Accenture plc ACN

Key data

The business

Accenture sells the people and the know-how to plan, build and run other companies' technology. Two engines drive it: consulting, the shorter advisory and build work, and managed services, running a client's systems for years under contract. It fields about 799,000 people across the Americas (49% of revenue last quarter), EMEA (37%) and Asia Pacific (14%). Its customers are the largest enterprises and governments on earth. The moat is switching costs stacked on scale: once Accenture runs your payments backbone or your supply-chain systems, tearing that out is slow, expensive and risky, and only a handful of firms can staff a global transformation at all. What the client actually holds is a signed multi-year contract and a team of badged consultants sitting in its own offices.

The numbers

Revenue growth has cooled from a boom to a walk, while profit and cash have held.

The latest quarter grew 5.6% on revenue and 8.9% on EPS against a year earlier, and EPS came in a touch above the Street each of the last four prints. The Q4 dip is seasonal, not a stumble. The longer arc matters more.

Revenue compounded about 8% a year over four years and EPS about 7%, but the growth is front-loaded: up 22% in FY2022, then 4%, 1% and 7%. The deceleration is the whole story. Against that, the cash is untouched: $10.9B of free cash flow in FY2025, a 9.5% free-cash-flow yield at today's price, on a net-cash balance sheet. Consensus models roughly 5% annual revenue growth out to FY2028. The variant view here is that the slowdown is a cyclical air pocket, delayed discretionary projects plus an AI-transition pause, not structural erosion, and the backlog is the tell.

Booked-but-undelivered work is building faster than reported revenue, which usually leads growth up. One caveat: the generative-AI bookings figure management once trumpeted has now been absent for a fourth straight release, so the cleanest read on AI demand is gone, exactly as flagged last time and still unresolved.

Management

Julie Sweet runs it. Insiders sold about $2.6M across ten open-market sales over twelve months and bought nothing; at that size, against a $114B company, it is routine and plan status is not disclosed, so read nothing into it either way. The real capital signal is the buyback. Accenture lifted its FY2026 repurchase program to $7.5B, a 62% jump from the roughly $4.6B it spent in FY2025, and said on the record that the share price does not reflect the company's financial strength or long-term growth. That is a rare, dated, specific statement, funded out of $10.9B of free cash flow rather than debt. A $308M severance charge in Q1 FY2026 closed a six-month cost program. Pay tracks earnings and the buyback is accretive at 15x. This is a disciplined operator putting cash behind its own words.

How it fails or surprises you

AI deflates the billable hour. Accenture bills for people. If clients use AI to need fewer consultant-hours per project, the core compresses. Watch organic local-currency growth against the 799,000 headcount: revenue per head falling while growth sticks near 6% means the machine is eating itself. The now-absent AI-bookings line would have shown it first.

The value-trap case. The market set a twelve-year-low multiple on 6% growth deliberately, pricing a slide toward flat. RPO at $38B argues against it, but managed-services backlog converts slowly and can mask soft consulting underneath. A single quarter of negative organic growth turns 15x from cheap into merely fair.

AI reaccelerates demand (right tail). Accenture is the largest place enterprises go to actually deploy AI at scale. If deferred projects convert and organic growth returns to double digits, a 15x multiple on a reaccelerating compounder re-rates hard. The $7.5B buyback, up 62% with management calling the stock mispriced, is insiders betting on exactly this. The print: bookings growth and one double-digit organic quarter.

Closing thoughts

Organic local-currency growth and bookings over the next two to three quarters resolve the bet. Both cases, AI erodes the labor model and AI reignites demand, run through those same two lines. The market has already marked the bear a long way, paying the lowest multiple in twelve years for a business still throwing off $10.9B of cash and growing, however slowly. That is where the asymmetry sits: modest downside if growth grinds flat, because the cash and the buyback set a floor, and a sharp re-rate if the $38B backlog converts. On this evidence the fatter tail points up, but only if backlog turns into faster revenue; if it does not, this is dead money with a dividend and a buyback, not a loss.

The bet is still that big companies keep hiring Accenture to do their technology work, and now to install AI across it, and that AI adds more work than it destroys. What breaks it is the pair to watch every quarter: organic local-currency growth against the $38 billion of booked work. If growth keeps sliding while that backlog stalls, the read is wrong and the cheapness was earned. If growth turns up while the backlog builds, you were paid to wait.