Compounding at the Speed of Flight

Private aviation has always innovated in the air. Faster aircraft, better avionics, connectivity beamed from space. On the ground, in the software that sources aircraft, moves money, and settles commissions, the industry has been waiting for someone to rebuild the stack.

At Goodwin, that's the job we signed up for. And over the last two quarters, something fundamental changed in how fast we can do it.

The change is measurable. So let's measure it.

The Data

We instrument engineering the way we instrument everything at Goodwin: every shipped feature, integration, and fix lands as a completed story in Asana. For seven straight quarters, from mid-2024 through the end of 2025, our team completed an average of roughly 95 stories per quarter. Steady. Respectable. Linear.

Then the curve broke upward.

In Q1 2026 we shipped 635 stories, up 452% year over year. In Q2 we shipped 983, up 740% year over year and another 55% on top of Q1. In six months, we shipped 4x more than we did in the entire 2025 calendar year.

Now overlay a second line on that chart: AI token consumption across OpenAI and Claude. Around a billion tokens in Q1. More than five billion in Q2. Engineering output and AI consumption are climbing the same wall, at the same time, for the same reason.

This is not a coincidence. It's a compounding loop. AI-generated code ships features faster. Faster shipping means faster feedback from brokers and operators. Faster feedback means better prompts, better context, and better internal tooling, which makes the next billion tokens worth more than the last. The cost of building software isn't just falling at Goodwin. The return on building it is rising.

Small by Design

Here's the context that makes those numbers matter: we didn't do this with a big team.

In mid-2024, Goodwin had four engineers. We ended 2025 with eight. Today we have twelve, and four of them joined this past quarter, two in the final weeks of June. Which means the 983 stories we shipped in Q2 were carried largely by eight or nine ramped engineers while a third of the team was still onboarding.

The per-engineer math tells the real story. Across 2024 and 2025, our engineers completed around 20 stories per quarter each. In the first half of 2026, that number is roughly 80. That's a 4x jump in individual throughput at the same time the team itself tripled. Multiply those together and you get the curve on that chart.

You don't get a curve like that by simply working harder. You get it by treating AI as core infrastructure: agents wired into our development workflow, models embedded in code review and testing, and an engineering culture that expects every engineer to orchestrate machines, not just write functions. It's also why new engineers contribute in their first weeks instead of their first months. The context a senior engineer used to carry in their head now lives in systems the whole team, and the AI, can draw on.

The lazy conclusion is that AI like this replaces engineers. Our data says the opposite.

David Sacks made this his most contrarian prediction for 2026 on All-In: "AI will increase demand for knowledge workers, not decrease it." His argument is Jevons Paradox. When the cost of a resource collapses, aggregate demand for it explodes, because you discover use cases that were never economical before.

That is exactly what's happening inside Goodwin. Every productivity gain has expanded our roadmap faster than it cleared it. Programmable banking, real-time ledgering, dynamic invoicing, smarter sourcing. Each of these was a someday project until AI-scale productivity turned it into a this-quarter project. And every one of them surfaced three more behind it.

So no, we are not shrinking the team. We're hiring for three new engineering roles right now. AI didn't reduce our demand for engineers. It made each engineer so valuable that we want more of them.

What This Means for the Industry

If you're a broker or operator, this is the part that matters. The pace you've experienced from Goodwin over the past six months, from instant payments and trip-restricted accounts to real-time portals and dynamic invoicing, is the new baseline, not a sprint. A platform partner shipping ten times more each quarter closes the gap between what you ask for and what you get in weeks, not roadmap-years.

If you're watching this industry from the outside: private aviation is a market where the aircraft are modern and the back office is not. The team that merges operational and financial data into one programmable, AI-accelerated ecosystem will set the standard for trust, transparency, and speed in this industry. We intend to be that team, and the curve says we're on pace.

I wrote last fall that every trip, every deposit, and every disbursement should be generated at the speed of flight. Add one more to the list: the software itself.

We're not just building faster. We're compounding.

Do Good. Have Fun. Get Wins.