year. If the company meets the top of every range in its own plan for 2030 and keeps that pace through 2036, the price offers about 9.3% a year. Siemens, at €276.10, asks for 6.6% to 9.1% a year, close to the high single digits its own framework targets, but that cash includes all of Siemens Healthineers, 31% owned by others at the last year-end, and interest received without the interest Siemens pays. Prysmian, at €129.80, asks for 10.3% a year on its 2026 guidance, 16.2% once a €550 million customer advance is set aside, and 18.7% on the €978 million it actually generated in the twelve months to June, a figure the company calls substantially stable on the year. The idea that would justify paying for that persistence, that today's equipment becomes tomorrow's recurring service, is not yet in the numbers: in the April to June quarter Schneider's field services grew 5% while its systems grew 28%, and Siemens's infrastructure services grew 7% while the business around them grew 13%.
A data center is built twice. First it rises in concrete and steel. Then, before a single server is switched on, it is built again out of electrical equipment: switchgear and transformers, uninterruptible power supplies, cooling, and cable by the kilometer. Three European companies sell a great deal of that second building, and in 2026 all three have been reporting records: Schneider Electric a record first half, Siemens a record third quarter, Prysmian its best quarter yet.
The three names, and the question, come from The Economist. On September 27 it argued that the damage to European industry is concentrated in a handful of sectors, carmakers above all, while other parts are thriving. It named Schneider Electric, Siemens and Prysmian among the winners of the artificial intelligence boom, as makers of the gear that keeps data centers running, and noted that Europe's manufacturers are selling more services, from maintenance contracts to software ("European industry is doing better than you may think").
The investment case has two halves. The first is the boom: data centers, grids and factories electrifying at the same time, with order books to match. The second is subtler, and worth more if it is true. Equipment that has been energized must be monitored, maintained, repaired, upgraded and increasingly run through software for as long as it carries current. If the company that sold it keeps the relationship, this decade's shipments become an installed base, and the installed base becomes an annuity that outlasts the boom.
The draft behind this note tested that second half with care, and its central sentence deserves to be kept: a plausible mechanism is not a measured profit pool. Between a purchase order and a shareholder's return lie five links, and any of them can break.
| The link | What would show it | What could break it |
|---|---|---|
| Orders to energized equipment | Commissioning and use, cohort by cohort | Permits, grid connections, labor, idle capacity |
| Equipment to service access | The share of the installed base under service | Independent service firms; customers who do it themselves |
| Access to recurring revenue | Renewals, retention and pricing | One-time commissioning mistaken for recurrence |
| Revenue to owner cash | Margins after the people, parts and capital service needs | Support costs, capitalized development, bought growth |
| Cash to an investor's return | Growth against the price already paid | A generous starting cash or terminal value |
Adapted from the draft behind this note.
This note asks Vista's usual question of the chain. Not what the shares are worth, which nobody can settle, but what their prices on October 2 already assume.
Adrian Slywotzky and David Morrison gave the second half of the thesis a name in The Profit Zone (1997): installed base profit. A company earns modestly on the machine and well on everything the machine keeps needing; among its practitioners the books name Otis Elevator and Gillette. Whether a company is winning at it comes down to one ratio:
Here F is follow-on revenue (service, parts, upgrades and software sold into the base), N is the equipment in use, and a is what each unit in use brings in a year. A base that grows while a falls is leaking to someone else.
The model has a second condition, and it is where Schneider Electric, the closest fit of the three, departs from the textbook: the follow-on business has to carry the profit. Schneider does not report profit by business model. It does report that products and systems were 82% of its second-quarter revenue, and that Energy Management, which sells most of that equipment, earned 22.4% of its revenue as adjusted operating profit in the first half. This is not a company that gives away the machine to sell the blades. It is an equipment business, and a very good one, with an installed-base ambition.
The ambition is on the record, which gives us a scoreboard. Schneider wants Software and Services, 19% of its revenue in 2025, to reach about 25% by 2030, and it aims to double the weight of recurring revenue. For the share to climb from 19% to 25% in five years, Software and Services must outgrow the group by a factor of (25/19)1/5, about 1.056 a year: 13.0% a year if the group grows 7%, 16.2% if it grows 10%. In the second quarter they grew 6% while the group grew 16.5%, and they were 18% of revenue. Field services grew 5%, led in Energy Management by installation and commissioning, which follow new shipments rather than the base.
On its own scoreboard, Schneider is not winning yet. But shipments are a flow and the installed base is a stock:
E is equipment shipped this year, R is equipment retired, and N is the equipment in use. Electrical equipment often runs for decades, so one year's shipments are a small part of the base. As an illustration, not a company figure: if equipment lasted twenty years and shipments had grown 5% a year before the boom, a jump to 18% growth would lift the growth of the base only from 5.0% to 6.0%. Services growing 5% to 6% are about what a steady revenue per unit would produce. The base is not visibly leaking. Nor is it yet paying the annuity the thesis needs.
Siemens reads the same way. In the April to June quarter, the service business of Smart Infrastructure brought in €1.256 billion, 19.7% of the segment's revenue against 20.8% a year earlier. It grew 7% on a comparable basis while the segment grew 13%, and the segment's orders, at €8.0 billion, ran at 1.25 times its revenue. The base is being built faster than it is being serviced, which is what a boom looks like from the inside.
So the model fits Schneider's plan better than its accounts. On the company's own scoreboard it is not yet winning; on the arithmetic of stocks and flows it is far too early to say it is losing. The scoring starts when this decade's shipments come off warranty and either stay with their makers or drift away.
This reminds me of the story in the Talmud of the sage Honi, who came upon a man planting a carob tree and asked how long it would be before the tree bore fruit. Seventy years, the man answered (Ta'anit 23a). An installed base is a carob tree. The equipment shipped in this boom bears its fruit as service over decades, and bears it for its maker only if the customer keeps coming back. The share prices are less patient than the planter. They need the cash to compound from next year.
Prysmian's machine is different. A cable, once laid, asks little of its maker, and the draft is right that a cable company does not inherit software economics because its customers build artificial intelligence. Prysmian earns by building capacity and filling it, which makes it a creature of an old and durable model, Mordecai Ezekiel's cobweb (1938):
Here q(t) is the capacity in service at time t, S is how much the industry builds in response to price, p is the margin that justified the building, and L is the lag between the decision and qualified output. Prysmian's own numbers show where p stands: in the second quarter Digital Solutions earned an adjusted EBITDA margin of 23.8%, against 16.8% a year earlier, and Transmission 21.2%, against 17.1%. Its new fiber program runs to 2031. Capacity approved at this year's margins will meet whatever margins prevail when it arrives, and every competitor is reading the same margins.
One feature of this cycle is new. On July 20 Prysmian signed an agreement with Molex, a Koch company, worth up to €5.5 billion over a period of up to ten years and supported by a €550 million upfront payment, a tenth of the ceiling. A customer willing to pay before the cable is made suggests that capacity, not demand, is the scarce thing. It is also lending: an advance is a loan repaid in cable, from capacity that still has to be built and filled.
Siemens, finally, is several machines under one roof: factory automation and industrial software, the electrical businesses of Smart Infrastructure, trains, a finance arm, and 69% of a listed medical-technology company at its last year-end. What a Siemens share owns is a sum:
Vind is the value of the industrial businesses and Dind their net debt and other claims; ESFS is the equity in Siemens Financial Services; s is Siemens's share of Siemens Healthineers and ESHL that company's equity. The model in this note values Siemens's consolidated cash instead, which is why it gives two answers for Siemens and why neither is the last word.
The draft asks one question of all three companies, in the manner of expectations investing (Mauboussin, 2006): how fast must each company's cash grow before a buyer at the October 2 close earns 10% a year? It takes a year of cash, grows it at a constant rate for nine years, lets it grow 3% a year forever after, and solves:
P is the October 2 close and N the shares, so P × N is the equity; C1 is the first year's cash and C10 the tenth year's; g is the growth the price requires in Years 2 to 10; r is the annual return, 10% in the main case; gT is the growth after Year 10, 3%. Vista does not publish what a share is worth. What a price requires is a question the evidence can answer.
| The case | Year-1 cash, €bn | Equity, €bn | Growth the price requires |
|---|---|---|---|
| Schneider Electric, trailing free cash flow | 5.792 | 175.3 | 14.56% |
| Schneider Electric, same cash, shares net of suspended votes | 5.792 | 170.7 | 14.15% |
| Schneider Electric, trailing net income converted at the company's own 100% target | 4.738 | 175.3 | 17.64% |
| Siemens, reported free cash flow | 11.848 | 213.7 | 6.62% |
| Siemens, free cash flow after interest paid | 10.118 | 213.7 | 9.05% |
| Prysmian, 2026 guidance midpoint | 1.700 | 38.8 | 10.26% |
| Prysmian, guidance midpoint less the Molex advance | 1.150 | 38.8 | 16.24% |
| Prysmian, free cash flow in the twelve months to June | 0.978 | 38.8 | 18.73% |
Growth a year in Years 2 to 10 at which each case's cash, discounted at 10% a year with 3% growth after Year 10, equals the equity at the October 2 close. Prysmian's equity is net of its June treasury shares; on all issued shares (€40.1 billion) the last two rows require 16.72% and 19.21%. Inputs and sources are at the end of this note.
Turn the same model around and it says what each price offers a year on a growth path you choose:
| If Year-1 cash grows a year at | 6% | 9% | 12% | 15% |
|---|---|---|---|---|
| Schneider Electric, trailing free cash flow | 7.1% | 8.0% | 9.0% | 10.2% |
| Schneider Electric, net income at 100% conversion | 6.4% | 7.2% | 8.0% | 9.0% |
| Siemens, reported free cash flow | 9.7% | 11.1% | 12.5% | 14.0% |
| Siemens, after interest paid | 8.8% | 10.0% | 11.3% | 12.7% |
| Prysmian, guidance midpoint | 8.4% | 9.5% | 10.7% | 12.1% |
| Prysmian, guidance less the advance | 6.7% | 7.5% | 8.5% | 9.5% |
| Prysmian, twelve months to June | 6.2% | 6.9% | 7.7% | 8.7% |
What each October 2 price offers a year: the discount rate at which each case's cash, growing at the rate shown in Years 2 to 10 and 3% after, equals the equity in the table above.
Five things stand out.
Schneider's price asks for more than Schneider's own plan. The company's targets for 2026 to 2030 are organic revenue growth of 7% to 10% a year and 250 basis points more adjusted operating margin. On 2025's margin of 18.7%, that is profit growth of 9.7% to 12.8% a year. Hold the top of that range for nine years instead of five and €303.00 offers 9.3% a year; let it fade after 2030 to 3% by 2036 and it offers 8.1%. For the price to offer 10% on the top of the plan, the first year's cash would have to be €6.50 billion, about 12% more than the trailing €5.79 billion.
And the starting cash is generous. The trailing €5.79 billion is 122% of the same twelve months' net income, against the 111% the company converted in 2025 and the "around 100%" it targets for 2026 to 2030. Converting trailing net income at that target gives €4.74 billion, on which the price requires 17.6% a year.
Siemens asks the least, and its definition of cash matters most. On reported free cash flow €276.10 requires 6.6% a year; after interest paid, 9.1%. Siemens targets high-single-digit growth in earnings per share, and at 9% the price offers 10.0% a year after interest and 11.1% before it. But Siemens's free cash flow includes interest received (€1.82 billion in the first nine months of its fiscal year) while leaving out the €1.50 billion its finance arm added to its loans in fiscal 2025, which sits in investing, and the interest Siemens pays (€1.14 billion in nine months), which sits in financing. It also includes all of Siemens Healthineers. Until the cash is separated by perimeter, the low number measures the definition, not cheapness.
Prysmian's price depends on which cash you believe. On guidance the price requires 10.3% a year; without the advance, 16.2%; on what the company actually generated in the twelve months to June, 18.7%. The difference between the first two rows is a single payment that will not recur. At 12% growth on cash without the advance, the price offers 8.5% a year.
Most of every price lies beyond 2036. At the growth each price requires, the cash after Year 10 accounts for 64% of Schneider's equity, 59% of Siemens's (after interest) and 65% of Prysmian's (without the advance). The next ten years pay for roughly a third to two-fifths of each price; the rest is a bet on the decades after.
The record is not kind to the growth these prices need. Chan, Karceski and Lakonishok (2003) found sustained high growth in earnings rare, and no more common than chance would produce; their evidence is on earnings rather than cash, but 14% to 19% a year for nine years is the kind of record they seldom found. Order backlogs, of which these companies have plenty (€25.4 billion at Schneider at the end of 2025, €132 billion at Siemens in June), do predict earnings, yet Rajgopal, Shevlin and Venkatachalam (2003) found that investors overweight them. The shares of companies that invest most heavily and grow their assets fastest have tended to return less afterward (Titman, Wei and Xie, 2004; Cooper, Gulen and Schill, 2008), a caution for a cable maker adding a fiber program and an acquisition in the same year. And part of the discount at which conglomerates trade is an artifact of acquisition accounting (Custódio, 2014), a reminder that Siemens's sum of the parts has to be measured, not assumed.
The strongest case for these prices does not need the annuity at all. Scarce capacity, pricing and productivity can compound cash for longer than a cautious screen allows: Schneider's adjusted operating margin rose 120 basis points organically in the first half, and Smart Infrastructure's growth outlook for fiscal 2026 has gone from 6% to 9% at the start of the year to 10% to 11%. A persuasive version of that case has to put numbers on those drivers and on the reinvestment they need; it cannot simply lower the discount rate or raise the value after Year 10. On the arithmetic above, Siemens needs the least of it.
The first half was a record by every measure the company reports: revenue of €21.2 billion, up 14.0% organically; an adjusted operating margin of 19.3%; and free cash flow of €1.63 billion, against €474 million a year earlier. The comparison flatters, and the company says why: the first half of 2025 carried a €207 million fine, and without it the rise is about 140% rather than 244%. In July it raised its 2026 target to organic revenue growth of 10% to 13%.
The bridge from that cash to a shareholder is not finished, and the draft says so plainly. Net debt was €15.36 billion at June 30, up from €13.72 billion in December, after a €2.4 billion dividend and €0.6 billion for the share buyback. Lease liabilities (€1.55 billion at the end of 2025), share-based pay (€167 million in the first half), pension contributions and dividends to minority holders (€48 million) are claims to reconcile once, not to deduct twice. Ownership changed too: in December 2025 Schneider paid about €5.5 billion for the 35% of its Indian subsidiary it did not own.
The company is also buying its way further into software and connected devices: Cognite, an industrial data and AI company with more than $170 million of revenue in 2025, for $3.1 billion in cash; about 90% of AiDASH at an implied enterprise value of $350 million, both awaiting approvals at June 30; and, agreed in September, an intended offer of €70.00 a share in cash for Shelly Group, about €1.2 billion. On September 28 Shelly announced that the offer had been registered with Bulgaria's Financial Supervision Commission. It still needed the commission's approval, acceptance would open only when the approved offer was published, and the 95% minimum acceptance threshold was unchanged; closing is expected by the first quarter of 2027. None of these businesses is in the trailing cash, and none of their financing is either.
Siemens reported a record third quarter, with orders up 14% on a comparable basis. Its financial framework, reset at the start of fiscal 2026 and excluding Healthineers, targets comparable revenue growth of 6% to 9% a year and high-single-digit annual growth in earnings per share before purchase accounting.
The cash still needs its perimeter drawn. Industrial net debt was €9.3 billion at June 30 and total financial debt €49.9 billion; the two describe different perimeters and cannot be swapped. Minority shareholders received €508 million in dividends in the first nine months, mainly from Siemens Healthineers if the prior year is a guide. In fiscal 2025 Siemens paid about €9.3 billion for Altair and €4.2 billion for Dotmatics.
The largest change is coming. Siemens plans a shareholder vote on a direct spin-off of Siemens Healthineers shares to its own shareholders at its annual meeting in February 2027 (its calendar shows February 11, to be confirmed), and in August it said it had received the binding tax decisions it needed to proceed. A spin-off changes what a Siemens share holds. It does not by itself create value, and taxes, retained interests and the costs left behind all count.
Prysmian had its best quarter yet: €730 million of adjusted EBITDA in the second quarter and €1.33 billion in the first half, on revenue of €11.24 billion. On July 30, ten days after the Molex agreement, it raised its 2026 free cash flow guidance to €1.65 billion to €1.75 billion, from €1.30 billion to €1.40 billion: €350 million more at the midpoint. None of the documents we read says whether the €550 million advance is in the new guidance. The draft treated it as included, and so do we, as a labeled assumption.
What Prysmian actually generated is steadier than its guidance: €978 million of free cash flow in the twelve months to June, against €979 million a year earlier. In the first half alone, free cash flow before acquisitions and disposals was −€390 million, as working capital absorbed €1.07 billion. Net debt was €4.08 billion at June 30, including €402 million of leases; a €1 billion hybrid bond counts as equity, and its interest (€13 million paid and €39 million accrued over the twelve months) sits below the free cash flow line.
Then there is Atkore. In August Prysmian agreed to buy the American maker of electrical conduit and cable trays for $95 a share in cash, an enterprise value of about $3.8 billion (€3.3 billion) and 9.8 times Atkore's 2025 EBITDA. The U.S. antitrust waiting period expired on September 14; Atkore's shareholders and other regulators must still approve, and closing is targeted by the end of 2026. On September 11 Prysmian sold 7,024,793 new shares at €121.00, about €850 million, as the equity part of the financing, taking its share count to 308,861,615. The equity in this note includes those shares, but the cash, like the 2026 guidance, excludes Atkore; leaving the new shares out would lower the growth required on cash without the advance from 16.2% to 15.9%. The mismatch is small, and the rest of the financing is still to come.
This reminds me of Brutus in Shakespeare's Julius Caesar, arguing for the march to Philippi: "There is a tide in the affairs of men / Which, taken at the flood, leads on to fortune." Every capacity cycle has its tide speech, and the lesson of the cobweb is that every rival hears it at the same moment. Prysmian is taking the flood twice, with a €1.25 billion fiber program that runs to 2031 and a €3.3 billion acquisition, and one customer has agreed to pay €550 million up front to sail with it. Whether the tide is still high when the new capacity is ready is a question for the margins of 2030, not those of 2026.
Likely, we can get some of it from primary research, which is what Vista is all about. The filings will report, a quarter at a time, whether the service annuity arrives. People who ran service businesses, critical facilities and cable plants through earlier build-outs can say now how such cycles usually unfold.
We would askAs a matter of general past practice, what share of equipment commissioned in a given year was under a paid service agreement three years later, and what did that agreement bring in each year compared with the original sale?
The answer that would change the view"Small, and falling, as customers brought maintenance in house."
We would askOnce a hall was energized and the warranty ran out, who maintained the switchgear and power systems: the maker, an independent firm, or your own staff? What decided it?
The answer that would change the view"Everything went multi-vendor after year two, and the maker's software was optional."
We would askWhich makers' equipment could independents fully service, and what kept them out: parts, firmware, software licenses or warranty terms? Did connected equipment make that harder or easier?
The answer that would change the view"Easier. Connected equipment opened the data to everyone."
We would askIn past expansions of preform, fiber and cable capacity, how long did it take from approval to qualified output, and what did yields and costs look like in the first two years?
The answer that would change the view"Ramp-up and customer qualification took twice as long as the plan."
We would askIn general practice, what does an upfront payment buy a customer under a long-term supply agreement (reserved capacity, a price, priority), and what happened to such agreements when demand fell short?
The answer that would change the view"The advance was credited against the price, and the supplier carried the volume risk."
For Siemens the first work is public accounting: separating its cash by perimeter needs filings, not interviews. Conversations follow only if Smart Infrastructure's service economics turn out to decide the answer, and then the first two roles above apply.
Every conversation would draw on past, general experience and public examples only: no current employees of Schneider Electric, Siemens or Prysmian or of their customers and suppliers, nothing confidential, and every Advisor screened for conflicts before a word is said. Before each call we would write down which answer moves which input in the model, and how far it moves what the price offers; a question that no plausible answer could act on is not worth asking (the value of information, Howard, 1966). For a client engagement, Vista would combine this note with those interviews: two or three Advisors for each question that matters, structured conversations, and one written brief that says where they agree, where they split, and what would change the answer.
We will score this note after Schneider Electric and Prysmian report on October 29, 2026 and Siemens on November 12, 2026, the dates on each company's own financial calendar, and publish the result on the scorecard whether it flatters us or not. The tests are fixed today, so they cannot drift:
Four or five passes, and the first evidence sides with the prices. Two or fewer, and the distance between what the prices need and what the businesses show has widened. Three is too early to say. We will also report what each price then offers on each case.
For now the boom is in the accounts and the annuity is in the plans. Schneider's price leans on the plans, Siemens's on whose cash it is, and Prysmian's on which cash you count. By mid-November the tests above will say which way the evidence is leaning. The decision, as always, belongs to the reader.
Each case starts from one year of cash, taken as the cash of Year 1, the twelve months from October 3, 2026 to October 2, 2027, received at its end. That cash grows at a constant rate over the nine intervals to Year 10, and Year-10 cash grows at 3% a year forever after; the terminal value is discounted ten years. What a price requires is the growth at which the cash, discounted at 10% a year, equals the equity at the October 2 close; what a price offers is the discount rate at which a stated growth path does. No debt is subtracted, because the cash bases already sit after interest, except Siemens's first case, which is shown both ways. No acquisitions, buybacks or share issues are modeled beyond those already in the share counts. Schneider's plan is translated into profit growth as revenue growth at the plan rate with the margin rising from 2025's 18.73% to 21.23% evenly over five years; the faded case holds that rate through Year 4 and steps down to 3% by Year 10. The model grows cash directly, so it cannot say what return each case implies on new capital: a limit of the method, stated rather than hidden. The draft's figures were rebuilt in code and reproduced to the second decimal before anything was turned around; every figure in this note was recomputed in code from these inputs.
| Input | Schneider Electric | Siemens | Prysmian |
|---|---|---|---|
| Close, October 2, 2026 | €303.00 | €276.10 | €129.80 |
| Shares | 578,605,363 issued; 563,428,459 net of suspended votes | 774.05m, third-quarter weighted average | 308,861,615 issued; 299,274,068 net of June treasury |
| Equity at the close (€bn) | 175.317; 170.719 | 213.715 | 40.090; 38.846 |
| Year-1 cash (€bn) | 4.635 + 1.631 − 0.474 = 5.792 | 10.812 + 6.541 − 5.505 = 11.848 | 1.700, guidance midpoint |
| Alternative cash (€bn) | 4.163 + 2.488 − 1.913 = 4.738 | 11.848 − 1.730 = 10.118 | 1.700 − 0.550 = 1.150; 0.978 |
| Return; growth after Year 10 | 10%; 3% | 10%; 3% | 10%; 3% |
Schneider's cash is its reported free cash flow for 2025 plus the first half of 2026 less the first half of 2025; its alternative is net income on the same basis. Its suspended votes (15,176,904 at August 31) are close to the 14,794,926 treasury shares it held at June 30. Siemens's interest paid is 1.733 + 1.144 − 1.147 = 1.730. Prysmian's €0.978 billion is its reported free cash flow for the twelve months to June 30: 2.099 − 0.210 − 0.703 − 0.216 + 0.008 = 0.978. Which hurdle is right is the reader's decision, so here is the growth each main case requires at four of them:
| Growth required, Years 2 to 10, at an annual return of | 8% | 9% | 10% | 11% |
|---|---|---|---|---|
| Schneider Electric, trailing free cash flow | 8.95% | 11.88% | 14.56% | 17.05% |
| Siemens, after interest paid | 3.79% | 6.54% | 9.05% | 11.39% |
| Prysmian, guidance less the advance | 10.52% | 13.51% | 16.24% | 18.77% |
This note is research, not investment advice. It states what market prices assume under labeled assumptions; it is not a recommendation to buy, sell or hold any security, and the decision belongs to the reader. As of October 3, 2026, Russ Rosenzweig, Vista's founder, owns no shares of Schneider Electric, Siemens or Prysmian. The note began as Russ Rosenzweig's research question and draft, which the Vista Research Desk challenged over several rounds, with corrections in both directions; the model and every figure were then rebuilt in code and checked against the companies' own reports and releases.