GaN Market Moves: Navitas's Strong Earnings and India's First Mass Production Point to the Next Frontier

Revenue Down 39%, Stock Up 27% – What This Paradox Reveals

In the first quarter of 2026, Navitas Semiconductor reported revenue of $8.6 million, a 39% decrease year-over-year. Despite this, the stock price surged 27.66% immediately after the announcement. While these figures may seem contradictory at first glance, they vividly illustrate what is currently happening in the GaN market.

What the market valued was not the absolute sales figures, but the shift in the origin of revenue. Navitas is intentionally scaling back its low-margin mobile and consumer business to focus on high-power segments such as AI data centers, grid energy infrastructure, and industrial electrification. Revenue from these segments increased by approximately 35% year-over-year. The structure of shrinking overall revenue while growing in key battlegrounds was evidently recognized.

Later that week, another piece of news emerged. Navitas announced a partnership with India's Cyient Semiconductors to launch India's first GaN power IC family. This collaboration, aligned with the Indian government's "Make in India" policy, raises questions across the entire supply chain as it explores a localized model for GaN.

Why AI Servers Have Become the Main Arena for GaN

The power consumption of AI servers is orders of magnitude higher than that of conventional servers. AI servers equipped with high-performance GPUs and large-capacity memory can reach a total power supply wattage of up to 18kW. For data center designers, this power density issue directly impacts equipment costs. The choice lies between increasing cooling capacity or improving conversion efficiency. GaN has emerged as a device to enable the latter.

The value GaN devices bring to power systems can be summarized in three key areas: efficiency, power density, and thermal management. Of particular interest is the ongoing transition to an 800V DC architecture in AI data centers. A leap from the traditional 48V/54V power supply to 800V requires high-voltage compatibility in a compact form factor. GaN achieves this while also enabling efficiency improvements through high-speed switching. The potential for approximately 30-50% size reduction compared to silicon devices with the same performance is a significant factor for data center operators looking to increase rack density.

Compared to silicon devices with the same output specifications, adopting GaN can reduce the size of the power supply unit by 30-50%. This is often cited as a design guideline. With AI server racks potentially consuming up to 18kW, this miniaturization capability is a crucial consideration for data center operators aiming to maximize rack density.

However, miniaturization is not the sole criterion for selecting power supply units. Stability and reliability to prevent power fluctuations and interruptions are equally important evaluation factors. The extent to which the reliability of GaN devices has been proven will be a key determinant in actual design-in decisions as their adoption expands.

However, the selection criteria for AI server power supplies are not solely based on miniaturization. Stability and reliability, which prevent power fluctuations and interruptions, remain equally important evaluation axes. As the adoption of GaN devices broadens, the extent to which their reliability has been demonstrated will heavily influence design-in decisions.

India's First Mass Production Signals Supply Chain Upheaval

It would be premature to view the partnership between Navitas and Cyient Semiconductors as merely an entry into an emerging market. The significance of this move can be understood on two levels.

The first level concerns the market itself. India is simultaneously facing rapidly expanding power infrastructure demand, renewable energy investments, and EV adoption policies, emerging as a promising demand hub for GaN devices. Local manufacturing under the "Make in India" policy can serve as a viable option for building a business in India, considering tariff structures and procurement risks.

The second level addresses the issue of diversifying the supply structure. Historically, the manufacturing of GaN-on-Si has been concentrated in specific regions and foundries. Establishing a manufacturing base in India also aligns with efforts to diversify supply risks. As the GaN market accelerates its growth, a pattern of geographical diversification is emerging across multiple companies and regions as a response to the risks inherent in supply chains dependent on specific locations.

Automotive GaN Nears Mass Production – The 2027-2028 Milestone

Alongside AI servers, this week's developments reaffirm the progress in automotive GaN. Automakers are anticipating the integration of GaN-on-Si devices into mass-produced models between 2027 and 2028. From a technical and certification perspective, this timeframe signifies that the countdown has already begun.

GaN-on-Si is first being adopted in automotive power systems, such as onboard chargers (OBCs) and DC/DC converters. OBCs are directly linked to EV charging efficiency, while DC/DC converters handle power conversion to the 12V system. Both applications have high demands for efficiency, miniaturization, and weight reduction, aligning perfectly with the characteristics of GaN-on-Si.

However, the hurdle of automotive certification is significant. Obtaining AEC-Q101, the reliability test standard for automotive electronic components, is the de facto gateway for GaN-on-Si devices to be integrated into mass-produced vehicles. The required operating temperature range is from -40°C to over 150°C, and in some cases, up to 175°C. Furthermore, resistance to tens of thousands of power cycles is also demanded.

This limited short-circuit withstand time is a critical issue for GaN-on-Si that directly impacts design decisions. While silicon IGBTs offered several microseconds of margin for short-circuit protection circuit design, GaN-on-Si requires assembly within a constraint of less than 1 microsecond. Challenges such as charge trapping (where charge accumulates between the gate and drain, causing device characteristics to fluctuate), gate stability, thermomechanical fatigue, and moisture resistance are particularly stringent in automotive environments. To target mass production in 2027-2028, addressing these technical challenges will be the core determinant in supplier selection.

The Path to the GaN Market in 2035 – Interpreting the Numbers

GaN semiconductor market size forecasts vary significantly among research institutions, with figures differing based on whether only power GaN is considered or if optoelectronics are included. Yole's projections estimate the power GaN device market to reach approximately $3 billion by 2030 (CAGR 42%), while broader market definitions suggest figures exceeding $10 billion by 2035. Rather than absolute values, understanding the "quality of growth" – which segments will be the drivers – is more beneficial for decision-making.

However, as Navitas's recent earnings indicate, this growth is not a straight line. The high-growth phase for consumer applications has peaked, and a transition period is underway, with the next wave driven by AI infrastructure, automotive, and industrial sectors, leading to restructuring of revenue streams. Navitas's net loss of $33.8 million in Q1 2026 (an increase from $16.8 million in the previous year) while maintaining approximately $221 million in cash on hand to drive this structural transformation can be viewed as a strategy that accounts for these transition costs.

A similar interpretation can be applied to the overall market. While a growth rate exceeding 20% is attractive, only players who have strategically positioned themselves early in the appropriate segments can capture it. For AI server power supplies, the speed of ramp-up is crucial, while for automotive, the timing of certification acquisition becomes the competitive axis.

How to Proceed – Decision Frameworks by Segment

The movements this week clearly indicate that the "next major battleground" for the GaN market has become defined. Regardless of whether one is involved in design, procurement, technology planning, or business development, considering the following structure will facilitate a clear decision-making process.

For automotive applications, the AEC-Q101 acquisition status and the track record of addressing inherent challenges such as charge trapping and short-circuit withstand time will serve as concrete metrics for evaluating supplier capabilities. It is worthwhile to verify the number of proven devices with acquired certifications and the existence of mass production adoption cases in OBCs and DC/DC converters, both technically and from a procurement standpoint.

For AI server power supplies, the core selection criteria will center on the specifications for 800V DC compatibility and the implementation benefits achievable through miniaturization and higher efficiency of the power supply units. As multiple suppliers accelerate their entry into this market, evaluation criteria should include not only spec comparisons but also production supply capacity and price outlooks.

The move towards manufacturing in India may not have an immediate impact on procurement decisions. However, in the context of diversifying supply chains, the question of "which region's foundry is manufacturing" is beginning to hold significance for medium-to-long-term risk management. The progress of the Navitas-Cyient partnership is worth tracking as a key indicator.