Oppenheimer Flags Six Power-Infrastructure Stocks Poised to Benefit as AI Data Centers Stress Grids

Oppenheimer analysts Colin Rusch and Noah Kaye have laid out a shortlist of six public companies they believe stand to gain from the unique power challenges created by artificial intelligence data centers. The analysts frame the issue as a combination of two pressures: first, AI training and inference workloads create sustained, large-scale duty cycles unfamiliar to the power sector; second, operators must keep power continuously available to ever-denser compute clusters while also being able to ramp demand up and down from moment to moment.

The firm highlights that as data centers increasingly adopt 800VDC-based power architectures, commercially viable solutions will tend to be modular and built from components that can tolerate intense switching demands while offering improved efficiency and reliability over traditional approaches.

Oppenheimer's six picks

Below are the six stocks singled out by the analysts, their Oppenheimer ratings and price targets, and the rationale provided for each company.

1. AMSC - Outperform, $71 price target. Oppenheimer points to American Superconductor's expertise in managing voltage and harmonics as a competitive asset for edge-level power solutions in data centers. The analysts specifically call out the firm's Comtrafu product line, which is qualified for North American sales, as relevant to facility-edge deployments where incoming power quality can materially affect internal efficiencies. AMSC's track record in producing rugged, small form factor products for difficult power-quality environments, together with existing relationships across grid, utility, and EPC channels, are cited as reasons the company could be a partner for operators managing local grid dynamics.

   Separately, the company reported fourth-quarter fiscal 2026 results that exceeded analyst expectations for both revenue and earnings per share.

2. ELVA - Outperform, $14 price target. Oppenheimer views ELVA's ceramic separator technology as a potential fit for rack-level storage, where voltage volatility is a concern. The technology's ability to manage heat during charge and discharge cycles produces performance characteristics that the analysts say resemble a hybrid of capacitor and battery behavior. Oppenheimer expects ELVA to begin trials before the end of 2026, with the likelihood of follow-on orders in 2027 for its 30-minute storage product.

3. ENPH - Outperform, $57 price target. Among newer entrants into the data center power-infrastructure market, Enphase Energy is highlighted for its potential to distinguish itself via quicker time to market and resilient margins. The analysts emphasize Enphase's ASIC design expertise and modular architecture as advantages, particularly as they are applied to the company's monolithic bidirectional gallium nitride (GaN) switches designed for demanding duty cycles.

   In related developments, Barclays upgraded Enphase to Equalweight, citing the company's solid-state transformer opportunity for data centers. Enphase also began production shipments of its IQ9S-3P Commercial Microinverter in the United States.

4. ENS - Outperform, $250 price target. Oppenheimer notes that data center customers already account for more than 10% of sales at EnerSys, and the firm sees additional growth potential across the company's storage and DC power offerings. Near-term catalysts identified by the analysts include potential business wins for EnerSys's newly introduced data center lithium-ion UPS product and its warehouse battery energy storage system (BESS) solutions.

   EnerSys also announced fourth-quarter fiscal 2026 earnings and revenue that beat consensus estimates, and the company received price target increases from both BTIG and Oppenheimer. EnerSys introduced DataSafe Noir, a new lithium-based energy storage system specifically designed for data center applications.

5. FPS - Outperform, $60 price target. The analysts describe FPS's data center revenue as weighted toward medium-voltage products and note that the company has progressed on offerings that align with 800VDC architectures, including DC-native switchboards. Oppenheimer believes FPS has the organizational agility to bring new offerings to market quickly.

6. VRT - Outperform, $353 price target. Vertiv is portrayed as pursuing a multi-path, evolutionary approach to DC power architecture. The company plans a range of offerings coming to market, including centralized medium-voltage UPS-BESS systems, rack-level sidecars for conversion, and both medium-voltage DC UPS and solid-state transformer (SST) products.

   Separately, Vertiv completed its acquisition of ThermoKey, a heat-exchange technology provider, to broaden its thermal management portfolio in Europe, the Middle East, and Africa. Bernstein also initiated coverage on Vertiv with an Outperform rating.

Context from the analysts

Oppenheimer frames the list around a narrow technical thesis rather than a broad mandate. The central point is that the combination of sustained duty cycles and the need to modulate demand rapidly creates a power-delivery environment unlike what many parts of the power industry have previously had to manage. The migration toward 800VDC within data centers, the analysts say, elevates the importance of modular system architectures and components capable of fast, efficient switching while maintaining reliability under stress.

Key points

• AI training and inference workloads impose sustained, large-scale duty cycles on data center power systems, creating new engineering demands.
• The industry trend toward 800VDC architectures makes modular, high-frequency switching components especially valuable for rack and facility power designs.
• Vendors with experience in voltage and harmonics management, thermal control, and compact, rugged power modules may have strategic advantages in serving data center operators.

Risks and uncertainties

• Execution risk for new products and trials - several companies are in trial or early production phases (for example, ELVA's expected trials before the end of 2026 and follow-on orders in 2027), and successful commercialization is not guaranteed.
• Market adoption and timing - the analysts' case relies on migration to 800VDC architectures and modular approaches, the pace and extent of which could vary and influence vendor opportunities.
• Competitive and technological risk - multiple approaches to DC power architecture are emerging, and not all product road maps will translate into market wins.

The analysts' selection highlights companies across a range of product sets - from ceramic separators and rack-level storage to medium-voltage systems and solid-state transformers - reflecting the multifaceted nature of the power challenges driven by AI workloads. Investors and industry participants should note the technical focus of the thesis: products that can manage voltage quality, heat, switching intensity, and modular scaling are the capabilities Oppenheimer cites as differentiators for the list above.