Gartner, Inc. IT

Key data

The business

Gartner sells opinions on a subscription. Its Research arm, 77% of revenue, publishes the analyst reports and rankings, the Magic Quadrants, that enterprise technology buyers use to shortlist vendors and then defend those choices to their boards. A seat runs into the tens of thousands of dollars a year. The other two pieces are Conferences, the Symposium/Xpo events that pack into the fourth quarter, and a shrinking Consulting arm. The moat is workflow: once a company's IT buying runs through Gartner seats and inquiries, ripping them out means re-teaching the whole organization how it decides, so renewals historically sit near or above 100% of prior contract value. That stickiness is the entire story, and it is exactly what the market now doubts.

The numbers

The last five quarters show a business growing slowly at the top and fast at the bottom:

Ignore the Q3 2025 GAAP collapse to $0.47; a one-time charge did that, and adjusted earnings that quarter were $2.76. The real read is Q2 2026: net income up 14% year over year, diluted EPS up 33%, the gap bought entirely by repurchase. Every one of the last four quarters beat consensus adjusted EPS, the latest $4.37 against $3.76.

Across fiscal years the top line has crept while earnings whipsaw on one-time items:

Here is the compounding math. Revenue grows low single digits, net income high single to low double, and Gartner retired 14% of its diluted shares in a year, 77.4 million down to 66.6 million, spending about $2.0 billion in FY2025 and another $1.1 billion in the first half of 2026. Stack those and per-share earnings compound in the twenties on a nearly flat top line, so long as the research engine holds.

That engine is what the segments answer:

Research at +2% is the number that took the stock from $266 to $186, and deferred revenue slipped to $2.76B from $2.81B at year end, a soft forward signal. the market prices Research as permanently broken, but if the deceleration is cyclical, macro plus federal cuts, rather than the analyst business being hollowed out by AI, then 17 times earnings and a 14%-a-year buyback is mispriced. The print that settles it is contract value growth over the next two or three quarters.

Management

Capital allocation is the whole management story, and it is aggressive: roughly $2.0 billion of stock repurchased in 2025 at an average price well below today's, which is why stockholders' equity is now negative, a deficit created on purpose by buying back more than the company earns. At 17 times earnings that is the right use of cash if contract value holds and value destruction if it does not. Net debt sits near $1.5 billion, about 1.1 times EBITDA, comfortable. Insider activity is quiet and one-directional: five small sales over twelve months totaling $1.4 million, no buys, the largest a $0.6 million sale by Robin Kranich in August, plan status not disclosed. Nothing there reads as either alarm or endorsement.

How it fails or surprises you

US federal cutbacks. Washington is a meaningful slice of Gartner's contract value, and government efficiency cuts hit exactly the renewals that drive Research. If federal seats do not renew, Research growth crosses from +2% into negative, and the buyback cannot outrun a shrinking core. Watch quarterly Research revenue and contract value growth.

AI eats the analyst (the real left tail). The syndicated-research product, a human expert summarizing a market, is what generative AI does cheaply. If clients decide an LLM answers "which vendor" well enough, retention breaks at renewal and shows first in falling deferred revenue, already down slightly. This is the structural bear the multiple is pricing.

The re-rate (right tail). If Research reaccelerates to high single digits while Gartner keeps retiring 10%-plus of its shares a year, EPS compounds near 30% and the multiple has room to travel from 17x back toward its historical mid-20s. That is a double over a few years without heroic assumptions, and nobody is paying for it at today's price.

Closing thoughts

This one resolves on a number, not a philosophy. Gartner is priced as if its research franchise is in permanent decline, cheaper than at any point in twelve years, yet it beats estimates every quarter and shrinks its share count faster than almost any company its size. What settles the argument is contract value growth: stabilize it above mid-single digits and the stock re-rates, let it turn negative and the buyback is only delaying the reckoning. An ambiguous print, Research stuck at 1% to 3%, leaves a value trap where repurchases carry per-share earnings but the multiple never comes back. My read is the right tail is fatter than the price implies, because the AI erosion is slow and visible in deferred revenue long before it is fatal, and the valuation floor is already here.

The bet is still that big companies keep paying Gartner every year for its research subscriptions, and they keep renewing even as the US government cuts spending and ChatGPT-style tools tempt people to skip the human analyst. What breaks it is Research revenue growth turning negative while deferred revenue keeps falling. Hold those two green and the buyback does the rest. Watch them roll over together and no multiple is low enough.