Roger Barga, SVP of AI & ML at Oracle Cloud, is a senior infrastructure leader with prior roles at Microsoft Azure and Amazon Web Services. He is interviewed by George Hoyem, formerly of In-Q-Tel.
Key takeaways
- AI infrastructure scale and economics
- AI capability is increasingly defined by access to electricity and GPU-powered compute rather than traditional cloud metrics.
- Oracle positions itself as operating at comparable scale to hyperscalers, with rapid expansion of multi-gigawatt data centre capacity.
- Differentiation lies in performance-oriented infrastructure (e.g. bare metal, low-latency networking), delivering lower training costs for large models.
- Constraints shaping the AI market
- Four key bottlenecks: energy, chip supply, data availability and latency.
- Chip manufacturing is viewed as the most binding constraint due to supply chain concentration.
- Data scarcity (especially high-quality text) may drive greater reliance on synthetic data.
- Enterprise adoption and workforce impact
- AI is reshaping tasks rather than eliminating roles; developers increasingly focus on testing, documentation and system understanding.
- Productivity gains can unlock new projects rather than reduce headcount.
- Effective upskilling is peer-led (one-to-two levels above), rather than top-down transformation.
- AI posture management as a new risk frontier
- Enterprises lack visibility over employee AI usage, data flows and agent deployment, creating new security vulnerabilities.
- Governance frameworks (analogous to cloud or data security posture management) are still immature.
- Shifting software and business models
- Core systems (ERP, CRM) remain durable, but interfaces are changing via agents and natural language layers.
- Emerging models focus on delivering outcomes rather than software licences, enabled by multi-agent workflows.
- Implications for corporate venture investors
- Strong opportunity in AI security, governance and “posture management” tooling (e.g. red teaming, sandboxing).
- Startups are likely to outpace enterprises in addressing emerging AI risks, creating a clear investment gap.
- Infrastructure differentiation (cost, latency, scale) remains a critical competitive axis, with partners such as OpenAI and NVIDIA shaping demand.
This is an AI-generated summary, which has been lightly edited by GCV staff.


