Syenta Raises $26M to Cut AI Chip Packaging Steps by 40%

SYDNEY: Syenta has raised $26 million in a Series A round to commercialise its advanced chip packaging technology for AI data centres. Playground Global led the round alongside Australia’s National Reconstruction Fund, which contributed $10.1 million. Investible, Salus Ventures, Jelix Ventures, Wollemi Capital, Blackbird VC, and In-Q-Tel also participated. Total funding now stands at $36.4 million across all rounds.

Pat Gelsinger is joining the board as part of the deal. The former Intel CEO is now a general partner at Playground Global. His involvement signals more than financial backing, Gelsinger spent four decades in semiconductors and led Intel through its most recent attempt to reclaim chip manufacturing leadership. His read on where packaging bottlenecks are most acute carries weight.

Also read: Dutch Quantum Startup QuantWare Lands $178M for Chip Production

Syenta’s technology targets one of the least visible but most consequential constraints in the AI hardware stack. Advanced chip packaging, the process of connecting multiple chips into a single high-performance system, has emerged as a critical bottleneck as AI workloads scale.

Current packaging methods are slow, expensive, and require large numbers of complex manufacturing steps. Syenta’s Localized Electrochemical Manufacturing technology cuts that process by 40%. It enables high-density chip-to-chip interconnects within existing semiconductor fabrication infrastructure, without requiring new fabs.

The company spun out of the Australian National University and is headquartered in Sydney. As part of its commercialisation push, Syenta is opening a US office in Arizona, deliberately positioned near Intel and TSMC facilities. That proximity matters for a company selling into a supply chain that is geographically concentrated. It targets high-volume production by 2028.

Gelsinger framed the problem directly, AI’s next scaling challenge is not just compute, it is how chips connect. Syenta’s bet is that solving the interconnect layer unlocks headroom that more GPUs alone cannot provide.