The Deal Is Closed. The Carve-Out Has Barely Started.

There is a particular kind of optimism in the room when a carve-out closes. The lawyers are done. The funds have moved. The separation plan has a row for IT, and that row is green, because there is a signed Transition Services Agreement and the seller has agreed to keep the lights on. Everyone moves on to the parts of the business they find more interesting.

Then it is month nine. Finance has its own systems. HR has migrated. The brand has a new website. And IT is still almost entirely dependent on the company you just separated from, with a TSA that expires in five months and an extension clause priced to make sure you never use it twice.

This is not an unusual story. It is closer to the default.

Why IT is the workstream that always runs late

A carve-out is, at its core, an exercise in pulling apart things that were never designed to come apart. IT is the connective tissue holding all of it together. Payroll runs on it. Finance closes the books on it. HR, customer support, order fulfilment, email, the badge that opens the front door. When people call a carve-out "operationally complex," what they usually mean, whether they know it or not, is that IT is complex.

The numbers back this up. KPMG's Global M&A Outlook 2026, based on a survey of 700 dealmakers, found carve-outs to be the defining deal theme of the year, with technology the single most-targeted sector. It also found that 52% of corporate dealmakers name operational disentanglement as their biggest execution challenge, and 40% point specifically to IT and data separation. Those are not two separate problems. The operational disentanglement largely is the IT.

Here is what makes IT different from every other workstream. When you separate finance, you are moving a function. When you separate IT, you are moving a function and discovering, in real time, every undocumented dependency the parent built over fifteen years. Shared identity. Licensing agreements that quietly do not survive a change of control. A managed service contract that covers both entities and cannot simply be cut in half. Data that lives in one tenant and is read by three systems in another. None of this is on a diagram, because nobody ever needed a diagram. It worked.

A TSA is a bridge, not a destination

The Transition Services Agreement is supposed to absorb this risk, and it does, for a while. The seller keeps providing IT services to the carved-out business for an agreed period, usually somewhere between six and eighteen months. It buys you time.

The mistake is treating that as the same thing as solving the problem.

A TSA does not separate anything. It postpones the separation and puts a price and a deadline on the postponement. The day it is signed, a clock starts. And of all the workstreams racing that clock, IT is the one most likely to still be running when it hits zero, because IT started from the hardest position and usually got the least planning attention before close. Advisory firms who work these deals have a standard phrase for what a signed TSA does to a buyer: it creates a false sense of security. The agreement looks like a safety net. It behaves like a countdown.

When the clock runs out and IT is not ready, you have two options, and you will not enjoy either. You extend the TSA, at a price the seller set specifically so that you would not want to. Or you exit on schedule and into a gap, where for some stretch of time the business you bought cannot do something it needs to do every single day.

Due diligence answers a different question

Part of the reason IT separation gets underestimated is that standard IT due diligence is built to answer a different question.

Diligence tells you what you are buying. The age of the infrastructure, the technical debt, the security exposure, the key-person risk, the quality of the codebase. All of it useful. All of it necessary. But "what is the state of this asset" and "what will it cost and take to make this asset stand on its own" are two different questions, and the deal model frequently holds a confident answer to the first and a vague one, or none at all, for the second.

The cost of standing alone is its own line item. It includes licensing renegotiated or repurchased in the carved-out entity's name. Contracts novated or replaced. An identity and access platform built from nothing. A cybersecurity baseline that no longer inherits anything from a parent. Data migrated, validated, and cut over without losing a day of operations. Roland Berger's recent study on carve-out costs put it plainly: the one-off costs of separation are consistently underestimated, and that is precisely where intended deal value quietly leaks away.

What getting it right looks like

None of this is an argument against carve-outs. Pulling a non-core business out of a structure that was holding it back is one of the more reliable ways to create value, which is exactly why so many boards are doing it now. It is an argument against treating IT as a box that the TSA ticks.

The carve-outs that land on time tend to share three habits. The first: day-one operability is named as its own target, separate from deal close. The deal closing and the business being able to run independently are two different milestones, and confusing them is how you end up surprised in month nine.

The second: the TSA exit plan exists before the TSA is signed. If you cannot describe how you are going to leave the agreement, you do not yet understand what you are agreeing to. The exit plan, with its own budget and its own timeline, is what makes a TSA a bridge instead of a trap.

The third: IT separation has a single accountable owner with real authority and a real budget. Not a workstream lead reporting green because the TSA exists. Someone whose actual job is to make the carved-out company independent, who treats the TSA end date as a hard deadline from the first week rather than the last.

The expensive version is the one that felt free

The costly path through a carve-out is the one that feels free at signing. A TSA gets signed, the IT row goes green, and the genuine work of separation gets quietly deferred into a future that arrives faster than anyone budgeted for. The value the carve-out was meant to unlock gets spent, instead, on emergency extensions and recovery.

An independent read on IT separation cost before the deal, and someone credible owning the separation after it, is not overhead. Set against the price of a carve-out that misses its own deadline, it is one of the cheaper forms of insurance you can buy.

If you are looking at a carve-out, on either side of the table, that is a conversation worth having early rather than late.

At Galactus, we help organizations get the IT side of a carve-out right, before the TSA is signed and long after the deal closes. If a separation is on the table for you, on either side, let's have a conversation.