GV’s Dave Munichiello On Qualcomm’s Modular Purchase, The Firm’s 10x Return And The Shift In AI Software

2026-06-30 · Source: Artificial intelligence - Crunchbase News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

GV Managing Partner Dave Munichiello discusses significant shifts in the AI infrastructure landscape. These are highlighted by Qualcomm's acquisition of Modular and SambaNova's \$800 million funding round, valuing it at \$10 billion. These developments underscore the increasing value of software layers. These layers enable AI models to run efficiently across diverse hardware, including AI-specific chips, CPUs, and GPUs, in a trend called "disaggregated inference." Munichiello, an early investor in both Modular and SambaNova, notes that hardware remains scarce and expensive. The focus is now on maximizing efficiency and squeezing value from existing chips. He adds that open-source models further expand the market. This allows enterprises to run models on their own hardware. Despite consolidation by tech giants, Munichiello believes a path to independent IPOs remains viable for startups. He cites 15-20 companies planning to go public in the next six months. His investment philosophy prioritizes fundamental technologies and supporting companies through "crucible moments."

Key takeaway

For Directors of AI/ML managing infrastructure costs, recognize that software layers enabling efficient model execution across diverse hardware are now paramount. Your strategy should prioritize solutions that optimize performance across heterogeneous compute environments, including AI-specific chips, CPUs, and GPUs. Consider adopting open-source models to gain control over inference and reduce reliance on external providers. This approach can significantly lower your total cost of ownership and mitigate hardware scarcity risks, positioning your organization for greater agility and long-term cost efficiency.

Key insights

AI software layers are gaining value, enabling efficient model execution across heterogeneous hardware amidst scarcity and the shift to disaggregated inference.

Principles

• AI inference requires heterogeneous chips.
• Software efficiency optimizes scarce hardware.
• Open-source models decentralize inference.

In practice

• Implement disaggregated inference for TCO.
• Optimize software for chip efficiency.
• Adopt open-source models for internal inference.

Topics

• AI Infrastructure
• Modular AI Acquisition
• Disaggregated Inference
• Open-Source AI Models
• AI Software Optimization
• Venture Capital Strategy

Best for: CTO, VP of Engineering/Data, AI Architect, Investor, Entrepreneur, Director of AI/ML

Related on AIssential

• The AI Industry Spent Billions Chasing Faster Chips. Inference Startup Kog Says They Were Solving the Wrong Problem. — Software optimization can enable standard GPUs to match specialized AI hardware performance for real-time inference.
• not much happened today — AI development is shifting towards multi-cloud distribution, cost-aware agentic workflows, and specialized hardware for inference.
• Token-Operations-Oriented Inference Optimization Techniques for Large Models — Token-operations-oriented inference optimization uses a four-layer architecture to make large model services operable.
• "But OpenClaw is expensive..." — A hybrid LLM architecture combining cloud frontier models with local open-source models significantly cuts costs and boosts privacy.
• Scaling Agentic Inference Across Heterogeneous Compute with Zain Asgar - #757 — Heterogeneous AI inference across diverse hardware optimizes cost and performance for token-intensive agentic workloads.
• [AINews] not much happened today — AI progress spans non-invasive BCIs, efficient inference, agent orchestration, and open-weight model commercialization, driven by infrastructure and evaluation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial intelligence - Crunchbase News.