Table of Contents
In this episode, Mike Maniscalco, CEO of QumulusAI, shares how his background as a computer scientist and software entrepreneur led him into AI infrastructure and ultimately to QumulusAI. Mike explains how the rapid advancement of AI, particularly reasoning models and agentic tools, made one thing increasingly clear: more capable AI requires more tokens, and more tokens require more compute infrastructure. That realization brought him back to the infrastructure layer powering the AI revolution.
Mike discusses QumulusAI’s differentiated approach to delivering GPU compute through a distributed, inference-first model built around “hyperspeed.” Rather than focusing exclusively on massive, multi-year hyperscale campuses, the company identifies smaller pockets of existing powered data center capacity that can be upgraded and brought online faster. This allows QumulusAI to respond to customers who need compute in months rather than waiting years for new infrastructure.
The conversation explores why inference represents such a significant opportunity as AI moves from training models to putting them into production. Mike also breaks down QumulusAI’s FACTS framework (flexibility, access, cost, trust, and speed) and explains how flexibility across locations, vendors, architectures, and supply chains can help accelerate deployment.
Looking ahead, Mike believes AI infrastructure demand is still in its early innings. As multimodal AI, agentic applications, robotics, autonomous vehicles, and other emerging use cases consume more compute, he sees distributed infrastructure becoming increasingly important, especially as latency and location matter more. QumulusAI’s ambition is to evolve into a true AI cloud capable of moving compute at the speed of demand.
Important Links
- Connect with Mike Maniscalco on LinkedIn
- Qumulus AI company website
- Let’s Get Digital with Carrie Charles
- Broadstaffglobal.com
About
Michael Maniscalco is the Chief Executive Officer of QumulusAI. He brings a rare combination of operator discipline and hands-on AI infrastructure leadership, with experience building and scaling digital infrastructure businesses through periods of rapid demand.
Prior to QumulusAI, Maniscalco served as Chief Technology Officer of Applied Digital (NASDAQ: APLD) from July 2023 to January 2025, where he helped drive consecutive quarter and year-over-year revenue growth while leading the company’s transition from blockchain data center development and operations to HPC data centers and GPU-as-a-Service. He played a key role in securing major customer contracts and overseeing the deployment of next-generation facilities to support rising AI compute needs. Earlier, he served as EVP of Technology at Applied Digital beginning in September 2021.
A serial entrepreneur with multiple startup exits, Maniscalco founded ventures in chronic disease management, IoT and remote systems management, and luxury smart home systems. His prior product experience includes launching several award-winning software solutions. Additional roles include entrepreneur-in-residence at StanleyX, product development, and engineering leadership positions spanning the AI, IoT, smart home, blockchain, and healthcare spaces. He holds a B.S. in Computer Science from Georgia Tech, is an experienced emerging technology investor, keynote speaker, and actively mentors startups through programs that include TechStars and FAU Tech Runway.
