Technology Investment J-Curve

Pay the penalty first

Brad Jacobs describes a two-phase technology strategy for building a company through acquisitions, where the cost is finite and upfront and the benefit compounds. In the first phase, roughly the first two years of a new company, he tells investors explicitly not to expect earnings, because he is building what he calls the central nervous system of the organization before buying significant revenue.

The components are a shared-service back office for accounting, FP&A, invoicing, and HR, plus the technology infrastructure for prediction, analytics, warehouse and transportation management, and route optimization, plus the integration capability that lets later acquisitions absorb cleanly. As he frames it, the goal is to be able to take on billions of dollars of revenue and still "get an invoice out and get paid."¹ The investment penalizes the P&L in a major way, and he says so deliberately and without apology, because the penalty ends.

Ride the curve

Once the infrastructure is in place, in Jacobs's account, every acquisition drops into a capable machine, and the unit economics of integration improve with each deal because the fixed overhead is already paid. "You have this huge J-curve after that, because now you've got all that set up and you can buy things of many billions of dollars in size and actually function well."¹ Technology in this phase actively strips cost through inventory management, demand forecasting, route optimization, and algorithmic pricing, so each marginal dollar of revenue runs through a lower-cost structure than the one before it.

Jacobs notes that he ran the same two-phase approach at United Rentals and XPO before QXO, which he presents as evidence that it is an architectural decision rather than an ad hoc one: build infrastructure capacity before buying revenue, not after. The alternative, buying revenue first and retrofitting the infrastructure, produces the duplicated functions and scrambled org charts he inherits when acquiring incumbents. The machine, not any individual acquisition, is the thing Jacobs builds and rebuilds.

Investor communication as part of the design

Telling investors upfront not to ask about earnings for two years is, in his telling, itself strategic. It sets expectations that match the actual plan, creates permission for the right short-term decisions, and filters for investors who understand the playbook, the repeat backers who, as he puts it, already get the joke. He presents this as the same discipline he applies to individual deals, extended up to the level of company-building, which connects it to his broader capital-allocation-discipline and to his approach to using corporate structure and stock as an allocation lever in the-outsiders-capital-allocation.

Speaking with FreightWaves about QXO, Jacobs names every layer of the stack being replaced across acquired building-products companies: ERP, CRM, e-commerce, WMS, TMS, and HRS, with hundreds of millions of dollars going in for the return.² He makes the analogy to the early XPO years explicit: people asked whether he was crazy to invest in technology before the revenue and profits could support it, and his answer then and now is that it is not a nice-to-have but a must-have. In his account the building-materials industry did not even have warehouse management and barcode scanning as a baseline, which he frames as making the phase-one penalty larger and the eventual curve steeper.