By Geoff McDonald, CEO, Ambassador
I'm an AI bull. We build on these tools every day at Ambassador, and I'd make the same bet again tomorrow. So read what follows as a builder doing the math out loud, because the math has stopped working.
On October 5, the 10-year Treasury yield touched about 5.35%, its highest level since April 2002. It started this year near 4.15%. That one number sits underneath the price of money for almost everything a business does.
The short version: money is expensive again, the AI boom is part of the reason, and the bill is about to land on every software budget. If you run customer growth, the wrong move this quarter is cutting the programs that earn revenue from customers you already paid to acquire. The right move is putting more behind them, and consolidating the scattered tools around them before your renewals reprice you.
Bond yields usually feel like someone else's problem. This quarter they belong on every growth leader's desk.
The 10-year crossed 5% on September 15 for the first time since 2007, then kept climbing into October. According to Reuters, the third quarter brought the biggest quarterly rise in Treasury yields this century. On September 16, the Federal Reserve raised its target range by a quarter point to 3.75 to 4 percent, its first hike since 2023, and the median projection from Fed officials points to one more increase before year end.
A soft September jobs report has traders expecting a pause at the October meeting, and I have no edge on what the Fed does next. What matters for planning is the level, and the level is the highest in a generation.
Here's why it reaches you. The 10-year sits underneath the cost of corporate debt, the discount rate investors apply to future cash flows, and the hurdle rate your CFO uses to approve spend. When it moves this far this fast, every growth plan gets re-underwritten, whether anyone announces it or not.
This is the part I find most interesting, as someone who builds on these models every day.
The reporting on the selloff keeps naming the same drivers: oil and the war in the Middle East, sticky inflation, heavy government borrowing, and the AI buildout. CNBC reported that bond issuance tied to AI spending is adding supply that competes with Treasuries, and Bloomberg named AI infrastructure investment as part of what is pushing borrowing costs up globally. The technology I'm most bullish on is part of what is raising the price of money.
Now look at the equity side. SpaceX listed on June 12 in the largest IPO in history, priced at $135 a share. Four days later it closed above $225. Within five weeks it briefly traded below its offering price. Anthropic filed confidentially on June 1, and The Wall Street Journal reports its listing has moved to November. Sam Altman told Fortune in September that OpenAI will not go public this year at all.
Put those together and you get what I've been calling the absorption problem. There is a finite pool of marginal buyers. Every record listing pulls dollars from somewhere, and behind the headline names sits a long line of application-layer companies priced on the assumption that capital stays cheap and patient. Capital just stopped being either one. Something gives.
You can already see the early signs in product quality. I'm watching tools that grew fast on cheap capital start to slip in reviews as customers notice the gap between the launch demo and the daily work. I won't name names, because the pattern matters more than any one logo.
When the repricing comes, the pain lands in a specific order, and the order is the whole point.
First, the largest funds. They marked these companies up, funded the next rounds themselves, and sell into the IPO. The public market ends up holding the shares at the top. That is how their model is designed to work, and they will be fine.
Second, the smaller funds that rode the same companies without the same horsepower. At these prices they can't defend their ownership in the next round, so they get diluted, then marked down, with no exit window when they need one.
Third, the investors behind those funds. Much of that is the newest money in the system: first-time allocators, family offices, first checks into a fund. The losses land on the people least able to afford the lesson.
Fourth, and longest, the customers. Companies that built real workflows on software priced for growth at any cost. When the vendor's math breaks, the vendor raises prices, sunsets a product, gets acquired, or disappears. The customer inherits the migration, the retraining, and the data that never comes along cleanly.
That fourth group is the one I care about most, because it is where our customers live. Your vendor's cap table is your risk, and most buyers never price it.
Repricing often shows up quietly, as a deal with no price attached.
This spring AppDirect acquired PartnerStack, a category leader in partner management software. Terms were not disclosed. PartnerStack's last public raise was a $34 million Series B in May 2021, and the deal was AppDirect's sixth acquisition in twelve months, following its purchase of Tackle.io in December. I respect both teams, and I have no inside view of how the integration is going. My read is about structure: a buyer absorbing that many companies at once has to decide which roadmaps move and which ones wait.
If you're a customer in that position, your product plan now lives inside someone else's integration plan. Watch the release notes over the next two quarters. That is always where it shows first.
Now pull this down to your own budget, because this is where my thesis turns into advice.
Software prices were already climbing before yields took off. Vertice, which tracks more than $75 billion of software spend, measured SaaS inflation at 16.4% in June 2026, the highest reading it has recorded and nearly four times US consumer inflation. Meanwhile, Gartner's 2026 CMO Spend Survey found marketing budgets essentially flat at 7.8% of company revenue, against 7.7% a year earlier.
Run that math on your own stack. If your tools inflate at anything close to Vertice's rate while the total budget holds flat, the difference comes out of your programs. Every point solution is its own renewal, with its own increase and its own negotiation. Ten tools means ten separate chances to get repriced.
I expect it to get worse before it gets better. Vendors that raised at peak prices now face expensive capital and a crowded exit window. The fastest way for them to fix their math is your renewal. Some will raise list prices. Others will add usage meters, AI surcharges, or tier changes that charge more for the same work. I can't tell you which vendor does what, so I'd plan for all of it.
I'll be direct here, and I'll own the bias up front. I run a customer lifecycle company. Weigh my view accordingly, then check the math yourself.
When money is cheap, a company can afford to rent growth. Buy the traffic, fill the funnel, worry about retention later. When money is expensive, the discount rate climbs, and revenue that arrives years from now is worth less today. Revenue you can earn this quarter, from customers you already paid to acquire, is worth more. Retention, expansion, referrals, and advocacy all live in that second bucket, because they earn on an acquisition cost you have already sunk.
So the reflex to treat lifecycle programs as a nice-to-have and cut them first runs backward right now. You would be trimming the cheapest revenue you have to protect the most expensive. And when the next budget review asks what the growth line returned, you would be the team without an answer.
That's the scramble I'm worried about. The companies that haven't built this muscle yet, or that still run it across five disconnected point tools, will reach for it in the middle of a budget squeeze, when every dollar has to be defended and every tool is repricing at once. A lifecycle program takes time to build well. The teams that start now get to build it calmly. Everyone else builds it in a panic.
If you take one thing from this post, run these four questions against every software vendor in your growth stack this quarter:
Then run the same test on your own lifecycle program. Count the tools it takes to run referral, incentives, retention, and the reporting that proves them. Count the renewals. Then ask whether the results show up in one place your CFO can read.
This is why we built Ambassador as one system instead of a stack of tools. Acquisition, retention, and proof run on the same spine, so a referral, the renewal that followed, and the revenue behind both sit in one record instead of three. HiroAI orchestrates the programs, and the outcome reporting shows what they returned. Across the programs we attribute cleanly, that adds up to more than $2.4 billion in attributed revenue to date, at a return of at least 3:1 and 5:1 or better at the top end.
Back in August I wrote that the market now pays for proof and discounts promises. The price of money just extended that rule to every budget line, including yours. When capital was free, a promise was enough to get funded. Capital isn't free anymore, and proof is what gets funded now.
So my advice runs opposite to the reflex. Protect the customers you already paid for, invest behind them, consolidate the tools around them, and make the return visible before your next renewal forces the question.
Grow. Keep. Prove.
Ambassador is The Customer Lifecycle Operating System, orchestrated by HiroAI. When acquisition, retention, and proof run on one system, your growth budget has fewer renewals to fight and one clear number to defend. See how it works.