Podcast Recap Episode 21: An Investor's Take on AI's Next Wave with Louis Toth
April 24, 2026
AI Factory
TL;DR
AI is still in its early innings, but it’s already reshaping infrastructure, business models, and speed of innovation.
The real opportunity lies at the edge where data is generated and processed.
Startups that combine strong tech differentiation with fast go-to-market execution will win.
Introduction
In this episode of the Aviz Podcast, Louis Toth shares his perspective on where AI is heading and why the edge is becoming the most critical layer in the stack. As AI adoption accelerates across industries, infrastructure is being rebuilt from the ground up.
This conversation dives into what’s driving that shift, why networking is becoming central again, and how startups can navigate a rapidly evolving market. If you’re building, investing, or operating in AI, this breakdown highlights what matters most right now and what will define success going forward.
Why is “AI meets the edge” becoming a critical investment theme?
AI’s next wave is being shaped at the edge where data is generated, processed, and secured.
Traditionally, data was pushed into the cloud to improve user experience. Now, massive volumes of data are being created at the edge through sensors, cameras, and mobile devices. Instead of blindly sending everything to the cloud, the edge is being rebuilt to capture, process, organize, and secure data locally.
This shift is creating a full-stack opportunity across semiconductors, hardware, operating systems, cybersecurity, DevOps, and AI/ML. The edge is no longer just a delivery layer it is becoming a core compute and intelligence layer.
Mini Framework: Edge Opportunity Stack
- Data generation: sensors, devices, transactions
- Processing: local compute at edge
- Infrastructure: hardware + networking
- Intelligence: AI/ML integration
- Security + transport
Where is AI actually going beyond the hype?
AI is a long-term secular trend with real products, real revenue, and an expanding ecosystem.
While there is noise and hype, AI is already driving tens of billions in chip sales and creating an entire ecosystem beyond GPUs. This includes infrastructure needed to build AI factories, improve models, and deploy them across enterprises and consumer use cases.
However, business models, cost-benefit trade-offs, and long-term value are still being tested. The opportunity is real, but the winners are still being determined.
Checklist: What defines real AI traction
- Actual products being sold
- Clear enterprise use cases
- Measurable cost or revenue impact
- Scalable infrastructure ecosystem
Why does networking matter again in the AI era?
AI infrastructure cannot scale without networking, making it a critical but underappreciated layer.
AI factories are often associated with GPUs, but they depend on a much broader system that includes switches, optical connectors, servers, and software. Networking enables data movement, efficiency, and scalability across these systems.
With increasing data volumes, computational intensity, and global connectivity, networking is becoming central again. It also plays a key role in reducing costs by enabling open ecosystems instead of proprietary systems.
Table: AI Infrastructure Components
| Layer | Role |
|---|---|
| Compute | GPUs and chips |
| Networking | Data movement and connectivity |
| Storage | Data management |
| Software | Model training and deployment |
| Edge | Data capture and processing |
Misconception: “AI is just about models and GPUs.”
Correction: AI is an entire infrastructure ecosystem, not just models.
Focusing only on models ignores the broader system required to make AI work at scale. From real estate and hardware to networking and software, every layer contributes to performance, cost, and usability.
Companies that understand and build across this ecosystem are better positioned to capture long-term value.
Key Takeaway
AI is not just a technology shift; it is an infrastructure transformation. The edge is becoming a central layer for innovation, and networking is re-emerging as a critical enabler.
Startups that combine deep technical differentiation with fast execution and clear go-to-market strategies will have the advantage.
The pace of AI innovation is accelerating rapidly, forcing startups to rethink strategies faster than ever. Capabilities that seemed impossible are becoming feasible within months.
This means founders must build for where the market will be in the next 3 to 7 years, not where it is today. Sustainable value will come from solving must-have problems that enterprises cannot ignore, especially in tighter economic conditions.
Conclusion
AI is still early, but already driving massive ecosystem growth The edge is becoming a key layer for data and intelligence Networking is critical to scaling AI infrastructure Speed, differentiation, and business models will define winners Startups must adapt quickly to a rapidly changing landscape
Frequently Asked Questions
1. Where is AI truly going besides the hype?
Although we're still in the very early days of AI, it's already leading to real products, revenue and a whole ecosystem growing. The long-term story will be infrastructure, deployment and adoption in enterprise.
2. Why does “AI meets the edge” matter more now?
The majority of data is born at the edge with devices, sensors and applications, thus bringing intelligence to where data lives will increase performance, speed, security, while decreasing cost.
3. What is changing at the edge compared to traditional cloud models?
Rather than bringing all of the data to the cloud, all of the data processing and logic is happening on the edge, and the edge is moving beyond simply the "last mile" to becoming a central compute and intelligence layer.
4. Why is the edge a full-stack opportunity?
Developing AI at the edge will require a system that is dependent on hardware, networks, operating system and security which will create a huge opportunity within semiconductors, Dev Ops, AI/ML, and infrastructure.
5. How can enterprise know if they should embrace an AI trend or is it just hype?
Traction will show up in the form of shipped product, clear use cases and undeniable business results. Strong adoption will be dependent on scalability of the infrastructure itself, not just models.
6. Why is networking becoming critical again in the AI era?
AI systems are utterly dependent on very fast, efficient movement of data within compute, storage, edge layers. An AI factory can't be scale efficient and affordable without strong networks.
7. What is the greatest myth surrounding the infrastructure of AI?
It appears that people often view AI as models and GPUs alone; however this is only a part of the equation. This would neglect a complex ecosystem which must work seamlessly to produce value.
Explore how your infrastructure strategy aligns with the evolving AI ecosystem and identify where your biggest opportunities lie in the edge and networking layers.
