Priced for no growth, and still growing

On the close of Friday 25 September 2026, Accenture traded at $176.11. Sperio values a business from its filed accounts and a discount rate, giving no credit of any kind to future growth, and arrives at $184.92.

The market is paying 95 cents for each dollar the accounts support without growth. That is barely a discount in the sense the last two pieces used the word, and it is nowhere near the twenty widest gaps in the index this week. It is the eighth cheapest of the 100 companies that clear our two gates, and the second on that list with no blocking flag — and it is not the name our machine put first. Why we publish it is most of this article, and the short version is this: a company whose revenue has risen every year for five years is priced as if it will never grow again, four days before it reports.

Why not the widest discount, and why not the machine's pick

The widest discount in the index this week belongs, again, to a fertiliser company whose net income is down 85%. Behind it sit a laboratory landlord and an office landlord whose net income is down by two thirds. Ranking by discount answers "what looks cheap", and it answers it reliably with whichever business is deteriorating fastest, because a collapsing denominator makes any price look like a bargain.

So the ranking applies two hard gates first. Does the business durably earn more on its capital than the capital costs — return on invested capital clearing ten percent in four fifths of the readable fiscal years and in each of the last three. And are both lines going the right way — revenue and operating income each compounding upward over the full readable history. Of the 453 companies the board has valued, 100 clear both.

The survivors are then ordered by cheapness: enterprise value to the median operating profit of up to fifteen fiscal years. That ordering is the only one in this project with any evidence behind it — a small backtest, an upper bound, described two weeks ago and not repeated here. The list this morning, cheapest first:

A. O. Smith was the subject two weeks ago, and the machine put it first last Sunday and again this morning. The honest reading of that is still "nothing has changed", and the scoring section at the end reports it. The editor went past it, and past a name blocked by a flag we argued with last week, to the eighth line: the second company on the list with no blocking flag, and — we measured this, we did not guess it — the one whose shares the public is actually asking about this month. That is an editorial choice, made for stated reasons, and the machine's own pick is printed next to it so that a reader can disagree.

Expensive on its own past, cheap on its last twelve months

Eighteen point three times is not cheap, and it is worth being exact about what the number is. The ranking divides today's enterprise value by the median operating profit of the last fifteen fiscal years. For a business that has tripled its operating profit over those fifteen years, the median is a profit it last earned around 2018, when it was a company half this size. On that measure Accenture reads expensive. Against the last twelve months of filed operating profit, the same enterprise value — about $103bn, after $5bn of net cash — is 9.8 times. Before the charges discussed below, 9.0.

Neither number is wrong. The full-cycle multiple is the frame's deliberate suspicion of any company's best recent years, and it is the ordering the backtest tested. It simply penalises growers, by construction, and a reader comparing this 18.3× to a screener's trailing multiple would be comparing two different things. We say which one we are using each time, and this piece uses both.

What it earns

Return on invested capital, on the ranking's measure: above fifty percent, which is what an asset-light business earns when its capital is mostly people. Return on equity 24.4%. A gross margin of 32% that has not moved in four years. And $10.2bn of cash against $5.1bn of debt at the last balance sheet date. The line that matters most for a valuation built on cash: in fiscal 2025 the company reported $7.68bn of net income and $10.87bn of free cash flow. Over the twelve months to May 2026, free cash flow was $12.58bn — about 11.6% of what the whole company costs at Friday's price.

The accounts, five fiscal years and the trailing twelve months, as filed with the SEC (fiscal years end 31 August):

Revenue is up 38% over the five years, a record in each of them. Operating income as filed is up 34%, net income 30%, earnings per share 33% — with the share count down 2%. Over the ranking's own four-year window, fiscal 2022 to fiscal 2025, net income is up 11.7% and shares are down 1.6%: the rise in earnings per share is earnings, not the buyback, which is the opposite of what the screen found on a health insurer three weeks ago. The buyback has since sped up — the share count fell 2.4% over the year to May — and the board's report notes a new $2bn authorisation, which is a reason to watch the per-share line more carefully from here, not a reason to credit it.

And the latest quarter, March to May 2026, against the same quarter a year earlier: revenue +5.6%, operating income +6.5%, earnings per share +8.9%. The business was still growing while the stock halved.

The charges that come every year

One line in that table deserves its own paragraph. Since fiscal 2023 the company has booked what it calls business optimization costs — severance, mostly — of $1.06bn, then $0.44bn, then $0.62bn, and $0.92bn over the last twelve months. Its own adjusted figures strip them out. Four years running, they are not one-offs, and we leave them in: the operating income in every figure above is as filed, charges included. Over the ranking's four-year window that line is up 9.2%; before the charges it would be up 15.7%. Take the filed line. A company that has paid to shrink its workforce in each of the last four years is telling you something about the business, and the adjusted figure is designed to make you not hear it.

What the price is saying

In February 2025 the market paid about $398 for a share of this company. In January 2026, $289. Then two legs down: −28% in the month of February, and −18% on 18 June, the day of the third-quarter results, from $156 to $128. A low of $124.44 at the end of June. On Friday, $176.11 — fifty-six percent below the high of last year, forty-two percent above the low of this summer. In between, the accounts printed the numbers in the table above: a record fiscal year, and a quarter that grew.

The market's case has a name, and it is artificial intelligence. If machines write the code and run the systems, a company that sells the hours of people who do that work will sell fewer of them, for less. The accounts hand the argument some support: revenue is growing at five and a half percent, where the last fiscal year grew at seven; and the restructuring charges have come four years running. A business that pays to shrink in the name of efficiency is making the market's argument and its own at the same time — the market says the work is going to machines, and the company says it intends to be the one that owns the machines. Last week's piece said the same of Cognizant. It is the argument of the whole industry this year, and no accounting line settles it.

The downside, priced — and a correction to our own table

A single valuation cannot be argued with. Three can.

The price sits just under the middle rung. Everything above it — the whole of the upside our frame can see — is the credit the frame gives a record of high returns on capital, and that credit is exactly what the market is disputing. The trough case rescales the no-growth value to the worst readable year of operating profit, about three fifths of the fifteen-year median, and it sits two fifths below the price: a buyer wrong about everything except the trough is down about 40%, not protected. We say that plainly because our own Sunday table does not.

Here is the correction. The table our machine prints each Sunday shows this company's middle rung at $295.87 and labels it no growth, and shows a credited value of $474.87 above it. The $295.87 is the board's headline valuation — and for this company the board's headline already carries the franchise credit: the no-growth models come out between $184.92 and $240.83, and the headline is the lower of those two times the same factor of 1.60. The ranking then applied the factor a second time. The figures in the table above are the corrected ones; the run's "trough above the price" for Accenture on 20 September, which briefly made it one of five names with a cushion, was the same error and is withdrawn. The last two companies we wrote about were not affected — their board valuations carried no credit. The fix to the table is ours to make and is not made yet; until it is, we check that line by hand for every name we publish. A frame that gives no credit to growth has to be able to say, for each number, whether that number is credited. On this one, it did not.

The rate sweep

At the cost of capital the ranking used this morning — 10.18%, built on the 5.18% ten-year Treasury as FRED reported it on 25 September — 100 companies clear both gates. Raise the cost of capital by a hundred basis points and 86 still clear; at two hundred, 77. Accenture clears at every rate. What the sweep also counts — how many names keep a trough above their price — is the line the correction above touches, and we do not use it for this company: on the corrected figures Accenture has no such cushion at any rate. On the uncorrected ones four names keep one today — one of them, Paychex, through the same double credit — and none keeps one at +100 basis points.

The case against, in the board's own words

Sperio runs a devil's advocate whose only job is to attack the thesis. From the board's report dated 25 September 2026, reproduced without softening:

Significant competitive disruption risk from AI-driven coding tools threatening premium margins, alongside vulnerability to cyclical contractions in client discretionary IT spending.

And from the same report's bear case:

The rapid commoditization of AI-driven coding and consulting tools threatens to erode the premium margins of ACN's human-capital-intensive service model. Additionally, a cooling macroeconomic environment could lead clients to defer or cancel large-scale digital transformation projects.

That is a real argument, and the accounts cannot answer it. They describe what the business earned; they say nothing about what a client will pay a person to do next year that a machine did not do this year. What the accounts can say is narrower. At Friday's price a buyer pays for the earnings the company already makes, held flat forever, and gets the record for nothing. Whether the record is worth the $111 a share our frame credits it with — or nothing — is the whole disagreement between the market and this piece, and nothing in the filed figures can settle it.

Why now

Two things, and one of them is a warning rather than a reason. The market has cut this company's price by more than half on accounts that moved by single digits, and it has done so in two lurches — one of them a single day. That is a fact about the price, not a prediction about it.

The warning: the company reports its fourth quarter and full fiscal year on Thursday 1 October 2026, before the market opens — four days after this is published. Every figure in this piece except the price will be superseded that morning. Consensus for the quarter is $3.18 a share. The number to read is not the beat or the miss. It is revenue growth — whether it is still five to six percent — and what the company says about the year ahead. We would rather publish four days early and say so than pretend the date is not there. Next Sunday's piece will re-read this one against the new accounts.

Last week's pick, scored

Last Sunday's piece said that this one would report whether Cognizant still clears the gates, and score it out loud.

Its shares closed Friday at $57.33 against the $59.87 of the previous Friday, down 4.2% in a week — a number that says nothing about the thesis and is reported because it was promised. It now sits first on the list above, still blocked by the two flags we argued with, and it still clears both gates this morning. A. O. Smith, the pick before it, closed at $58.65 against $56.95 the week before, up 3.0%, and is the machine's pick for the second Sunday running. Neither number means anything at a week; both are recorded so that in a year they can.

What this is, and is not

Every figure here comes from accounts filed with the SEC — the 10-K for the fiscal year to 31 August 2025, filed on 10 October 2025, and the 10-Q for the quarter to 31 May 2026, filed on 18 June 2026 — from a board valuation dated 27 September 2026 (the devil's advocate and bear case quoted above are from the board's report of 25 September), and from a ranking run dated 27 September 2026. The price used is the close of Friday 25 September 2026. The cost of capital is built from FRED observations. One note on the company page on this site: for fiscal 2023 to 2025 it currently shows the company's adjusted operating income rather than the figure as filed, a data-source mix we are correcting; the table above uses the 10-K and the 10-Q.

The valuation gives no credit to growth beyond what the franchise record earns, and the ranking that produced this list orders the survivors of two gates by their full-cycle operating multiple — a way of comparing them to each other, tested on a small sample, not a forecast of what any of them will do. This week the editor went past the machine's pick, for reasons stated above, and corrected a label in the machine's own output; both are reported so that a reader can disagree.

The board that produced these figures will run on any company you name. What it will not do is decide for you — nothing here is investment advice.