Madrona’s 2025 CEO Summit brought together portfolio company leaders for a day focused on acceleration — moving beyond probabilistic thinking, letting go of anchors to embrace what’s possible and accelerate their mission and their companies.
The event featured conversations with Bob Muglia (founder of Snowflake), Thomas Dohmke (former CEO of GitHub), Nikita Shamgunov (former CEO of Neon, now VP at Databricks), and Ron Gabrisko (CRO of Databricks), offering insights on scaling from startup to massive enterprise success. In addition, our CEOs participated in a “growth mindset” session led by Loren Alhadeff (former CRO of DocuSign) and Anna Baird (former CRO and COO of Outreach).
Against the backdrop of unprecedented AI investment by big tech — with Microsoft, Meta, Amazon, and Alphabet committing $360 billion in AI CapEx this year — 4 major themes emerged that we think will shape how companies accelerate in 2025.
The Scale Paradox: Fast Growth Demands Different Rules
The most striking theme of the day was how rapidly scaling companies must reinvent themselves at every stage. Thomas shared how GitHub navigated explosive growth after Copilot’s launch, emphasizing that every time he was ready to leave Microsoft, they always had something else for him to do. The key insight: successful scaling requires accepting that your company will become fundamentally different every 12-18 months. I completely agree and would add that in this fast-paced AI-native world, product-market fit could be an ephemeral as opposed to a permanent state.
Ron reinforced this with Databricks’ journey from “begging companies to give us like $18K” to becoming a $4+ billion revenue company. As he put it: “75 to 80% of the revenue is going to be in the enterprise,” but getting there meant completely reimagining their go-to-market motion, moving from developer-first PLG to enterprise sales while maintaining product velocity.
Takeaway for Founders: Plan for organizational reinvention, not optimization. Build systems and culture that can handle complete transformation every 12-18 months. Don’t get attached to your current operating or business model. It’s temporary by design. The companies that thrive are those that embrace becoming a different company at each stage rather than trying to preserve what got them there.
Building Teams That Scale With You
One of the most memorable insights came from Nikita: “The team you build is the company you build.” This wasn’t just about hiring; it was about creating systems that enable new people to succeed at unprecedented speed.
Ron shared a counterintuitive approach: “Your existing people need help, but they’ll be okay. You need to make your new people successful.” At Databricks, this meant investing most leadership time in onboarding and enabling new hires rather than managing existing top performers. This philosophy enabled them to double their sales force annually while maintaining quality.
Takeaway for Founders: Flip your time allocation. Spend 70% of your leadership bandwidth on scaling people, especially the new hires and stars. Build onboarding systems that compress months of learning into weeks. Your existing team will figure things out — your new people are the constraint to hypergrowth. Create promotion pathways that reward people for successfully onboarding and developing others, not just individual performance.
The Infrastructure-Application Feedback Loop
A recurring theme was how breakthrough applications drive infrastructure needs, which then enable the next wave of applications. Thomas noted that GitHub Copilot launched “a year and a half before ChatGPT,” giving them unique insight into how developer workflows evolve. The lesson: being first with infrastructure often matters more than being perfect.
This creates opportunities for companies building the picks and shovels, as well as the applications above them in the stack. I noted that we’re in a “super cycle” where “the cost of intelligence and reasoning is asymptotically going to zero,” which should unlock unprecedented productivity for the years ahead. Moving fast and iterating quickly is the key to staying ahead in this race.
Takeaway for Founders: Don’t wait for perfect infrastructure. Build on what exists and help shape what comes next. If you’re building applications, contribute to open-source infrastructure projects that your success depends on. If you’re building infrastructure, stay close to application builders and solve their immediate pain points, not theoretical future needs. The best companies are founded by people who felt the pain firsthand.
The AI-First Operating Model
Perhaps most importantly, successful companies are already AI-native in their operations. Ron shared that everything he does is “Databricks on Databricks” with their own Lakehouse containing all customer data: “I can predict my revenue within 1 or 2%. I can see which customers might churn. We have next best action on every single customer.”
This isn’t just about using AI tools. It’s about rebuilding core business processes around AI capabilities. The companies that figure this out first will have sustainable advantages in efficiency, decision-making, and customer success.
Takeaway for Founders: Become your own best customer immediately. Use your product in your own operations from day one — it will reveal product gaps faster than any customer feedback. Build AI into your core business processes (sales, marketing, customer success) before your competitors do. The winners in this cycle won’t just build AI products — they’ll operate as AI-native companies. Start tracking leading indicators that AI can act on, not just lagging metrics that humans review.
“Lockdown Mode”: Crisis as Your Ultimate Accelerator
A powerful insight came from Thomas’ tactical approach to galvanizing GitHub’s entire organization around a single objective when facing competitive threats. In October 2024, GitHub realized they were losing developer mindshare to Cursor, whose public growth numbers were outpacing its own. Thomas and GitHub’s leadership invented what they called “lockdown mode”, focusing a large part of their organization on only one single thing: getting Copilot back on track from a growth perspective. But the key wasn’t just declaring focus; it was defining clear exit criteria to return to “normal mode,” and bringing complete visibility to every detail that mattered to Copilot’s success.
The results were immediate. They shipped Copilot Free three days before Christmas — something Thomas said would have never happened without lockdown mode, because people always raise concerns.
Takeaway for Founders: When facing existential competitive threats, don’t just work harder. Restructure everything around a single objective. Define clear success metrics and exit criteria upfront so your team knows this isn’t permanent. Create forcing functions, such as daily reporting to leadership, to make status transparent and decision-making fast. Most importantly, use crisis moments to cut through normal organizational friction and get direct access to the resources you need. The best way to accelerate is to temporarily abandon everything that isn’t mission-critical
Your Next Move
Looking ahead, the message was clear: This is the year for offense. As I emphasized on stage, “acceleration isn’t a choice. It is something that we have to do or risk irrelevance.” The companies that lean into growth while embracing the discomfort, invest in their people, and build AI-first operations will emerge as the category winners in this unprecedented cycle of innovation and capital deployment. The infrastructure is being built. The models are improving rapidly. The question isn’t whether massive opportunities exist — it’s who will move fast enough to capture them.
Thank you again to all our amazing CEOs for joining us this year for “Accelerate 2025!”