Introduction
On 5 November 2025 SolarEdge Technologies (a smart‑energy leader) and Infineon Technologies (a power‑semiconductor giant) announced that they would jointly develop a 2–5 MW solid‑state transformer (SST) platform for AI and hyperscale data centres. The product combines Infineon’s high‑voltage silicon‑carbide (SiC) switches with SolarEdge’s DC‑conversion and control technology, enabling direct conversion of medium‑voltage grids (13.8–34.5 kV) to 800–1,500 V DC with >99 % efficiencybusinesswire.com. This section analyses how the collaboration could reshape the SST market, the broader economic implications of surging data‑centre power demand and what it means for customers.
Market Impact
Growth of the solid‑state transformer market
- Accelerating SST market: Mordor Intelligence estimates the global solid‑state transformer market will grow from USD 181.35 million in 2025 to USD 339.03 million by 2030, a compound annual growth rate (CAGR) of 13.33 %mordorintelligence.com. Momentum comes from utilities, rail operators, data‑centre owners and electric‑vehicle (EV) chargers, which value bidirectional power flow, real‑time voltage regulation and compact footprintsmordorintelligence.com. Medium‑voltage equipment (2–36 kV) accounted for 56 % of market revenue in 2024 and remains the dominant segmentmordorintelligence.com. The Asia‑Pacific region commands 40 % of revenues and is the fastest‑growing market due to government stimulus toward grid resilience and rail electrificationmordorintelligence.com.
- Data‑centre segment: Data centres represent one of the fastest‑growing SST applications. DataIntelo’s Solid‑State Transformer for Data Centres report estimates this segment reached USD 1.14 billion in 2024 and will expand to USD 8.88 billion by 2033 (CAGR 22.7 %)dataintelo.com. Growth is driven by digital‑infrastructure expansion, AI workloads and edge computingdataintelo.com. SSTs appeal to operators because they provide enhanced power quality, bidirectional power flow, reduced footprint and seamless integration of renewablesdataintelo.com.
- Drivers and restraints: Mordor Intelligence identifies several drivers that will affect SST growth. A rapid smart‑grid rollout contributes 2.8 percentage points to the market’s CAGR, renewable‑integration requirements add 2.5 points and EV fast‑charging hubs add 2.0 pointsmordorintelligence.com. The AI data‑centre power‑density push contributes 1.8 points, reflecting rising demand for high‑power DC architecturesmordorintelligence.com. Key headwinds include high capital cost, with SST units costing 2–3 times more than oil‑filled transformers, lack of standards/interoperability and SiC/GaN wafer supply bottlenecksmordorintelligence.com.
Impact of the SolarEdge–Infineon collaboration
SolarEdge and Infineon are positioning themselves at the intersection of two booming markets: power semiconductors and data‑centre power infrastructure.
- Technology leadership: The jointly developed 2–5 MW SST building block will convert grid‑level voltages directly to 800–1,500 V DC, avoiding multiple AC‑DC conversion stages. This enables >99 % efficiency and significant reductions in size, weight and CO₂ footprintbusinesswire.com. Compared with legacy AC architectures, 800 V DC distribution can improve end‑to‑end efficiency by ~5 %, reduce copper requirements by 45 % and lower maintenance costs by 70 %, lowering total cost of ownership by up to 30 %developer.nvidia.com.
- Market penetration: SolarEdge brings 15 years of expertise in DC‑coupled solar‑inverter technology, while Infineon contributes advanced SiC switches and GaN devices. The partnership allows SolarEdge to expand beyond residential solar into hyperscale data‑centre marketsbusinesswire.com. If successful, the collaboration could accelerate SST adoption, set new efficiency benchmarks and pressure competitors (e.g., ABB, Hitachi Energy, Mitsubishi) to follow with similar solutions.
- Potential revenue: For SolarEdge, entry into the data‑centre market provides a new revenue stream at a time when residential‑solar demand is volatile. The company has signalled that data‑centre products could ship in 2027–2028, with revenue potential in the hundreds of millions of dollars by the end of the decade (per analyst notes from earnings calls). Infineon’s SiC division, already growing thanks to EV inverters, would benefit from additional demand for high‑voltage power semiconductors.
Economic Implications
Surging energy demand from AI data centres
The economic context for SST deployment is shaped by the explosive growth of AI workloads and the resulting strain on electricity infrastructure:
- Global consumption: The International Energy Agency (IEA) estimated that data centres consumed 415 TWh of electricity in 2024, around 1.5 % of global consumption, and projected usage to 945 TWh by 2030. OPEC’s estimate in the IMF’s Chart of the Week is higher: global data centres used 500 TWh in 2023 and could triple to 1,500 TWh by 2030—comparable to India’s entire consumptionimf.org.
- US perspective: Pew Research reports that U.S. data centres consumed 183 TWh in 2024, more than 4 % of national electricity, and are expected to reach 426 TWh by 2030, a 133 % increasepewresearch.org. AI‑centric hyperscale centres can consume as much electricity as 100,000 households, and forthcoming facilities could use 20 times that amountpewresearch.org. In states like Virginia, data centres already consumed 26 % of the state’s electricity supply in 2023pewresearch.org.
- European outlook: A Finnish government‑commissioned study suggests that if data‑centre capacity in Finland rises from 300 MW today to 2,500 MW by 2030, average electricity prices could increase 10 %, with hourly prices spiking above €0.50 /kWh for about 100 hours annuallyyle.fi. The European Central Bank notes that AI‑related energy consumption in data centres is currently ~20 TWh (0.02 % of global energy) but could rise by 90 TWh by 2026, equivalent to ~4 % of the EU’s current electricity useecb.europa.eu. Under a scenario where all additional demand is met with natural gas, gas prices in Asia/Europe could rise by about 9 %, with AI‑driven centres contributing about 2 percentage points of that increaseecb.europa.eu.
- Environmental and macroeconomic impacts: The IMF warns that, under current policies, AI‑driven electricity demand could add 1.7 gigatons of greenhouse‑gas emissions between 2025 and 2030, roughly Italy’s energy‑related emissions over five yearsimf.org. Supply‑side responses matter: abundant investment in generation and transmission can moderate price increases, whereas sluggish supply leads to steeper cost hikes that could dampen economic growthimf.org.
Price effects on electricity customers
- Rising bills and grid upgrades: As data‑centre loads grow, utilities must expand transmission lines and capacity, costs often passed on to households. Pew Research notes that data‑centre demand in the PJM electricity market boosted the 2025–26 capacity‑market costs by US$9.3 billion, translating into residential bill increases of US$16–18 per month in Maryland and Ohiopewresearch.org. A Carnegie Mellon study cited by Pew Research projects that data centres and cryptocurrency mining could raise the average U.S. electricity bill by 8 % by 2030 and 25 % in the highest‑demand marketspewresearch.org.
- European price pressure: The Finnish study projects a 10 % increase in average electricity prices by 2030 if data‑centre construction accelerates, with prices spiking to €0.90 /kWh during some hoursyle.fi. Such price spikes could discourage industrial competitiveness and trigger political pushback.
- Potential downward pressure: Not all analyses conclude that data‑centre demand raises rates. A Brattle Group study (via PBS NewsHour) argues that if utilities already have spare capacity, large new customers such as data centres can spread fixed grid costs over more kilowatt‑hours, potentially reducing per‑unit rates. However, this benefit exists only when demand growth does not trigger expensive upgradespbs.org.
Implications for semiconductors and supply chains
The SolarEdge–Infineon partnership highlights another economic dimension: demand for SiC and GaN power devices. These wide‑band‑gap semiconductors offer higher efficiency but are currently in short supply. Mordor Intelligence cites SiC/GaN wafer supply bottlenecks as a restraint on SST market growthmordorintelligence.com. Demand from EV inverters, renewable‑energy inverters and now data‑centre SSTs will put upward pressure on wafer prices. If supply expands (e.g., through Infineon’s planned fab in Kulim, Malaysia, and new foundries by Wolfspeed and Onsemi), costs may fall over time, enabling broader adoption.
Customer Impact
For data‑centre operators and hyperscale cloud providers
- Efficiency and cost savings: By converting medium‑voltage AC directly to 800–1,500 V DC with >99 % efficiency, the SolarEdge–Infineon SST eliminates multiple intermediate transformers and AC/DC conversion stages. According to NVIDIA’s analysis of 800 V DC architectures, such configurations can reduce copper usage by 45 %, maintenance costs by 70 % and total cost of ownership by up to 30 %developer.nvidia.com. Reduced conversion losses also mean less heat, lowering cooling requirements.
- Higher power density and modularity: The 2–5 MW modules enable racks or clusters of GPUs/TPUs to draw hundreds of kilowatts per rack. This supports next‑generation AI workloads that require >100 kW per rack and allows operators to use smaller footprints while delivering more compute per square metre.
- Reliability and sustainability: SSTs provide real‑time voltage regulation, harmonic mitigation and fault isolationdataintelo.com, improving uptime in mission‑critical environments. Their bidirectional power flow facilitates integration of battery storage and onsite renewables, enabling micro‑grids and reducing dependence on fossil fuels. Because the modules are lighter and smaller than oil‑filled transformers, they can be installed closer to server halls, reducing installation complexity.
- Capital costs: SSTs are more expensive upfront (2–3 × the price of conventional transformers), but lower operating costs and energy savings can deliver a 20‑year lifecycle paybackmordorintelligence.com. Early adopters will need to justify capital expenditures based on long‑term energy and cooling savings.
For consumers and businesses buying AI services
- Potential price moderation: More efficient power conversion reduces the cost per AI operation. As data‑centre operators adopt high‑efficiency DC architectures, the cost of AI inference and training may decline, making AI services more affordable. This can stimulate new products and productivity gains across the economy.
- Environmental benefits: Higher efficiency and the ability to integrate renewables lower the carbon footprint of AI computations. Given that AI‑driven electricity demand could add 1.7 Gt of CO₂ emissions by 2030imf.org, efficiency gains from SSTs are essential for meeting corporate climate commitments and regulatory requirements.
- Risk of higher electricity tariffs: If grid expansion and generation do not keep pace, rising data‑centre loads could raise electricity rates, indirectly affecting all electricity users (households and businesses). The balance between positive and negative effects will depend on national energy policies, grid investment and local market conditions.
Conclusion
The SolarEdge–Infineon collaboration arrives as AI’s hunger for power reshapes the electricity landscape. By offering high‑efficiency, medium‑voltage‑to‑DC conversion modules, the partnership positions itself to capture a sizeable share of a rapidly growing SST market – a market expected to multiply both overall and within the data‑centre segment. For data‑centre operators, the technology promises lower energy losses, higher rack power densities and improved sustainability, potentially reducing the cost per AI computation and enabling greener AI services. For the broader economy, however, the story is more nuanced: data‑centre power demand is set to double or triple this decade, raising concerns about electricity prices, grid stability and emissions. Efficient SSTs and DC architectures are necessary but not sufficient; complementary investments in generation, transmission, storage and regulatory reforms will determine whether AI‑driven energy demand becomes an economic boon or a drag. Ultimately, the collaboration underscores that semiconductor innovation and power‑system design are now integral to AI economics and to ensuring that the digital revolution remains sustainable and affordable.
