Kandou AI, a Swiss semiconductor company that builds chip-to-chip interconnect technology, has raised $225 million in what it calls a Series A round, led by Maverick Silicon with strategic participation from SoftBank, Synopsys, Cadence Design Systems and Alchip Technologies. The round values the company at $400 million. The label is worth pausing: Kandou was founded in 2011 and previously raised more than $163 million in Series B and C rounds under the name Kandou Bus. The “Series A” designation reflects a change in branding and leadership, not a new beginning.
The company’s new CEO, Srujan Linga, a former CEO of Goldman Sachs, replaced founder Amin Shokrollahi in 2025, an EPFL mathematics and computer science professor who invented the core technology. Shokrollahi’s contribution, a signaling method called Chord that sends correlated signals over multiple cables to increase bandwidth by a factor of two to four while halving power consumption, remains the technical foundation. The rebranding to Kandou AI and repositioning toward AI infrastructure is Linga’s doing, and it appears to have worked: The $225 million raise is the largest in the company’s history and puts SoftBank, one of the most aggressive AI infrastructure investors, on the cap table.
The bet against the light
What makes Kandou AI’s position unusual is not the problem it is trying to solve but the material with which it proposes to solve it. The AI industry’s interconnection bottleneck is real and well documented. As models scale to hundreds of billions of parameters and training sets expand to tens of thousands of GPUsThe speed at which data moves between processors and memory has become the binding limitation on performance. At signaling speeds of 224 gigabits per second, traditional copper interconnects consume approximately 30 percent of the total cluster power, with signal degradation so severe that range is limited to less than one meter without amplification.
The predominant industry response has been to move to optics. Ayar Labs raised $500 million in March 2026 at a $3.8 billion valuation for its co-packaged optical interconnects. Marvell completed a $3.25 billion acquisition of Celestial AI in February, purchasing photonic fabric technology that claims 25 times the bandwidth of copper alternatives at one-tenth the latency. The optical interconnection market for AI data centers It is projected to grow from $3.75 billion in 2025 to $18.36 billion in 2033.
Kandou AI is betting that copper is not finished yet. Its Chord signaling technology, the company claims, can achieve capacity efficiency on the way to Shannon, reducing power consumption and system costs by a factor of ten, while extending copper links to 448 gigabits per second and beyond. If that claim holds up, it would mean that the billions being spent on optical interconnection transitions are at least partially premature, and that existing copper infrastructure can support several more generations of hardware at a fraction of the cost.
Strategic investors tell the story
The composition of the investor syndicate matters more than the headline figure. Synopsys and Cadence are the two dominant providers of electronic design automation tools. Their involvement is not purely financial; It points to a possible integration of Kandou AI’s serializer/deserializer intellectual property into the design flows that chip architects use to build processors and memory controllers. Alchip, a Taiwanese ASIC design services company, offers a path to manufacturing. SoftBank, which has invested more than $100 billion in AI-adjacent companies through its Vision Fund and direct investments, adds scale capital and strategic network.
The practical implication is that Kandou AI technology could appear inside chips designed by other companies rather than requiring customers to adopt Kandou’s own silicon. This is a licensing and intellectual property model, similar in structure to Arm’s focus on mobile processors, and is a more capital-efficient path to market dominance than making and selling chips directly. The central question is whether Kandou can execute that model at a $400 million valuation and $225 million in fresh capital, versus optical competitors valued at ten times that much.
The valuation gap
At $400 million, Kandou AI is valued at about one-tenth of Ayar Labs and one-eighth of what Marvell paid for Celestial AI. That gap could reflect market skepticism about the longevity of copper in AI infrastructure, or it could reflect the fact that Kandou’s technology, if it works as claimed, does not require the industry to remove its existing wiring. Copper is already in all data centers. If Kandou’s signaling technology can make it fast enough for another generation of AI workloadsthe adoption curve would be faster and cheaper than an optical transition.
The risk is that “another generation” will not be enough. AI model sizes and training set scales are growing at a rate that consistently exceeds infrastructure predictions. What is adequate at 448 gigabits per second today may be inadequate at the terabits per second speeds that next-generation models will require in two or three years. Optical interconnects, despite their cost and complexity, offer a higher theoretical ceiling.
Kandou AI’s $225 million gives it time to prove that the roof can wait. The company’s 15-year history and the technical credibility of Chord signaling, which has been commercially implemented in consumer electronics since the mid-2010s, give substance to the bet. But the AI infrastructure market has a pattern of rewarding ambition over incrementalism, and a company that argues that the existing stuff is good enough faces a tougher narrative sell than one that promises to replace it entirely. Investors in this round appear to be banking on engineering pragmatism. Whether the market agrees will depend on how quickly the optical transition matures and whether Kandou copper can keep pace with an industry that has shown little interest in waiting for anything.
