FC Insights | July 2026
The Next Semiconductor Wave AI Shifts Demand, Specialization Defines the Market
Foreword
Semiconductors are entering a new growth phase as AI and robotics drive demand for higher hardware performance, while exposing constraints around power, memory, and energy supply. As innovation shifts from production scaling toward compute efficiency, new opportunities are emerging for asset-light, design-led companies.
In Southeast Asia, more companies are building around chip design, fabless models, and specialized architectures such as chiplets, photonics, and edge inference. We expect this to create more room for innovation and fundraising opportunities across the semiconductor ecosystem.
In Southeast Asia, more companies are building around chip design, fabless models, and specialized architectures such as chiplets, photonics, and edge inference. We expect this to create more room for innovation and fundraising opportunities across the semiconductor ecosystem.
FC Insights
The semiconductor industry is entering a new growth phase, as AI raises demand for higher hardware capability requirements, shifting innovation from pure production scaling toward greater compute efficiency. This shift creates new opportunities for asset-light, design-led companies, particularly in Southeast Asia, to capture value through specialized chip architectures, IP, and system-level design as the market moves toward chiplets, photonics, and edge inference.
We expect this shift to drive further innovation and fundraising opportunities across the semiconductor ecosystem.
How do you view the semiconductor and hardware space evolving?
The rise of AI and robotics has created unprecedented demand for semiconductors. This demand is now running up against real physical constraints, particularly power and memory, as energy supply struggles to scale at the same pace of AI compute growth. This requires continued innovation in hardware and processes to keep up.
Although incumbents are already present, there are still chances for newcomers to enter the market and fulfill specific demand requirements created by the ever-evolving developments of technology and AI.
Although incumbents are already present, there are still chances for newcomers to enter the market and fulfill specific demand requirements created by the ever-evolving developments of technology and AI.
Where do you see the potential opportunities across the semiconductor value chain?
An effective agent for enterprise-level is rarely a single piece of software, but a vertical stack consisting of multiple layers:
Why is Building an AI Agent an Infinity Loop?
Semiconductors have a broad value chain, from design to manufacturing to assembly, with specific niches addressed by different players. Each segment can be attractive to different types of investors, depending on their investment angle.
Specifically, we have recently seen chip design and fabless models gain more market attention. These design companies carry lower capex commitments and have a margin profile closer to that of software companies. Southeast Asia, in particular, has seen a growing number of teams building this design-led model, along with overseas founders setting up entities in the region.
Specifically, we have recently seen chip design and fabless models gain more market attention. These design companies carry lower capex commitments and have a margin profile closer to that of software companies. Southeast Asia, in particular, has seen a growing number of teams building this design-led model, along with overseas founders setting up entities in the region.
What are the key drivers behind this vertical?
We see a significant driver coming from AI. While AI chips represent a small share of total units produced, they contribute close to half of total chip industry revenue.
AI development is driving innovation toward chiplets, photonics, and edge inference. Chiplets divide a chip into smaller, specialized components rather than one piece doing everything, improving yield and energy efficiency. Photonics uses light instead of electricity to move data, easing power and heat bottlenecks. Edge inference shifts processing closer to the device, reducing reliance on data centers. Together, these shifts let design teams specialize in a narrower part of the value chain without owning the full stack.
What's the outlook from here?
We see continued innovation through up-and-coming models. We notice that design models are seeing more development in chiplets, photonics, and edge inference, while the supporting ecosystem around them is also growing. As more founders set up their businesses in Southeast Asia, we look forward to seeing more deals and advancements in this space.
At Favour Capital, we have served leading growth-stage technology companies across the region, supporting their regional and international expansion through equity financing. We help businesses prepare for capital raising and connect them with the right strategic partners to accelerate their growth and execute their expansion plans.
At Favour Capital, we have served leading growth-stage technology companies across the region, supporting their regional and international expansion through equity financing. We help businesses prepare for capital raising and connect them with the right strategic partners to accelerate their growth and execute their expansion plans.