Elastic Path is the API-first, AI-ready commerce platform for brands with complex requirements: large catalogs, multi-tier pricing, and multiple brands and markets. Manufacturers, distributors, and retailers run account hierarchies, quoting, and negotiated pricing natively, not through bolt-on apps and workarounds.
TL;DR
Top News of the Week: For years, software vendors counted on customers not paying close attention to the bill. AI agents check the cost of every call.
From Last Week’s News: Home Depot brought in $2 billion more in sales this quarter than a year ago, on about 800,000 fewer customer transactions. Management expects the same pattern through year end, with roughly 1.4% comparable growth.

TOP NEWS OF THE WEEK
SaaS Pricing Was Built For Buyers Who Weren’t Watching The Bill
Jason Lemkin wrote this week that B2B companies rebuilding themselves for AI (think Hubspot refounding) should cut their prices, in some cases by a lot. I agree with him, and I also expect very few of them to do it.
For most of the last twenty years, a person bought business software, rolled it out to a team and then stopped paying close attention to it. Half the seats would go quiet, nobody worked out the cost per action, and the renewal showed up with another price increase that got grumbled about and signed anyway. Lemkin’s point is that a lot of SaaS pricing quietly depended on that kind of neglect, and agents don’t neglect anything. They see the cost of every call they make.
He uses SaaStr as the example. The company says it runs on three people and more than 20 AI agents, and one of its core vendors saw the bill go up 40% this year, nearly all from agent traffic of around 35K calls a day. SaaStr responded by cutting those calls by roughly 90% while still needing the product just as much. These are SaaStr’s own figures and SaaStr sells an agent playbook built around them, so I’d treat them as one company’s experience. The pattern makes sense to me anyway.
They also replaced Marketo with a cheaper email API, and they’re moving away from some AI tools that work well but cost too much per call for agents that need to hit them hundreds or thousands of times a day. For the vendor, the damage starts well before renewal, because the customer uses the product less every day in the run-up, and that lower usage is what they look at when the contract comes up.
Lemkin’s model says that if you cut the price per action by 80% and agent usage rises 20x, revenue from that customer goes up 4x. The arithmetic checks out. The 20x is the assumption doing all the work, though, and I’d want evidence from more than one company before a board signs off on it.
Databricks gives him better support. Ron Gabrisko, its CRO since the company was under $1M in ARR, said on SaaStr’s podcast that Databricks once charged $1K per user on top of consumption, and customers told him that fee was holding back wider rollouts. He dropped it and consumption went up sharply. Lemkin values Databricks at $200B today.
The resistance is easy to understand. Margins take a hit at first because inference isn’t free, NRR models assume annual price increases, and sales comp is still mostly tied to seats. A price cut looks like giving back revenue that’s already been booked.
His practical suggestions include pricing agent and API usage separately from human seats, getting rid of punitive overages and tracking calls per customer per day as their own metric. He also suggests testing a 50% to 80% cut on a single tier to see what happens to consumption.
THE BIG IDEA
In commerce, merchants are already connecting agents to catalog, pricing and ad tools that still bill as though a human logs in twice a week. Lemkin says the drop in usage shows up months before the churn does. I haven’t seen a retail tech vendor publish agent pricing that reflects any of this yet, and I’m curious who will be willing to explain that margin dip to their board first.

THE TRUTH ABOUT LAST WEEK: Home Depot Beat Its Plan While Customer Traffic Kept Falling
Home Depot’s third quarter came in ahead of the company’s own plan. Sales were $43.4B, up 4.9%. Comparable sales rose 2.2%, so new stores and acquired businesses account for a good share of the growth.
Comparable transactions fell 0.5% and average ticket rose 3.0% to $93.09. Customers came in less often and spent more per visit, which points to pricing and larger pro orders more than any pickup in housing activity.
Operating margin was 12.6%, down from 12.9% a year ago. Adjusted EPS rose 3.2% to $3.86, behind sales growth of 4.9%.
Full-year guidance now calls for comparable sales growth of about 1.4%. Management is planning for a soft housing market through the end of the fiscal year.
Incoming CEO Richard McPhail said the results exceeded expectations, and on the company’s own terms they did. The pro business Home Depot bought through SRS and GMS is carrying more of the load while DIY traffic slips.
WHY IT MATTERS
Home Depot beat its own forecast for the quarter and grew both sales and earnings. Margins still slipped by 30 basis points, and DIY customers are coming in less often than a year ago. The company has spent $1.8B on acquisitions so far this year, nearly all of it aimed at contractors. It has said much less about what it plans to do for homeowners doing their own projects.



