Analog Devices buys Empower for $1.5B cash

Analog Devices is buying Empower Semiconductor for $1.5 billion in cash to improve AI power delivery.

Analog Devices said on May 19, 2026 that it will acquire Empower Semiconductor in an all-cash deal valued at $1.5 billion. The target close is the second half of 2026, and the move is aimed squarely at one of AI hardware’s messiest problems: getting power to chips fast enough, efficiently enough, and close enough to the processor.

The announcement matters because it is about more than adding a smaller chip company. ADI is buying technology that sits inside the power delivery chain for AI systems, where every millimeter and every watt can affect performance, heat, and cost. That is especially important as hyperscalers keep packing more compute into tighter racks.

Why ADI wants Empower now

AI chips are getting faster, but the power system around them is becoming harder to design. That is where Empower fits in. Its Integrated Voltage Regulator work and silicon capacitor technology help move power conversion closer to the processor, which shortens the path and reduces losses.

ADI framed the deal as a response to a very specific constraint: power density. In plain English, the issue is not only how much electricity a data center has, but how efficiently it reaches the chip that needs it. That is a big deal for AI training clusters, inference servers, and any system that runs hot under constant load.

Vincent Roche, ADI’s CEO and chair, put it bluntly in the company’s release: “AI infrastructure is fundamentally reshaping how power must be delivered, with energy now the most persistent constraint to scaling next-generation systems.”

That quote gets to the heart of the acquisition. ADI already has a strong power-management business, but Empower gives it more specialized silicon for the next layer down, where voltage regulation and capacitance become system-level bottlenecks.

• Empower’s silicon capacitors are already in production.
• IVR programs are advancing with hyperscalers and AI silicon providers.
• Tim Phillips will keep leading IVR technology efforts after the deal closes.
• ADI says the combined platform reaches from grid to core power delivery.

What changes for AI hardware buyers

The practical question is whether this deal changes what customers can build. For AI system designers, the answer is probably yes, but gradually. ADI is not buying a consumer brand or a software layer. It is buying deeper control over the electrical plumbing that feeds accelerators, memory, and surrounding logic.

That kind of acquisition can matter in three ways. First, it can reduce the number of vendors a customer has to manage. Second, it can speed up integration between regulator, capacitor, and package-level design. Third, it can help ADI sell a more complete power stack into large accounts that care about efficiency as much as raw performance.

Empower’s own framing is just as direct. Tim Phillips, Empower’s CEO, said: “Empower was founded to solve the hardest problem in AI power delivery – the power bottleneck that is limiting AI throughput.”

That line is useful because it explains why a power company is getting such a large price tag. In AI infrastructure, power delivery is no longer a back-office component. It is part of the performance story.

• ADI said the transaction was approved by both boards.
• The deal still needs customary closing conditions.
• It also needs the waiting period under the Hart-Scott-Rodino Antitrust Improvements Act of 1976 to expire.
• Closing is targeted for the second half of 2026.

How the numbers compare

The $1.5 billion price is large for a specialized power-tech acquisition, but it is small next to ADI’s own scale. That matters because it suggests the company can keep investing in AI infrastructure without betting the entire balance sheet on one move.

Here is the comparison that jumps out: ADI reported more than $11 billion in FY25 revenue, then $3.16 billion in Q1 FY26 revenue. Against that backdrop, the Empower deal is strategic rather than transformational. It is meant to strengthen a high-value part of the portfolio, not to redefine the whole company.

Investors should also keep the timing in mind. A second-half 2026 close means the market will have more than a year to watch whether Empower’s technology keeps gaining traction with hyperscalers. That is a long runway for a deal that is supposed to help with near-term AI demand.

• ADI generated more than $11 billion in FY25 revenue.
• Empower is being bought for $1.5 billion in cash.
• ADI’s recent Q1 FY26 revenue was $3.16 billion.
• The company said the close should happen in H2 2026, not immediately.

What to watch next

This acquisition tells you where the AI hardware race is heading: closer to the power rail, not just the chip core. If ADI can fold Empower’s IVR and capacitor work into its broader portfolio without slowing execution, it could become a stronger supplier to the companies building AI servers and accelerator platforms.

The main risk is timing. A deal announced in May 2026 but expected to close in the second half of the year leaves plenty of room for regulatory delays, integration questions, and customer churn. Still, the strategic logic is easy to see, and ADI has the balance sheet to wait.

For now, the key question is whether more semiconductor deals will follow the same pattern: not buying compute itself, but buying the power, packaging, and thermal pieces that make compute usable at scale.

If that happens, this deal will look less like a one-off and more like a preview of where AI infrastructure spending is heading next.

Related coverage: OraCore news