Vertical Semiconductor has successfully secured $11 million in seed funding to tackle a major challenge in artificial intelligence: power-hungry data centers. Led by Playground Global, this investment will fuel the development of the company's innovative Vertical GaN AI chips, which promise to make AI more sustainable and efficient. Other investors in the round include Jimco Technology Ventures, Milemark-Capital, and Shin-etsu.
Tackling the AI Power Bottleneck
Modern AI data centers face a significant power delivery bottleneck. As AI workloads become more advanced, current chip designs struggle to provide enough power without generating excessive heat and consuming vast amounts of energy. This core problem not only limits the pace of AI innovation but also creates major operational and environmental challenges.
How Vertical GaN AI Chips Revolutionize Power Delivery
Vertical Semiconductor’s solution is a fundamental redesign of the power transistor. Instead of traditional lateral designs, the company creates Vertical Gallium Nitride (GaN) AI chips.
This innovative approach involves stacking transistors vertically, which provides several key advantages:
- Higher Density and Voltage: The vertical structure allows more transistors to be packed onto a single chip and supports much higher voltages, delivering more power in a smaller space.
- Improved Thermal Management: This design is far more effective at dissipating heat, which is a critical factor in high-performance data centers.
- Built-in Surge Protection: The technology enables a self-protection mechanism called "avalanche," allowing the chips to handle sudden power surges efficiently and safely.
Ultimately, this technology aims to improve overall data center efficiency by up to 30% and reduce the power footprint by half. The company has already demonstrated its technology on standard eight-inch wafers using conventional CMOS manufacturing methods, paving the way for scalable production.
What's Next on the Roadmap
With the new funding, Vertical Semiconductor is moving quickly to bring its product to market. The company plans to begin sampling packaged devices for early customers by the end of this year. Looking further ahead, a fully integrated solution is targeted for release before the end of 2026.
