Computacenter told the market on 8 September 2026 that adjusted profit before tax for the full year would be "no less than £380 million" — roughly 11% above the £340.9m company-compiled analyst consensus it was upgrading away from. The interim numbers behind that guide were extraordinary: first-half revenue up 71.6% to £6.85bn and adjusted profit before tax up 87.0% to £152.4m. The shares touched a record 6,015p intraday and closed at around 5,825p, up roughly 4%, valuing the group at about £5.87bn.
An 11% guidance upgrade that moves a share price 4% is a market that had already half-priced the news and has moved on to a different question. That question is not whether Computacenter is growing. It is what kind of profit the growth is made of.
What the first half contained
Revenue of £6.85bn for six months compares with £3.99bn in the first half of 2025 and with £9.19bn for the whole of last year. Computacenter did three-quarters of a full prior year in a single half.
Almost all of the increment came from one line. Technology Sourcing revenue — the product resale business — rose to £5.91bn from £3.18bn, driven by demand from hyperscale and data-centre customers for AI infrastructure, concentrated in North America. Adjusted operating profit rose 86.5% to £153.1m and adjusted profit before tax rose 87.0% to £152.4m from £81.5m.
The forward indicator is the committed product order backlog, which reached a record £9.3bn at the half-year against £7.1bn at the end of December 2025, an increase of 29.5% in constant currency. To put that in proportion: the order book alone is now larger than the group's entire revenue for the whole of 2025.
Why a 4% move on an 11% upgrade
Three things explain the gap, and none of them is scepticism about the numbers.
The first is that the upgrade was partly anticipated. Computacenter had already raised expectations after a first-quarter beat earlier in the year, and the shares had run hard through the summer — more than 70% higher year-to-date by late August, before this week's gain. A stock that has already re-rated that far has, by definition, paid for a good deal of the good news in advance.
The second is the shape of the profit. Revenue grew 71.6% and adjusted operating profit grew 86.5%, which does mean operating leverage was positive — the adjusted operating margin went from about 2.06% in the first half of 2025 to about 2.24% this time, on Openbook's arithmetic. But the absolute level is the point. A 2.2% operating margin is a distribution margin, not a services margin. For comparison, Computacenter's full-year 2025 adjusted operating margin was closer to 3.0%, on £274.7m of adjusted operating profit against £9.19bn of revenue. Half-year and full-year margins are not directly comparable because of seasonality, so the honest half-on-half read is that margin improved slightly. The honest structural read is that the business mix is moving toward the thinnest-margin line in the group at extraordinary speed.
The third is customer concentration. "Hyperscale and data-centre customers in North America" is a small number of very large buyers. Backlog of £9.3bn built on that base is genuinely committed, but it is committed by a handful of counterparties whose own capital expenditure plans are the single variable that determines whether the backlog converts on schedule.
Set against that, the implied second half is demanding rather than conservative. Adjusted profit before tax of no less than £380m for the year, with £152.4m banked in the first half, implies at least £227.6m in the second — against roughly £190.5m in the second half of 2025, or growth of about 19.5%. That is a real acceleration commitment, not a hedge, and it is why the guide was worth making.
The Openbook read
Momentum is unambiguously the strongest of the five factors and this release extended it. A record intraday high, a close up around 4%, and a year-to-date gain that was already north of 70% by late August describe a stock in a sustained uptrend confirmed by fundamentals rather than by flow alone. The caveat is the standard one for momentum at this stage: the factor is strongest precisely when the reversal risk is highest, and a 4% response to an 11% upgrade is the first mild evidence that the marginal buyer is becoming harder to find.
Growth scores about as well as it is possible to score, and the quality question sits inside it rather than beside it. Revenue up 71.6%, adjusted PBT up 87.0%, backlog up 29.5% in constant currency to a record £9.3bn, and a full-year guide implying adjusted PBT growth of nearly 40% on 2025's £272.0m. What tempers it is durability: this is order-book growth from an AI infrastructure build-out, and build-outs have a shape. Backlog conversion is visible for a year or more, which is more than most businesses have, but it is not an annuity. The growth score is exceptional and should be read as cyclical-exceptional, not structural-exceptional.
Profitability is the factor the market is genuinely arguing about, and it deserves precision rather than a slogan. On the reported half-year, profitability improved: adjusted operating margin rose roughly 18 basis points half-on-half. On the mix, the risk is real: Technology Sourcing went from about 80% of half-year revenue to about 86%, and that line earns the least. The reconciliation is that scale in product resale carries its own economics — bigger deals, better vendor terms, fixed-cost absorption — and this half shows those working. Profitability scored on returns rather than margins looks better still, because a distribution business consumes relatively little capital for the profit it generates. The number to watch is not the margin percentage but whether it holds as the mix keeps shifting.
Solvency is not stressed but it is doing more work than it was. Product distribution at this scale is a working-capital business: inventory and receivables expand with revenue, and revenue expanded 71.6%. Computacenter has historically run a conservative balance sheet with a net cash position — adjusted net funds stood at £606.0m at the end of 2025 — and nothing in the guide suggests that has changed materially, but a £9.3bn backlog converting through 2027 will absorb working capital in a way a £7.1bn backlog did not. This is the factor most likely to move on the next disclosure, and the least discussed.
Reward/Risk resolves the other four into a single, uncomfortable symmetry. At 5,825p against 2025 adjusted diluted earnings per share of 175.1p, the shares trade on roughly 33 times trailing adjusted earnings — Openbook's arithmetic — for a business that has historically been valued as a mid-single-digit-margin IT reseller. The forward multiple on a £380m adjusted PBT is materially lower, which is the bull case in one line: the earnings are arriving fast enough to grow into the rating. The bear case is that the same speed works in reverse. Because the profit is backlog-driven and the backlog is customer-concentrated, a deferral by one or two hyperscale buyers does not slow the growth rate — it removes a slice of it. A 33-times trailing multiple on a 2.2%-margin revenue base leaves very little tolerance for that, and the modest 4% reaction to a large upgrade suggests the market has already worked this out.
The read-across
The most direct comparators are the other UK-listed IT resellers. Softcat and Bytes Technology sit in an adjacent flow, though both are more weighted to software and services and to UK corporate and public-sector customers rather than North American hyperscalers. That difference is the point of the read-across rather than an obstacle to it: Computacenter's half demonstrates that the AI infrastructure build is reaching the channel, but it is reaching a specific part of it. Investors should not assume the same revenue acceleration in peers whose customer base is not building data centres.
The broader signal runs up the supply chain. A record £9.3bn committed product order backlog at a European reseller is a corroborating data point on hyperscaler capital expenditure plans for 2027, independent of what the hyperscalers themselves say on their own calls. Backlog at the distribution layer is a leading indicator that is harder to manage than a guidance statement, because it reflects orders customers have actually placed. For anyone tracking whether AI infrastructure spending is decelerating, this is a useful and under-followed series.
Computacenter joined the FTSE 100 at the June 2026 quarterly reshuffle, moving up from the FTSE 250 with a market value at the time of about £4.7bn; after Tuesday's move it is worth about £5.87bn. That makes it one of the blue-chip index's more direct AI-linked constituents, and it gives the FTSE 100 an exposure it did not have a year ago — a genuine sensitivity to hyperscaler capital expenditure, which UK investors have historically had to seek offshore. The Openbook screener is the place to see how the rest of the UK technology cohort has re-rated alongside it.
What to watch next
- Backlog conversion, not backlog size. The £9.3bn is a stock, not a flow. The number that matters at the full-year results is how much of it converted to revenue in the second half and how much remains — a backlog that grows because orders are being placed is very different from one that grows because orders are being deferred.
- The margin on the mix. Adjusted operating margin of about 2.24% in the half is the anchor. If the full-year figure holds near that level with Technology Sourcing still around 86% of revenue, the scale-economics argument is proven. If it slips, the mix concern is confirmed.
- Working capital and the net cash position. The full-year statement is the first look at what a 71.6% revenue increase did to inventory and receivables. This is the cleanest test of whether the growth is self-funding.
- Full-year results, due around March 2027. That release confirms whether adjusted PBT cleared the £380m floor and, more importantly, what management says about 2027. Guidance for the year after the build-out year is the disclosure that will determine how the market rates Computacenter from here.
One further point of context. Computacenter's 2025 was itself a record — revenue up 32.0% to £9.19bn, adjusted PBT up 7.1% to £272.0m, adjusted diluted EPS up 9.5% to 175.1p, with North America rising to 39% of group operating profit before central costs from 24%. The gap between 32% revenue growth and 7% profit growth in 2025, versus 71.6% and 87.0% in this half, is the clearest single measure of how much the operating leverage has improved. Whether that persists once the build-out normalises is the question the next two reporting periods exist to answer.

Discussion
Log in to join the discussion