B2B GTM for AI Enterprises: Ed (Databricks, AWS)

GM (APJ) @ Databricks, ex-VP (APJ) @ AWS, ex-GM (SE Asia) @ VMware

Ed has spent the last two decades building and scaling go-to-market engines for some of the world’s most transformative cloud and data companies.

He led VMware during the virtualization boom, helped AWS become the dominant cloud provider across APAC, and now serves as GM (APJ) at Databricks (US$60B+ valuation), where he’s driving the adoption of unified data and AI platforms across the region.

In this conversation, Ed shares why Asia is not one market, how to solve GTM as a supply-side challenge, and what it really takes to drive AI readiness when every country is at a different stage.

Q: You’ve scaled companies like VMware, AWS, and Databricks across Asia. What’s changed about building GTM muscle in the region?

Ed: In truth, not a whole lot. The business problem is still the same. Whether it was AWS 10 years ago or Databricks today, we’re working within seismic shifts in how people produce and consume technology.

In these shifts, you don’t have a demand problem. You have a supply problem.

Take Databricks. Everyone is moving their fragmented, complex, expensive data estate into a unified, cloud-based data platform where AI is a first-class citizen. This is inevitable - with or without us. So we’re not out there trying to generate demand. We’re solving for how to build, deliver, and support that transformation. That’s a supply-side GTM challenge.

Q: You’ve said market selection is the most important GTM decision. How do you approach it?

Ed: Too many companies still talk about APJ like it’s a monolith. It’s not even one single place. It’s a collection of deeply different markets, and you cannot succeed by going wide too early.

You need to pick your markets.

I break Asia into two buckets:

• Easy Asia: India, Southeast Asia, ANZ

• Hard Asia: Japan, Korea, China

“Easy” doesn’t mean the deals are easy - but it’s easier to start. You can get by with English, hire faster, and scale more predictably. In “hard” Asia, you need deep localization, language fluency, and white-glove enterprise support. You’re dealing with closed markets and longer enterprise sales cycles. You don’t walk into Japan with one person and figure it out.

Q: You mentioned the center of gravity has shifted. Why is Singapore such a strong starting point?

Ed: Historically, American companies entered Asia through Australia. But winning Australia doesn’t help you win Asia.

Singapore is geographically central. You can run India, Southeast Asia, ANZ, and even parts of North Asia from here. It’s business-friendly, globally connected, and welcoming to tech companies. And because all the hyperscalers came here first, the infrastructure is world-class.

There’s also a network effect - the more APJ HQs that are based here, the easier it is to hire, partner, and build. At Databricks, the Singapore ecosystem has been incredibly supportive of our regional buildout.

Q: So once you pick a market, how do you enter and build?

Ed: We follow a phased model:

1. Start with a Minimum Deployable Unit.
   A tiny, cross-functional team - maybe a salesperson, a pre-sales architect, a partner manager, and a demand gen lead. Just enough to “soften the beach” and validate the opportunity.

2. Track traction against key metrics.
   That could be pipeline growth, customer acquisition, partner engagement, or early revenue. If the signals are strong, you earn the right to scale.

3. Evolve into a Market-Making Unit.
   That means dedicated leadership in-market, support, pre-sales, services, and full local capability. At that point, you're not testing anymore. You’re building.

This model has served me at VMware, AWS, and now Databricks. It works because it respects the diversity and scale of Asia without overextending resources.

Q: With limited resources, how do you prioritize functions - sales, marketing, partners, services?

Ed: You don’t. You bring them all together from the beginning.

Founders often think of GTM like a funnel: first sales, then marketing, then partnerships. That’s wrong. The magic happens when every GTM function is aligned on the customer experience from the very first conversation.

At Databricks APJ, I have leaders across all GTM disciplines - sales, marketing, customer success, partners, pre-sales, professional services - on my leadership team. We meet multiple times a week to plan, measure, and execute. Our only two goals: acquire customers and grow customers. Everything maps back to those.

Q: Every market is talking about AI now. How do you tailor the GTM for markets at different maturity levels?

Ed: The truth is: it’s not an AI problem. It’s a data problem.

You can’t do AI if you don’t trust your data. If you don’t know your lineage, can’t govern it, or can’t unify it - your AI initiative is going to fail. That’s why Databricks has always been the data + AI company, even before AI became hot.

We already have access to more data inside our customers than almost anyone. That means we can say: don’t move your data to a model - bring the model to your data. Start simple: maybe batch inference or risk analytics. Then evolve to fraud detection, personalization, supply chain forecasting, etc. But you only get there if your data platform is strong.

Q: Any final thoughts?

Ed: Don’t try to do everything at once.

Pick your market. Build your team. Earn the right to scale. And remember: Asia is not one market - it’s many. Respect the nuance, and go deep before you go wide.