On October 5, C.H. Robinson agreed to buy RXO for about $5.8B, or $30.25 a share: $17.25 in cash plus 0.0856 of a C.H. Robinson share. In the press release it said it would apply its 「Lean AI operating model」 to RXO and expects about $300M of net run-rate cost synergies within two years of closing. Closing is expected in the first half of 2027. RXO stock rose about 23% on the day and C.H. Robinson fell about 10%, TT News reported.
Neither company gave a headcount for the deal. That is the gap this piece works inside.
Where the $300M comes from
The investor deck filed with the 8-K splits the savings 80/20. The 80% is cost-to-serve savings, described as operating leverage from the Lean model and a fleet of AI agents deployed across workflows. That is about $240M by our arithmetic. The other 20% is shared services, real estate, insurance and vendor procurement, the usual merger overlap.
Put the $300M next to RXO’s own numbers. The deck shows RXO’s 2026 estimates at $6.8B gross revenue, $1.0B adjusted gross profit and $137M adjusted EBITDA. The savings target is more than twice that EBITDA and about 30% of that gross profit. A buyer does not pay a 29% premium to find savings of that size in office leases.
What the model did at C.H. Robinson
The acquirer’s own filings show the track record. C.H. Robinson’s FY2023 10-K reports 15,246 employees at year-end. Its FY2025 10-K reports 11,855. That is 3,391 fewer people, down 22%, over two years in which its North American volume kept outgrowing the Cass Freight Shipment Index.
The deck gives the other side of the ratio: adjusted gross profit per employee in its North American surface transportation segment rose from $162K in FY2023 to $331K in FY2025, and it says more than 100 AI agents now automate quote-to-cash tasks. The Q2 2026 earnings summary claims productivity up more than 60% since the end of 2022. Together those are the company’s case that it needs about half the people per dollar of gross profit it needed three years ago.
RXO sits at the other end of that curve. Its FY2025 10-K counts 9,218 workers at year-end: 6,906 regular employees and 2,312 temporary. Divide the $1.0B 2026 gross profit estimate by 9,218 and you get roughly $108K per worker. That mixes an estimate with a year-end count and is not like-for-like with the $331K segment figure, so read it as direction, not a ratio. The direction is a gap of about three to one, and the $300M is the price of closing it.
Not the first synergy with AI in it
Merger savings have long come out of headcount. What changes here is the stated mechanism. When Paramount and Warner Bros. Discovery cleared their last hurdle on September 21, the $6B of synergies was a duplication story. Here 80% of the target is attributed to software that does the work. BMW said on September 30 it would cut management roles by 20% by mid-2027 and tied the leaner structure to agentic AI, also without a headcount. Acrisure’s 2,250 cuts in May were an insurance broker doing the same arithmetic. The pattern in all three: the AI claim arrives first, the number of people arrives later.
For scale, Challenger counted 3,961 U.S. job cuts attributed to AI in September, fifth among reasons. A deal like this never shows up in that count, because nothing is announced as a layoff.
What it means for the jobs
The release names no roles. The deck names the work: quote-to-cash, meaning pricing quotes, load tendering, order entry, invoicing and collections. In a brokerage those tasks are done by carrier sales reps, customer service reps, pricing analysts and billing clerks. That is our read of where the cost-to-serve line lives, not something either company said.
Timing is the part you can plan around. Closing is first half 2027 pending approvals, and savings are targeted within two years after that, so the heaviest effect on RXO’s roles falls between 2027 and 2029. The first buffer is probably RXO’s 2,312 temporary workers, a quarter of its workforce, who can leave without a layoff notice. C.H. Robinson’s own two-year path, 22% fewer people while it outgrew the freight market, is the closest preview of what its integration team will try to repeat. The unanswered question is how much of the $240M is people and how much is third-party services, and nothing filed on October 5 says.
