Who pays for the AI upgrade — and the one number every franchisee can now demand

The question that usually goes unanswered

Every franchise system eventually hands its franchisees a technology mandate. New point of sale, new kitchen display, new scheduling tool, now an AI layer. The brand explains the benefit in adjectives and the franchisee pays in capital. Ask what the return per outlet is and the answer is usually a case study from three unnamed pilot locations.

This week gave the industry a different template, and it is worth separating what is genuinely new from what is merely large.

What was actually disclosed

At its investor day on 23 September 2026, McDonald s committed $8.5 billion of NEXT partnering support through 2036, delivered to franchisees as rent relief and capital support, per 24/7 Wall St. Food Chain Magazine reported on 28 September 2026 that about $5 billion of that is expected by 2030; that the per-restaurant spend runs about $800,000 for a US drive-thru and $650,000 to $700,000 internationally; that an average US restaurant is expected to gain about $100,000 of annual cash flow; and that the implied payback is roughly four years. The stated system-level target is about 250 basis points of gross restaurant-level efficiency, on a programme combining redesign, kitchen equipment, revised procedures and ArchIQ, a generative-AI-enabled platform.

The AI is not the interesting part. Voice ordering, demand forecasting and scheduling tools are widespread. What is unusual is a franchisor publishing cost per unit, benefit per unit and payback period for a programme it wants its franchisees to adopt.

Why the disclosure matters more than the technology

Three things follow from putting those numbers in public.

First, it makes the programme falsifiable. A four-year payback is a claim that can be checked against actual profit and loss statements in 2030. A brand that publishes it has accepted being measured against it, which is a meaningful constraint on the usual optimism.

Second, it sets a negotiating benchmark for every other system. A franchisee asked to fund a mandated upgrade can now reasonably ask the same three questions: what is the cost per outlet, what is the expected annual cash-flow effect per outlet, and over what period does it pay back. Any franchisor who cannot answer those has not finished the analysis, and that is useful information by itself.

Third, and most substantively, it reframes who carries modernisation capex. Rent relief and capital contributions mean the brand is absorbing part of the investment rather than simply requiring it. That is a different bargain from the classic remodel clause, where the franchisee funds the upgrade and the brand collects royalty on whatever uplift follows.

The arithmetic a franchisee should actually run

The headline payback is an average, and averages are where unit-level decisions go wrong. Anyone evaluating a similar programme should rebuild it locally:

• Recompute with your own revenue. A benefit quoted for an average restaurant scales with volume. A location at 70% of system average volume does not get the average cash-flow gain, and its payback stretches accordingly.
• Separate the technology from the building work. A bundle of redesign, equipment, procedures and software reports one blended return. Ask which portion of the gain comes from the AI layer, because that is the part that might be available without the construction.
• Price the disruption. Closure or reduced-capacity weeks during a remodel are a real cost that rarely appears in the payback figure.
• Check what the support actually is. Rent relief reduces ongoing cost; a capital contribution reduces upfront outlay. They affect cash flow on entirely different schedules, and a franchisee near a financing limit cares which one arrives.
• Ask what happens if the gain does not materialise. A published estimate is not a guarantee, and the loan on the fit-out does not adjust.

The counter-argument, stated fairly

Markets did not applaud. On 23 September McDonald s shares fell 4% to $239.56 while Wendy s slipped 1% and Yum Brands eased 0.4% — a spread indicating the reaction was specific to the company rather than the sector. The bear case reported alongside is straightforward: the spending starts now and the benefits arrive across a decade.

That scepticism is worth holding onto rather than dismissing. A programme that improves restaurant-level efficiency by 250 basis points is a real operational gain, but the same reports note second-quarter 2026 comparable sales rose only 1.3% globally and 0.8% in the US. Efficiency and demand are different problems, and capital committed to the first does not solve the second.

What we do not know

• The reported figures are company estimates published at an investor day, not audited outcomes. Nothing yet demonstrates the $100,000 annual gain or the four-year payback in practice.
• The split of the $8.5 billion between rent relief and capital contributions is not broken out in these reports, so the cash-flow timing for a franchisee cannot be modelled precisely.
• No breakdown attributes the 250 basis points between ArchIQ, the kitchen equipment, the redesign and the procedure changes, so the AI contribution on its own is unknown.
• No figures were disclosed on franchisee margins, so it is not possible to say what share of an average operator profit the $800,000 represents.
• Whether the programme is mandatory, and what happens to an operator who declines, is not addressed in these reports.

This article analyses publicly available information. It is not investment advice and promises no level of return.

Get next week's brief by email. Free.