Counts tell a strategy, and the 2020 grid-control counts tell one of fragmentation. Looking at grants and publications in the H02J 3 family — circuit arrangements for AC mains generation and distribution — the named-assignee list reads like an industrial roll call rather than a platform war.
On the incumbent side: Rolls-Royce's US10530163B2 (microgrid control), Jiangsu Goldwind's US10637242B2 (microgrid dynamic stability), and GE's US10811882B2 (solar inverter grid emulation) anchor the named grants. These are exactly the companies you would expect — turbine makers, power-conversion houses, and grid-equipment firms — each fencing off a control method adjacent to its existing hardware business.
The more telling datum is the tail. A large fraction of 2020 grid-control publications carried no assignee at all — individual inventors and small entities filing method claims. US20200273120A1 (grid decision-support, machine learning) and US20200036748A1 (decentralized autonomous grid) are unassigned filings that read like academic or startup work.
What the fragmentation signals: in 2020, grid control was not yet a concentrated IP battleground. Contrast the battery landscape, where a handful of cell makers were already stacking H01M filings into defensible portfolios. Grid control was still method-rich and assignee-thin — lots of distinct ideas, little consolidation.
The caveat this desk always attaches: a snapshot of one year and one CPC family is a direction, not a verdict. Publications and grants are different populations, and an unassigned publication may later assign on grant. But the 2020 picture is consistent — grid-control IP was diffuse, which is itself a strategic fact about where the smart money was (and wasn't) building patent walls.