CBRE Group, the world’s largest commercial real estate services firm, announced on October 20 that it has signed a definitive agreement to acquire Onnec Group Limited, a London-based provider of data-center infrastructure services, for £460 million (approximately $610 million) in cash up front, plus a multi-year earnout of up to £300 million (approximately $400 million) tied to growth thresholds. The total consideration, if the earnout pays in full, is roughly $1 billion. The deal is expected to close around the end of the year, subject to customary regulatory approvals — expected, not certain, and the approvals are the one formal step standing between announcement and ownership.
Onnec is not a developer, not a landlord, and not a power producer. Founded in 1986 and based in London, it designs, deploys, and maintains network, fiber, and data-center infrastructure for the world’s largest technology companies. Its approximately 800 employees serve hyperscalers across the United Kingdom, Continental Europe, and Asia Pacific. In plain terms, Onnec is the firm that shows up inside the building — cabling, connectivity, fit-out, and ongoing maintenance — after the shell is built and before the servers can do useful work. CBRE will fold it into its Critical Infrastructure Services line of business.
The bottleneck has moved
The investment case for this deal starts with a shift in where the data-center buildout actually jams. For the last several years, the binding constraints that investors talked about were capital and power: who would finance the next wave of campuses, and whether the grid could deliver the megawatts. Those constraints have not disappeared — interconnection queues in constrained regions still stretch for years, and the cost-allocation fights over who pays for grid upgrades are still live. But capital has, by and large, shown up. The hyperscalers are among the most creditworthy corporations in the world, and the contracted cash flows behind their load have drawn infrastructure funds, project-finance lenders, and utilities into the buildout, as this publication has covered in its data-center-power-as-an-asset-class series.
What has not scaled as easily is execution. A hyperscaler building in Virginia, Frankfurt, and Singapore at the same time does not need three different ways of installing a network. It needs one way, repeated identically, with the same standards, the same documentation, and the same handover discipline — and it needs the capacity delivered on the date the compute is scheduled to arrive. Research firm Dell’Oro Group put it directly in coverage of the deal: the bottleneck has moved from capital to execution, from getting capacity financed to getting it designed, deployed, and handed over on time, the same way in every market. That is the problem Onnec solves, and it is the problem CBRE just paid up to $1 billion to own more of.
This matters for power-sector investors because it reframes where the investable layer sits. The first phase of the data-center trade was the electrons: generation, PPAs, tariffs, and the contracted revenue streams behind them. The second phase, now underway, is the physical delivery layer: the firms that turn a powered shell into a working facility. Those firms do not sell megawatts. They sell certainty — that a hall will be ready on Tuesday, that the fiber plant will test clean, that the handover from construction to operations will not strand a quarter’s worth of compute. Certainty, in a buildout running at this speed, is a scarce product with a price.
Reading the price
The headline numbers deserve a careful read, because the structure of the deal tells you as much as the size. CBRE expects Onnec to generate approximately £370 million ($498 million) in revenue and more than £50 million ($67 million) of core EBITDA in 2027, and expects the acquisition to be immediately accretive to its core earnings per share. Measured against those expected figures — CBRE’s own forecasts, not realized results — the initial £460 million cash price is roughly 1.2 times expected 2027 revenue and roughly nine times expected core EBITDA. If the earnout pays in full, the total £760 million is roughly twice expected revenue and roughly fifteen times expected EBITDA.
That is a full price for a services business, and the earnout is doing real work in it. By making roughly 40% of the total consideration contingent on growth thresholds, CBRE is paying the base price for the business Onnec is today and reserving the premium for the business Onnec becomes if the buildout keeps running at its current pace. The structure shares the demand risk with the seller: if hyperscale deployment slows, the earnout does not pay, and CBRE’s effective multiple falls. If it accelerates, Onnec’s owners capture part of the upside they helped create. For investors evaluating the deal, the earnout is the honest part of the price — it is CBRE admitting, in the structure itself, that it is buying into a cycle whose duration nobody can guarantee.
The scale of what CBRE is assembling around Onnec is the other number to note. CBRE expects its Critical Infrastructure Services line to generate nearly $4.5 billion of total revenue in 2027, including Onnec’s contribution. Its other services businesses — most notably Project Management and Advisory Services — are expected to generate almost $2.5 billion of infrastructure services revenue in 2027, for total estimated 2027 infrastructure services revenue of nearly $7.0 billion. That figure excludes periodic data-center land sales in the development business. In other words, CBRE is not dabbling in infrastructure services. It is building a business line whose revenue approaches that of a mid-sized utility, assembled from services rather than from regulated assets.
The second piece of a deliberate pattern
Onnec is not CBRE’s first move into this layer. The company previously acquired Direct Line Global from private equity firm Guardian Capital, a Fremont, California-based provider of mission-critical data-center infrastructure services — design, installation, maintenance, and management for data-center owners and operators, primarily in the United States, focused on the hyperscale market. The logic of the two deals is a mirror image: Direct Line gave CBRE the technical services capability in the U.S. market, where it already had facilities management scale; Onnec gives it the equivalent capability across the UK, Continental Europe, and Asia Pacific, where its existing data-center services footprint was thinner.
CBRE’s chief operating officer, Vikram Kohli, framed the strategy in the announcement in terms this publication’s readers should recognize: the company is pursuing an explicit strategy to build a global infrastructure services business to complement its real estate services business, and the acquisition is consistent with its effort to build resilient businesses that benefit from secular tailwinds. Onnec’s chairman, Barry Horgan, framed the seller’s side the same way from the other direction — joining CBRE gives Onnec the platform to broaden its capabilities for clients whose deployments are global, not regional.
Put together, the pattern is a roll-up of the execution layer, assembled geography by geography. CBRE already manages facilities; it bought the technical installation capability in the U.S.; it has now bought the equivalent in Europe and Asia Pacific. The result, if integration works, is a single provider that can take a hyperscaler from site selection through construction management, technical fit-out, and ongoing operations — in every region where that hyperscaler is building. That end-to-end claim is the pitch to the customer, and it is the reason the two acquisitions are worth more together than apart: a global customer would rather manage one relationship than three regional ones.
Why this belongs in a power investor’s notebook
It is fair to ask why a real estate services acquisition belongs in an energy investing column. The answer is that the data-center power story and the data-center execution story are the same story at different points in the value chain. Every megawatt of contracted load eventually becomes a room full of equipment that someone has to install, test, and hand over. The power sector’s analysis has concentrated, understandably, on the front of that chain — the generation, the interconnection, the tariff, the PPA. But the economics of the whole chain depend on the back of it: a campus that is powered but not fitted out earns nothing, and a handover that slips by a quarter pushes revenue, and the returns on the capital behind it, by a quarter.
There is a second, more structural reason. The firms in the execution layer are the earliest indicators of where the buildout is actually going. Developers announce pipelines; utilities file load forecasts; but the installation firms see the purchase orders. When their backlogs grow, the buildout is real. When their utilization falls, the announcements were ahead of the work. For investors trying to distinguish contracted revenue from contracted hope — the distinction at the heart of the data-center asset class — the services layer is a leading indicator worth watching, and CBRE’s quarterly disclosures on its infrastructure services revenue will now be one of the cleaner windows into it.
There is also a competitive read-through. CBRE is not the only real estate services firm moving into this layer — rivals have been expanding their data-center teams as well — and the competition for the execution layer will itself become a market worth tracking. The firms that win it will be the ones that can staff 800-person deployment organizations across three continents and hold quality constant while doing so. That is a human-capital business, not a balance-sheet business, and it consolidates differently: you cannot build it by writing a check alone, which is why CBRE is buying operating companies rather than hiring teams one by one.
What could go wrong
The risks here are not the risks of a power-plant investment, and they should be underwritten differently. The first is integration. Onnec’s value sits in its people and its customer relationships — relationships that, by its own account, in many cases span more than 30 years. A services roll-up that loses the technicians and the customer trust it bought has bought revenue that walks out the door. CBRE’s track record with Direct Line Global is the relevant precedent, and investors should watch retention and customer concentration disclosures in the quarters after closing.
The second risk is the earnout itself, read from the other side. Earnouts align incentives, but they also create a period in which the acquired business is managed partly toward threshold targets rather than purely toward long-run value. The thresholds are not disclosed, so outside investors cannot judge how demanding they are — only whether they pay out, in arrears, over the coming years.
The third risk is cyclicality, and it is the one the deal structure acknowledges most directly. The data-center buildout is a secular trend with cyclical amplitude: hyperscale capital spending can pause, deployments can slip, and a services business geared to installation volume feels a pause faster than a business geared to contracted power revenue. Onnec’s expected 2027 revenue and EBITDA are forecasts made at the top of a historic buildout; if the cycle turns, the earnout protects CBRE’s downside, but the base business still has to be managed through the trough.
Finally, there is the regulatory step. The transaction is subject to customary regulatory approvals and is expected to close around the end of the year. “Customary” should not be read as “automatic” — cross-border services deals touching critical digital infrastructure attract scrutiny in multiple jurisdictions — but nothing in the announcement suggests an approval fight, and the expected year-end timetable implies both sides consider the path clear.
What to watch
Three things will tell investors whether this deal earns its price. First, the close itself: does it land around year-end as expected, or does regulatory review stretch the timetable? Second, the 2027 delivery: does Onnec produce the roughly £370 million in revenue and more than £50 million in core EBITDA that CBRE’s price assumes? Third, and most important over time, the cross-selling: does the combined Direct Line–Onnec platform win global, multi-region mandates that neither firm could have won alone? That third test is the strategic thesis, and it will show up in CBRE’s infrastructure services revenue line and in the win announcements long before it shows up in any single quarter’s earnings.
The broader lesson is the one the price itself teaches. When a company like CBRE pays up to $1 billion for 800 people who install infrastructure, it is telling you where it thinks the scarcity in the data-center economy now lives. It is not in the capital — there is plenty of that. It is not, for the moment, in the power, which the industry has learned to contract and finance. It is in the execution: the capacity to deliver identical, working infrastructure, on time, in every market at once. The electrons made the data-center buildout an asset class. The firms that install the infrastructure are becoming one, too — and CBRE just bought itself a larger share of it.