Kandou AI raises $225 million to scale AI chip infrastructure
Kandou AI, a Swiss semiconductor startup pivoting toward artificial intelligence infrastructure, has raised $225 million in a funding round backed by SoftBank, Synopsys, Cadence Design Systems and Alchip Technologies.
The Lausanne-based company is now valued at about $400 million, marking a major step in its shift from consumer electronics to AI-focused hardware. Founded in 2011 as a spin-off from EPFL, Kandou originally specialized in high-speed connectivity for consumer devices before refocusing on data center and AI applications.
The strategic pivot accelerated after Srujan Linga took over as chief executive in 2025. The company rebranded as Kandou AI and expanded its focus to key segments of AI infrastructure, including inference, training systems, memory platforms based on CXL and rack-level connectivity.
Kandou’s core technology, Copper MIMO, has been deployed in more than 28 million chips globally since entering production in 2018. The company is now applying this expertise to address one of the main constraints in AI systems: data movement between processors.
The semiconductor industry faces a widening gap between computing power and interconnect speeds. While computing performance has increased dramatically over the past two decades, data transfer capabilities have lagged behind, creating bottlenecks that limit system efficiency. Kandou AI claims its low-power, high-bandwidth interconnect solutions can reduce inference system costs by up to 12 times.
The participation of major chip design firms such as Synopsys and Cadence highlights growing industry interest in solving these infrastructure challenges. The funding will be used to scale production of connectivity chips and strengthen partnerships with large data center operators.
With more than 350 patents and operations across Europe, North America and Asia, Kandou AI is positioning itself within a rapidly expanding segment of the semiconductor market focused on enabling large-scale AI deployment.
-
17:50
-
17:30
-
17:20
-
17:00
-
16:50
-
16:45
-
16:30
-
16:20
-
16:18
-
16:15
-
16:00
-
15:54
-
15:50
-
15:45
-
15:45
-
15:30
-
15:27
-
15:20
-
15:15
-
15:06
-
15:00
-
14:50
-
14:45
-
14:30
-
14:20
-
14:15
-
14:00
-
13:50
-
13:45
-
13:30
-
13:20
-
13:15
-
13:00
-
12:50
-
12:20
-
12:15
-
12:00
-
11:50
-
11:45
-
11:30
-
11:20
-
11:17
-
11:15
-
11:10
-
11:00
-
10:50
-
10:45
-
10:30
-
10:20
-
10:15
-
10:00
-
09:55
-
09:50
-
09:45
-
09:30
-
09:21
-
09:20
-
09:15
-
09:00
-
08:50
-
08:45
-
08:37
-
08:30
-
08:20
-
08:15
-
08:00
-
07:50
-
07:45
-
07:30
-
07:15
-
07:00