Navitas Semiconductor disclosed a definitive agreement Tuesday to acquire Claros, Inc. for up to $232.8 million, marking the GaN power semiconductor firm's largest acquisition and a direct push into the precision analog layer beneath AI compute workloads. Shares rose in premarket trading. The deal price includes earnouts tied to revenue milestones, though Navitas did not disclose the base cash component or timing assumptions.

Claros brings analog mixed-signal IP focused on high-voltage power management for data centers, a domain where Navitas already sells gallium nitride power ICs into server and telecom racks. The combined entity targets an $8 billion addressable market opportunity spanning AI training clusters, inference servers, and next-generation power delivery architectures. Navitas CEO Gene Sheridan said the acquisition "accelerates our roadmap by 18 to 24 months" in prepared remarks, suggesting the company was building similar capability internally but opted to buy velocity. Claros revenue run-rate was not disclosed, nor were customer overlap details.

The move matters because power efficiency is now the binding constraint in AI infrastructure buildouts. Hyperscalers are designing racks that draw 100 kilowatts per cabinet, double the prior generation, and every point of conversion loss compounds cooling costs and limits rack density. GaN devices switch faster and waste less energy than silicon MOSFETs, but integrating them with precision analog control loops requires IP Claros already ships. Navitas is betting that owning both layers—the GaN power stage and the analog control fabric—positions it for sole-source design wins at Meta, Microsoft, and the Tier 2 cloud builders now scrambling to add capacity. The company has not named customers, but its prior investor presentations listed "hyperscale data centers" as a $3.2 billion segment by 2027.

The acquisition also signals compression in the power semiconductor M&A market. Claros was venture-backed, likely in late-stage rounds, and exiting to a public acquirer at $232.8 million suggests limited appetite for another funding round or standalone IPO. Navitas itself has traded between $2.50 and $4.20 over the past year, giving it a currency for deals but not the balance sheet of an Analog Devices or Texas Instruments. The earnout structure protects Navitas if Claros customer traction lags, but it also implies the seller saw limited alternatives. Worth noting: Navitas went public via SPAC in 2021, and this is its first major use of that capital for inorganic growth.

Allocators should watch for two follow-on events. First, Navitas is expected to host a transaction call within ten days where management will disclose the base purchase price, earnout triggers, and pro forma revenue impact. Second, the company's next quarterly report—due in early May—will show whether its organic GaN revenue is growing fast enough to justify absorbing Claros's operating expenses before the earnout milestones hit. If organic growth is decelerating, the acquisition looks more like a hedge than a springboard.

The $8 billion market figure is a 2028 projection from Navitas's internal TAM model, not third-party validated. But the direction is correct: AI power is a design bottleneck, and the companies that solve it will capture disproportionate economics as hyperscalers rush to add inference capacity ahead of o3 and Gemini 2.0 rollouts. Navitas just paid for a seat at that table.