From converter to controller
In 2015 the inverter converted DC to AC and stepped out of the way. A decade later it is the operating system of a distributed energy asset. Part of the history record behind the 2026 roadmap.
The record, sourced
Figures below are drawn from Wood Mackenzie's global inverter shipment rankings, IHS Technology, S&P Global, NREL, semiconductor-industry references and policy reporting: each figure names its source. This page is historical only: the projections live on the roadmap.
Milestones of the evolution
| Year | Milestone | What it tells us |
|---|---|---|
| 2015 | Huawei tops global shipments with a 13.3% share in a fragmented field: SMA, ABB, TMEIC and Sungrow contest the top five (IHS Technology) | The starting grid: a passive, commodity component market with no dominant platform |
| 2015–19 | Silicon IGBT defines the era: peak efficiency tops out around 96–97%; the string inverter has no battery awareness and no grid-signal response beyond anti-islanding (semiconductor-industry references) | The 2015 device was a power adaptor with a kilowatt-hour counter |
| 2016–20 | Microinverters and optimisers mature in parallel: module-level electronics capture 5–25% more energy on shaded or complex roofs (NREL), while peak-efficiency parity with strings narrows to about one point | Architecture niches are won on real-world yield and monitoring granularity, not on peak conversion efficiency |
| 2020–23 | The battery integration moment: DC-coupled hybrids (95–97% storage round-trip path) displace AC-coupled retrofits (90–94%) as the default new-build architecture (industry datasheet comparisons) | Removing a conversion stage beat improving one: system architecture outweighed component efficiency |
| 2023–25 | Wide-bandgap adoption accelerates: SiC MOSFETs go from 18% (2023) to 35% (2025) of new utility-scale string inverter shipments, reaching ~99% peak efficiency; GaN stays under 2%, concentrated in microinverters (industry shipment data; Nexperia) | A two-to-three-point efficiency ceiling lift in two years: and the thermal headroom that made the compute payload possible |
| 2024 | Global shipments reach 589 GWac; Huawei (176 GWac) and Sungrow (148 GWac) hold 55% combined: the first time in Wood Mackenzie's decade of rankings that no other vendor exceeds 5%; nine of the top ten are China-headquartered (Wood Mackenzie, July 2025) | A fragmented 2015 field consolidated into a two-firm market as the engineering bar rose |
| 2024 | VPP policy matures: US utility and regulatory frameworks begin treating aggregated behind-the-meter fleets as dispatchable resources rather than pilots (Utility Dive, February 2025) | The inverter's communications and control payload became a regulatory expectation, not a feature |
| 2025 | SiC reaches system-level cost parity with silicon IGBT at 250 kW and above; IGBT retains 65% of utility string and 95% of central inverter shipments below that threshold (industry shipment data) | Coexistence, not replacement: the old technology gets pushed down-market, exactly as p-type silicon did in modules |
The decade in numbers
| Metric | Then | Latest reported | Source |
|---|---|---|---|
| Peak conversion efficiency ceiling | 96–97% (silicon IGBT era) | ~99% (SiC designs) | Semiconductor-industry references |
| Top-vendor concentration | Leader at 13.3% share (2015) | Top two at 55% of 589 GWac (2024) | IHS Technology; Wood Mackenzie, July 2025 |
| SiC share of new utility-scale string shipments | 18% (2023) | 35% (2025) | Industry shipment data |
| Storage conversion path efficiency | AC-coupled: 90–94% | DC-coupled hybrid: 95–97% | Industry datasheet comparisons |
| Switching frequency | IGBT: 15–30 kHz | GaN HEMT: 100–300 kHz; power density 0.5 → 4 W/in³ | Nexperia, November 2024 |
| Device job description | DC–AC conversion + kWh count (2015) | Conversion + battery management + backup switching + VPP dispatch + tariff optimisation (2026) | Product documentation across the benchmark set |
Where that leaves 2026
The inverter now sits at the intersection of five roles: power converter, battery manager, grid-service provider, VPP participant, and an autonomous decision layer weighing tariffs, forecasts and dispatch signals. That is why the benchmark scores semiconductor technology, backup and islanding capability, VPP readiness, open-protocol maturity, and firmware and security disclosure: the criteria follow the device's accumulated responsibilities. The projections live on the roadmap's conversion-technology section.
Why the record earns the projections
Two lessons. First, semiconductor transitions in this category run on cost-parity thresholds, not lab records: SiC went from premium option to 35% of utility string shipments within three years of approaching system-level parity: the yardstick for reading the roadmap's SiC-to-residential horizon. Second, function accumulates toward the device that already owns the connection: every new responsibility of the decade (battery coordination, backup, VPP dispatch, cybersecurity) landed on the inverter rather than a new box: which is why the roadmap treats grid-forming capability and platform lock-in as the decisive frontiers, not raw efficiency.
← All history segments · The 2026 roadmap →
Condensed from the Solar Analytica decade audit "The inverter decade"; sources named inline. Last review: July 2026.