Key Takeaways:
A $4.3 billion power-chip maker has quietly become the best-performing semiconductor stock of 2026 by solving a problem Nvidia's GPUs create: how to move electricity through AI data centers without wasting it.
Navitas Semiconductor has surged 148% year-to-date to about $17, far outpacing Nvidia's roughly 10% gain, as investors bet its one-stage power conversion technology can cut energy losses in AI data centers. The company, trading under the ticker NVTS on Nasdaq, now carries a market capitalization of about $4.3 billion.
"This closes the efficiency gap between high-voltage input and chip-level power delivery in a single step, which is what AI racks need as they scale to 800V architectures," a Morgan Stanley analyst wrote in May after raising the firm's price target on Navitas to $12.50 from $4.20.
In March, Navitas introduced an 800V-to-6V DC-DC power delivery board that eliminates the traditional 48V intermediate conversion stage. The company says the design maximizes system efficiency and reliability for Nvidia's AI infrastructure. Baird followed with a more aggressive target of $20, up from $9. The stock's 52-week range of $5.44 to $34.17 reflects how quickly expectations have shifted.
The bet is that power efficiency has become a bottleneck for AI scaling. Data centers consume enormous electricity, and every percentage point of conversion loss compounds across thousands of racks. Navitas, with a market cap of about $4.3 billion, is tiny compared with Nvidia's $4.7 trillion, but its technology addresses a cost problem that grows with every new GPU generation.
How GaN chips solve a data center problem
Navitas makes gallium nitride (GaN) power semiconductors, a material that switches faster and wastes less energy as heat than traditional silicon. Its new board converts 800 volts directly to 6 volts in one power stage, removing the intermediate 48V bus that adds components, heat, and space requirements. For data center operators, that translates into lower cooling costs and more usable computing density per rack.
The competitive landscape is shifting. While Nvidia, AMD, and Micron dominate the AI chip conversation, the infrastructure supporting those chips — power delivery, cooling, networking — is attracting a new wave of investment. Navitas sits at the intersection of two themes: AI infrastructure buildout and energy efficiency. Every major cloud provider, including Microsoft and Amazon, has pledged to reduce data center power consumption, and regulators in the European Union are tightening efficiency standards for large computing installations.
Valuation leaves no room for error
The stock's valuation has stretched dramatically. Navitas trades at about 92 times sales, compared with its five-year average of 11.8 times. That multiple implies the market has already priced in years of rapid revenue growth from AI data center design wins. If Navitas fails to convert its technology advantage into shipments at the pace investors expect, the stock could mean-revert sharply.
Morgan Stanley's $12.50 target sits below the current share price, suggesting some analysts see limited near-term upside. Baird's $20 target implies roughly 18% upside from current levels. The divergence reflects uncertainty about how quickly Navitas can scale production and win contracts with major data center operators.
For investors, Navitas represents a high-conviction bet on the power-efficiency sub-theme within AI infrastructure. The technology addresses a real bottleneck, and Wall Street's rapid target resets signal institutional interest. But at 92 times sales, the stock leaves no room for execution missteps. Investors considering a position may want to wait for a pullback or gain exposure through a broader semiconductor ETF that includes power-management names.
This article is for informational purposes only and does not constitute investment advice.
