Beyond the AI hype: Understanding Cloud Computing's Real Moats
Wed Dec 10 2025
In this Montaka Spotlight Series podcast, Andy sits down with Amit Nath to challenge the prevailing narrative that AI is in a bubble. As a prelude to Montaka's forthcoming white paper on AI infrastructure, they explore critical misperceptions about hyperscale cloud providers.
Perfect for: Investors, tech analysts, and anyone seeking to understand the real economics behind the AI infrastructure boom beyond the headlines.
Related Content: Watch for Montaka's comprehensive white paper on AI bubble arguments, coming soon.
Key topics covered:
00:00:33 - The Neo-Cloud Misconception: CoreWeave vs. True Hyperscalers
00:01:41 - Barriers to Entry: Are They Really Falling?
00:03:01 - Oracle's Gamble: A Case Study in Value Destruction
00:04:14 - Why Training Infrastructure ≠ Production Infrastructure
00:05:36 - The Fungibility Myth: When Cloud Infrastructure Matters
00:07:00 - Google's TPU Cost Advantage: 44% Cheaper Than Nvidia
00:08:36 - Breaking Free from the Broadcom Tax: MediaTek Partnership
00:12:47 - Software Moats: Attacking Nvidia's CUDA Dominance
00:13:05 - Inference vs. Training: Why Ecosystems Win
00:15:38 - The Dual-Use Advantage: Internal Demand De-Risks Buildouts
00:20:40 - Real Options: From Transformers to Quantum Computing
00:22:29 - Google vs. Nvidia: The Disruption Opportunity
Tags: AI infrastructure, hyperscalers, cloud computing, Google TPU, Nvidia, Microsoft Azure, Amazon AWS, neo-clouds, CoreWeave, Oracle, OpenAI, AI bubble, investment analysis, technology investing, semiconductor industry, R&D accounting, depreciation, quantum computing, PyTorch, CUDA, inference vs training, Montaka
Disclaimer: Issued by Montaka Global Pty Ltd ABN 62 604 878 533, AFSL 516942. This information is general in nature and does not take into account your specific needs or circumstances. You should consider your own financial position, objectives and requirements and seek professional financial advice before making any financial decisions.
More
In this Montaka Spotlight Series podcast, Andy sits down with Amit Nath to challenge the prevailing narrative that AI is in a bubble. As a prelude to Montaka's forthcoming white paper on AI infrastructure, they explore critical misperceptions about hyperscale cloud providers. Perfect for: Investors, tech analysts, and anyone seeking to understand the real economics behind the AI infrastructure boom beyond the headlines. Related Content: Watch for Montaka's comprehensive white paper on AI bubble arguments, coming soon. Key topics covered: 00:00:33 - The Neo-Cloud Misconception: CoreWeave vs. True Hyperscalers 00:01:41 - Barriers to Entry: Are They Really Falling? 00:03:01 - Oracle's Gamble: A Case Study in Value Destruction 00:04:14 - Why Training Infrastructure ≠ Production Infrastructure 00:05:36 - The Fungibility Myth: When Cloud Infrastructure Matters 00:07:00 - Google's TPU Cost Advantage: 44% Cheaper Than Nvidia 00:08:36 - Breaking Free from the Broadcom Tax: MediaTek Partnership 00:12:47 - Software Moats: Attacking Nvidia's CUDA Dominance 00:13:05 - Inference vs. Training: Why Ecosystems Win 00:15:38 - The Dual-Use Advantage: Internal Demand De-Risks Buildouts 00:20:40 - Real Options: From Transformers to Quantum Computing 00:22:29 - Google vs. Nvidia: The Disruption Opportunity Tags: AI infrastructure, hyperscalers, cloud computing, Google TPU, Nvidia, Microsoft Azure, Amazon AWS, neo-clouds, CoreWeave, Oracle, OpenAI, AI bubble, investment analysis, technology investing, semiconductor industry, R&D accounting, depreciation, quantum computing, PyTorch, CUDA, inference vs training, Montaka Disclaimer: Issued by Montaka Global Pty Ltd ABN 62 604 878 533, AFSL 516942. This information is general in nature and does not take into account your specific needs or circumstances. You should consider your own financial position, objectives and requirements and seek professional financial advice before making any financial decisions.