In this episode of the MI&S IT Talk Podcast, Matt Kimball and Jason Andersen dig into two of the most important (and misunderstood) shifts happening right now in enterprise tech: agent platforms and custom silicon.
The conversation moves beyond hype and into what’s actually changing — how AI agents are forcing a rethink of software platforms, and why the explosion of specialized chips may reshape the data center, but not in the way most people expect.
From Salesforce’s Agentforce strategy to the rise of hyperscaler silicon from Amazon Web Services, Google, and Microsoft, this episode connects the dots between AI infrastructure, enterprise workflows, and where control—and complexity—will ultimately land.
Key topics covered:
- The emergence of agent platforms as a new enterprise software layer
- Why agents are fundamentally different from traditional applications
- How Salesforce is positioning Agentforce as a platform play
- The shift from user-based pricing to consumption-based AI models
- The growing divide between deterministic systems vs. probabilistic AI agents
- Why agent orchestration is still an open question (LLM vs. control plane vs. platform layer)
- How AI is reshaping product management workflows (vibe coding, rapid prototyping, feedback loops)
- The risk of confirmation bias in LLM-driven workflows
- The rise of custom silicon and chip diversity in AI infrastructure
- New players like SiFive and the push behind RISC-V
- Hyperscaler silicon strategies (TPU, Trainium, Inferentia, etc.)
- Whether silicon fragmentation is a long-term trend or short-term bubble
- The future of abstraction layers—and why hardware may become invisible
- The emerging opportunity for a control plane for heterogeneous compute environments
Watch the full episode here:
Disclaimer: This show is for information and entertainment purposes only. While we will discuss publicly traded companies on this show. The contents of this show should not be taken as investment advice.
