For a decade sodium-ion was the battery that would matter once lithium got expensive. In 2026 it picked up something it never had before: order books measured in gigawatt-hours. CATL signed a three-year deal in April to supply Beijing HyperStrong, one of the world’s five biggest storage integrators, with 60 GWh of sodium-ion cells. In the US, Peak Energy holds a phased agreement with Jupiter Power for up to 4.75 GWh between 2027 and 2030, and Alsym Energy has a 500 MWh deployment lined up with developer Juniper Energy, mostly in California.

A data centre can secure land, announce a supply agreement and order computing equipment while still lacking the grid connection needed to operate at the intended scale. That gap makes local electricity infrastructure central to the AI power debate. National generation totals matter, but a project needs deliverable power at a particular place, with an acceptable connection date and a workable operating arrangement. The International Energy Agency’s Electricity 2026 report places grid constraints and connection queues among the major issues facing expanding electricity systems.

An offshore wind project starts spending money long before its turbines produce electricity. Developers need surveys, engineering, equipment orders, port access and vessels capable of transporting and installing enormous components at sea. Suppliers make commitments of their own, often years ahead of construction. When federal policy leaves the future of a project uncertain, the consequences travel through that entire chain. A delayed permit can unsettle a vessel reservation. A cancelled lease can weaken the case for a port expansion.

For a year the story of US clean energy policy looked like a straight line downhill. Credits cut, grants cancelled, offshore wind ordered to stop. Look closer and the line has kinks in it, and most of them were put there by federal judges. CrossGov.com, a site that covers US government with a federal focus, has just published four posts that trace what’s actually left of federal support for renewables and electric vehicles.

Northvolt raised more than $13 billion on the promise of a European battery industry. It filed for Chapter 11 in November 2024 with $5.8 billion of debt and about $30 million in cash, and went into Swedish bankruptcy in March 2025, the largest in modern Swedish industrial history. Eighteen months on, the pieces have nearly all been sold. On September 25 the last valuable one found a buyer. Refinyx, a Stockholm startup, came out of stealth by buying Northvolt’s recycling technology: a portfolio of 134 patents and the 13,000-square-foot pilot and lab facility in Västerås.

Energy Vault is buying the project portfolio of Goshe Energy Storage, a Boulder, Colorado developer of utility-scale batteries. The deal adds 15 US projects totaling more than 2.3 GW to Asset Vault, the unit through which Energy Vault owns and operates storage instead of just building it for others. No price was disclosed. The shares rose on the news. Most of that 2.3 GW is pipeline, and pipeline is cheap to announce.

H2SITE has won €3.4 million (NOK 39.1 million) from Enova, Norway’s state climate fund, to build a commercial-scale ammonia-to-power system for ships. The project, called HydraNord Power, will produce a 700 kW net unit that cracks ammonia into hydrogen on board and feeds it to fuel cells. The system can make up to one tonne of hydrogen a day. It’s aimed at auxiliary and hybrid power on deep-sea vessels and offshore operations: the electricity a ship needs for everything except main propulsion.

Two next-generation geothermal companies moved closer to the grid on the same day. Quaise Energy was picked for up to $25 million in Department of Energy support for its superhot plant in Oregon. Sage Geosystems chose Ormat’s Blue Mountain power plant in Nevada as the site for its next commercial-scale enhanced geothermal system (EGS). Different technologies, different business models. Same destination: firm power, around the clock, from rock that conventional geothermal can’t use.