Every hyperscaler partner programme is built around the same job title. Partner Development Manager. The person who's supposed to know your product, know AWS (or Azure, or GCP), and sit in the middle making the relationship actually work.

But it's worth asking the question directly: is a role built around developing partners actually a role that can scale?

I don't think it is. Not in its current form.

What a PDM Actually Is

Strip away the title and the job description, and a PDM is effectively an in-house consultant. They're expected to be a subject matter expert on your business — your product, your GTM motion, your gaps — while also being fluent in the internal mechanics of the hyperscaler: ACE, funding programmes, field priorities, Marketplace mechanics.

That's not an admin function. That's a strategic one. And strategic expertise doesn't scale by adding headcount linearly. It scales the way consulting scales — slowly, expensively, and with a hard ceiling on how many clients (or in this case, partners) one person can actually be good for.

Why the Model Breaks

If you have a handful of high-value partners, a dedicated PDM is genuinely worth it. The economics resolve themselves. One person, deeply embedded, driving outsized outcomes for a handful of strategic accounts — that works, and I've seen it work.

The problem shows up at scale. If you're AWS, Microsoft, Google — or, I'd speculate, increasingly Anthropic, Nvidia, and OpenAI as they race to build out partner ecosystems of their own — you don't have a handful of partners. You have thousands. And you want thousands more.

You can't hire your way out of that. Every PDM you add is still capped at being a genuine expert for a small handful of companies. Multiply that constraint across a partner base in the thousands, and the maths simply doesn't work. This is exactly why we've seen wave after wave of PDM layoffs and restructures across the hyperscalers over the last few years. Not because partnerships stopped mattering — they clearly matter more than ever — but because the model of "one expert human per partner" was never going to scale to the size these programmes are chasing.

The Outsourcing Trap

The obvious answer is to outsource the function. Bring in a firm that specialises in being the PDM layer. And to be clear — that's a real improvement. A specialist partner-development firm can operate more efficiently than an internal hire, spread expertise across engagements, and bring pattern recognition from working across many partners rather than one.

But it doesn't solve the underlying problem. It just moves it. A consultancy still has senior operators with finite time. You can add junior staff, but the strategic value — the actual "development" part of Partner Development Manager — still runs through a small number of people who genuinely know what they're doing. The capacity ceiling is higher. It's not gone.

Is AI the Answer?

Partially — but not in the way most people hope.

AI is already helping consultancies like ours scale further than a pure headcount model would allow. And the hyperscalers are clearly investing in their own AI layers to try to solve this exact problem — better internal tooling, matching engines, and AI-assisted account guidance (see Week 24 on the Algorithmic Matchmaker).

But from what I've seen, AI is good at surfacing information and speeding up admin. It is not yet good at proactive strategy. It can tell a partner what a Better Together story should include. It struggles to walk into a boardroom, read the politics, and push a reluctant sales leader to actually commit resources. That's the part of the PDM role that was always the hardest to replace — and it's still the hardest to replace.

What About Workshops?

Workshops scale in a way one-to-one PDM relationships never will. One session, dozens of partners, consistent content, low marginal cost per attendee. It's the closest thing to a genuinely scalable delivery model in this space.

But workshops come with their own gap: engagement, follow-through, and execution support after the session ends. A workshop can hand a partner the framework. It can't force them to go build a Cloud Desk, fix their ACE hygiene, or get sales leadership to sign off on a co-sell motion. Without something on the other side of the workshop — a structured follow-on, a gated next step, some form of accountability — the content sits in a deck nobody opens again.

The Uncomfortable Truth for ISVs

Here's the part that doesn't get said enough. A lot of ISVs treat a PDM as a catch-all. As if having a dedicated hyperscaler contact will compensate for their own unwillingness to commit internal resources to learning the motion and actually executing it.

It won't. It never did, even when PDM coverage was more generous than it is today. A PDM — internal, outsourced, or AI-assisted — can only accelerate a partner that's already prepared to do the work. No model, no matter how well resourced, fixes a partner that wants the relationship to do the heavy lifting for them.

So What's the Actual Answer?

I'm not going to pretend I've solved this for organisations with partner bases in the thousands. I haven't, and I'm skeptical anyone fully has yet.

What I do think is true: above a certain partner value threshold, a dedicated PDM (or several) is still the right economic call — the outcomes justify the cost. Below that threshold, the current default model doesn't hold up, and pretending otherwise is how you end up with the layoffs we've seen.

The gap in between is where the real conversation needs to happen — probably some blend of scalable content (workshops), AI-assisted enablement, and a much clearer, harder line on what a partner actually needs to bring to the table before development resources get pointed at them. The mission for hyperscalers building programmes today — and, I'd guess, for the growing wave of AI-native players like Anthropic, Nvidia, and OpenAI now building similar infrastructure — is to figure out that segmentation, not just how many PDMs to hire.

Tip of the Week

If you're an ISV waiting on your PDM to build your co-sell motion for you, stop. Build your Better Together story, get your ACE hygiene in order, and show up with a defined ask. The partners who scale their AWS relationship fastest aren't the ones with the best PDM — they're the ones who would have made progress with a mediocre one.

One to Read: Go back to Week 24 (The Algorithmic Matchmaker) — it's the clearest early signal of how hyperscalers are trying to use AI to plug this exact gap, and where that approach still falls short.