Demand Aggregation

Could routing spend become the moat

Ramp already routes roughly one hundred billion dollars of business spend, a volume large enough that Eric Glyman raised a specific possibility on the Cheeky Pint: that scale alone could make Ramp a pricing arbiter, the way Costco is for the small retailers that shop there. Glyman, investor Alex Rampell, and host John Collison discuss whether aggregating customers' demand to extract vendor discounts could become the moat as Ramp scales.¹ Rampell frames the mechanism through Groupon, which began as ThePoint.com, a collective-action site built on the idea that if ten thousand people each committed a dollar, a cause could become a discount: get fifty people to commit to a pizza place and get a lower price. The Costco version is that every small restaurant shops there, and Costco uses collective bargaining to tell Coca-Cola "give us a very low price," passing most of it to customers. With that volume of spend to direct, the same logic says Ramp could ask a vendor for a 20 percent discount and mean it.

The forms the leverage takes

Glyman maps the shapes this could take. The first is the group purchasing organization, born in healthcare and now common in private equity: aggregate demand, negotiate bulk discounts for specific purchase types. Ramp already sees dozens, and growing, of merchants where it routes billions of dollars, the client's own money going where the client chooses, but Ramp can credibly tell a vendor how many dollars will flow to it across the Ramp buyer base over the next year and ask for a discount in return. The second, in Collison's framing, is the preferred-vendor bake-off: movie chains and fast-food chains run a Coke-versus-Pepsi bake-off every five years and serve only one, and Ramp could hold preferred rates the same way. "We should try everything." The third is merchant-specific pre-committed balances, "commit two thousand dollars of next year's spend for eighteen hundred," the same pattern Costco uses when it sells one hundred dollars of Starbucks gift cards for eighty, banking locked-in spend and breakage. It is a loyalty and risk-management play: de-risked, pre-committed demand is worth a discount even to a merchant like Amazon, with its low cost of capital, because it removes churn risk from their plan. A fourth is advertising-adjacent surfacing: telling a customer whose renewal is 90 days out about a vendor's 20 percent complimentary welcome rate.¹

The caveat that keeps it honest

Glyman does not treat aggregation as the core thesis, and his own caveat is the important part. Negotiating a shipping deal is interesting, he grants, but then comes the funnel, getting people to know, sign, and use the offer, and the marginal hour is often better spent automating accounting and doing the financial work. This is where the concept connects to banks sell money, Ramp sells time: no bank can match the time vector, so "they sell money, we sell time" remains the durable edge, and aggregation is a "try everything" adjacent to it rather than a replacement for it. The bargaining leverage is real but built on data, since knowing per-seat prices and where dollars flow is what makes the discount ask credible in the first place, the same proprietary-data foundation that underwrites the dark-matter moat. It is best read as a candidate second-stage moat, a scale-derived advantage passed to the customer once early product-led differentiation matures, and it sits close to the market-expanding logic of demand generation versus demand capture in treating routed demand as an asset.

Where it strains

The sharpest tension is channel conflict. It is the client's money, and steering spend toward preferred vendors risks the neutrality that makes Ramp trusted, with the advertising-adjacent version the most exposed edge of that. Aggregation also carries a funnel cost, awareness to signup to usage, that Glyman flags as the reason it may lose to simply automating the work, and the group-purchasing model invites a take-rate that can feel misaligned with the promise of saving the customer money. The concept is therefore held on the platform as an explored option with a stated caveat, not as a settled strategy.