- AngelsRound
- Posts
- Madrone ❄️
Madrone ❄️
Cooling infrastructure for AI data centers
Spotlight
What if the companies enabling AI become more valuable than the companies building AI?
Quick Pitch: Madrone helps AI data centers unlock more compute by replacing traditional chillers with a cooling system that uses less power and water.
The Problem
Power Is the Limit: AI data centers are running out of electricity before they run out of GPUs.
Cooling Steals Compute: Up to 30% of available power is spent keeping servers cool.
The Grid Moves Slowly: New power connections can take years to secure.
Every Watt Counts: Every watt spent on cooling is one less watt available for AI.
Why It Matters
For years, AI infrastructure was framed as a chip story. That story is changing. GPUs can be bought, but power cannot be added overnight. Every watt spent on cooling is a watt that cannot be used to run AI.
This shifts innovation beyond models and chips toward the infrastructure that powers them. As AI scales, physical constraints become just as important as software breakthroughs.
Snapshot
Industry: AI Infrastructure
Headquarters: San Francisco, CA
Founded: 2025 (YC P26)
Investors: Y Combinator
Funding: Raising $5M (Seed)
Traction: Rapid prototype development, a paying pilot, early customer commitments, and active discussions with hyperscalers.
Founder’s Edge
Erik Meike spent four years designing thermal systems for Apple's iPhone hardware before building robots for data centers.
Akshay Trikha combined materials science research at Berkeley with machine learning infrastructure experience at QuantumScape.
What Users Love
Free more electricity for revenue generating GPUs.
Lower both capital and operating costs.
Reduce water usage without sacrificing performance.
Deploy faster than conventional cooling systems.
Playing Field
Mechanical Chillers: Proven but energy intensive.
Cooling Towers: Efficient but water hungry.
Adiabatic Cooling: Lower cost but less effective in extreme heat.
Madrone's Edge: Combines lower power and water usage without relying on mechanical chillers.
Analysis
Bulls Case 📈
AI is exposing infrastructure as the next bottleneck.
Customers gain more compute from existing power.
The market extends beyond data centers.
AI infrastructure spending continues to accelerate.
Bears Case 📉
Customers will require years of reliability.
Hardware is difficult to scale.
Incumbents have entrenched relationships.
AI infrastructure spending could slow.
Verdict
The real opportunity is not better cooling. It is changing how cooling is valued.
Historically, cooling was a cost to minimize. AI turns it into a way to increase compute without adding more power. Madrone's success depends on whether buyers begin measuring cooling by the compute it unlocks rather than the energy it saves.
Who's Hiring — Benchmark Portfolio Startups Right Now
Benchmark backed eBay, Uber, Twitter, Snap, Discord, Dropbox, and Instagram.
Cerebras — AI chips — hiring across hardware, software, and infrastructure
Starcloud — Space compute — hiring across engineering and operations
Sierra — AI customer agents — hiring across engineering and GTM
Mercor — AI recruiting platform — hiring across engineering and operations
Fireworks AI — AI inference — hiring across engineering
The Startup Pulse
Another happening week in startup funding. Three signals from this week:
AI is moving deeper into healthcare operations
Enterprise software is becoming increasingly AI native
Infrastructure remains one of the strongest investment themes
Assort Health — Raised $120M Series C to scale AI agents that automate patient scheduling, intake, and care coordination across healthcare systems.
Taktile — Secured $110M Series C to help banks and fintechs automate credit, fraud, and risk decisions using AI.
Runpod — Raised $100M Series A to expand GPU cloud infrastructure powering AI training and inference.
You’re gonna want to see this live
On July 16th at 1PM ET, beehiiv is unveiling the next chapter for audience-led businesses.
For years, creators and brands have been forced to stitch together bloated stacks of tools just to publish content, grow an audience, and make money online.
Newsletters in one platform. Websites in another. Podcasts somewhere else. Analytics scattered everywhere.
beehiiv thinks there’s a better way. Now, they’re ready to show it off at their Summer Release Event.
Written by Ashher
