GaN epitaxial growth (MOCVD)

Every gallium nitride device India makes for its radars and satellites begins on a machine India cannot yet build. As of 2025, 100% of the MOCVD reactors used to grow GaN's crystalline layers — along with the ultra-pure precursor gases they consume — are imported from the United States, Japan, Germany and Europe.

Metal-organic chemical vapour deposition is the dominant process for growing GaN epitaxial layers, and it is unforgiving. A reactor holds trimethylgallium and ammonia at 950–1100°C while regulating gas flow, pressure and temperature tightly enough to deposit atomically uniform films across a wafer. The chemistry is precise; the tolerance for error is small. This is why only a handful of nations do it well. Japanese firms lead on wafer uniformity, China accounts for over 60% of GaN MOCVD deployments in Asia-Pacific, and the world's reactor supply comes from roughly six to eight equipment makers.

India's capability is real but early. DRDO's Solid State Physics Laboratory announced in November 2024 the design and fabrication of GaN HEMTs up to 150W and MMICs up to 40W at X-band — the frequencies that matter for radar and electronic warfare. DRDO's GAETEC centre fabricates GaN MMICs for RISAT-series radar satellites, running batches on an 80-day cycle. IISc Bangalore operates a GaN fab at research scale. All of this is genuine process mastery. None of it is production-scale epitaxy: there is no private commercial GaN manufacturing facility in India, and GAETEC relies on materials from SSPL and imported substrates rather than its own MOCVD growth.

The structural reason the gap persists sits beneath the device. Epitaxy is the single bottleneck for all GaN power, RF and optoelectronic supply chains, and it depends on capabilities India does not yet hold: the reactor itself (a single GaN MOCVD system costs $1.5–3 million), the >99.99999%-purity precursor gases, and a skilled workforce concentrated in Japan, the US and South Korea. Around 12% of production units in developing economies report a shortage of engineers able to run this equipment.

Movement has begun. On 5 May 2026 the Cabinet approved Crystal Matrix Limited to build an integrated compound semiconductor fab at Dholera, Gujarat, offering GaN foundry services with epitaxy on 6-inch wafers. It is at pre-construction stage.

What closing the gap would take is visible in that dependency list: domestic reactor access, indigenised precursor chemistry, and a trained epitaxy workforce built alongside the fabs — the foundations that turn demonstrated device design into homegrown crystal growth.