Female-founded Semiconductor Ai Startup Sixsense Raises $8.5m

A Singapore-based heavy tech startup called SixSense has developed an AI-powered level that helps semiconductor manufacturers foretell and observe imaginable spot defects connected accumulation lines successful existent time.

It has raised $8.5 cardinal successful Series A bringing its full backing to astir $12 million. The information was led by Peak XV’s Surge (formerly Sequoia India & SEA), pinch information from Alpha Intelligence Capital, Febe, and others.

Founded successful 2018 by engineers Akanksha Jagwani (CTO) and Avni Agarwal (CEO), SixSense intends to reside a basal situation successful semiconductor manufacturing: converting earthy accumulation data, from defect images to instrumentality signals, into real-time insights that thief factories forestall value issues and amended yield.

Despite nan sheer measurement of information generated connected nan fab floor, what stood retired to nan co-founders was a astonishing deficiency of real-time intelligence.

Akanksha brings a heavy knowing of manufacturing, value control, and package automation done her acquisition building automation solutions for manufacturers for illustration Hyundai Motors and GE and led merchandise improvement astatine startups for illustration Embibe. Agarwal adds method acquisition from her clip astatine Visa, wherever she built large-scale information analytics systems, immoderate of which were later protected arsenic waste and acquisition secrets. A skilled coder pinch a beardown inheritance successful mathematics, she had agelong been willing successful applying AI to accepted industries beyond fintech.

image credits: sixsense

Together, nan duo evaluated sectors from aviation to automotive earlier landing connected semiconductors. Despite nan semiconductor industry’s estimation for precision, inspection processes stay mostly manual and fragmented, Agarwal told TechCrunch. After speaking pinch much than 50 engineers, it became clear there’s important room to modernize really value checks are done, she added.

Fabs coming are filled pinch dashboards, SPC charts, and inline inspection systems, but astir only show information without further analysis, Agarwal said. “The load of utilizing it for decision-making still falls connected engineers: [they must] spot patterns, analyse anomalies, and trace guidelines causes. That’s time-consuming, subjective, and doesn’t standard good pinch expanding process complexity.”

SixSense provides engineers pinch early warnings to reside imaginable issues earlier they escalate pinch capabilities specified arsenic defect detection, guidelines origin analysis, and nonaccomplishment prediction.

SixSense’s level is besides specifically designed to beryllium utilized by process engineers alternatively than information scientists, Agarwal said. “Process engineers tin fine-tune models utilizing their ain fab data, deploy them successful nether 2 days, and spot nan results — each without penning a azygous statement of code. That’s what makes nan level some powerful and practical.”

The competitory scenery includes in-house engineering teams utilizing devices for illustration Cognex and Halcon, inspection instrumentality makers integrating AI into their systems, and startups including Landing.ai and Robovision.

SixSense’s AI level is already successful usage astatine awesome semiconductor manufacturers for illustration GlobalFoundries and JCET, pinch much than 100 cardinal chips processed to date. Customers person reported up to 30% faster accumulation cycles, a 1–2% boost successful yield, and a 90% simplification successful manual inspection work, nan founders said. The strategy is compatible pinch inspection instrumentality that covers complete 60% of nan world market.

“Our target customers are large-scale chipmakers — including foundries, outsourced semiconductor assembly and trial providers (OSATs), and integrated instrumentality manufacturers (IDMs),” Agarwal said. “We’re already moving pinch fabs successful Singapore, Malaysia, Taiwan, and Israel, and are now expanding into nan U.S.”

Geopolitical tensions, particularly betwixt nan U.S. and China, are reshaping wherever chips are made, driving caller manufacturing investments crossed nan globe.

“We’re seeing fabs and OSATs grow aggressively successful Malaysia, Singapore, Vietnam, India, and nan U.S. — and that’s a tailwind for us. Why? Because we’re already based successful nan region, and galore of these caller accommodation are starting caller — without bequest systems weighing them down. That makes them acold much unfastened to AI-native approaches for illustration ours from time one,” Agarwal told TechCrunch.