Industrial power buyers stop waiting for the grid

Tata Group switched on a 100-megawatt captive solar plant in India while a hybrid generator in the Democratic Republic of Congo began supplying baseload power to a copper mine, and neither event looks like a grid story until you notice what is missing: no interconnection agreement, no utility offtake, no queue position. The buyer is the generator.

Large industrial energy buyers are no longer waiting for grid operators to connect them; they are building, buying, or contracting their own generation, and that shift is re-pricing what a queue position is worth. The scarcity premium that once attached to a spot in the interconnection line is migrating to something less visible: the capability to supply oneself.

Tata's 100 MW is a self-supply asset for a conglomerate rather than a merchant purchase of power, insulating industrial loads from congestion pricing, curtailment, and the allocation politics of a strained grid. For a company of Tata's scale, it is a planning tool first and an investment second; every captive plant of this kind removes a potential anchor customer from the merchant queue.

The DRC hybrid plant makes the same point from a harder geography, where baseload supply to a copper mine—key terms undisclosed—shows captive power in an environment where grid reliability cannot be priced. Miners whose ore bodies are scheduled in years, not decades, cannot afford a connection timeline set by a utility; owning the generation lets a mine own its schedule.

Cape Town's 70 MW municipal solar PPA sidesteps Eskom by contracting directly for generation instead of waiting for a national utility to deliver it, using procurement as a route around a failing grid. It is self-supply by signed contract rather than construction, and it expands the pool of buyers taking matters into their own hands.

Brookfield's Solarity added 5.2 MWp of Chilean commercial and industrial solar, a transaction small enough to be invisible next to gigawatt headlines but the same behavior in smaller denominations: businesses attaching generation to their own operations rather than queuing for a wire. Distributed solar transacts in human-scale increments because the buyer's problem is next month's power bill, not the grid's ten-year plan.

None of these buyers is joining a queue; they are contracting around it. The old model sits on the other side of the week.

Genesis made a US$74m solar FID in New Zealand, a capital commitment that buys a place in the connection queue. The offtake question decides what that place is worth: capital committed before revenue, time-in-line as the asset, the anchor tenant assumed to arrive later.

The FID may be the rational price of a queue position where new connections are years away, but it is a different asset class from the Tata plant, which never needed a third party to make the project bankable. One is a bet that the customer will show up; the other has already become the customer.

Natural Power won a ground survey contract for a 2 GW British HVDC link, a small award in pounds that nonetheless commits real money to surveys, consents, and planning years before a cable earns anything. It is a queue position carried as a cost, pricing the permission to connect rather than the power that will eventually flow.

Project Cayman shows what the old model looks like when the anchor tenant does not arrive: a 2.88 GW gas-and-battery campus in southern Louisiana with a site and a public-meetings calendar but no signed data center customer to pay for it. The project is deep in scale and shallow in contracted revenue, exactly the profile that looks riskier as captive power removes the deepest-pocketed buyers from the market.

ERG's 22.5 MW French repowering is the missing middle: the plant starts up without joining the interconnection queue because it trades existing grid rights for larger capacity at the same point. The asset being upgraded is the queue position itself, not a new permission—a template for adding generation where new connections are the binding constraint.

Elawan's 20-MW Spain filing puts permits first: the hybrid storage permit application is a claim on a grid slot, not a power plant yet. In a market where permission to connect is the binding constraint, the filing is the deal, and the turbines, batteries, and financing follow.

Catalonia's 300 MW solar announcement clears a political hurdle; the offtake and grid slot will decide whether it clears a financial one. Across the sector, consent is the asset, and cash flows are the consequence of having consent in the right place.

Sunotec broke ground on a 100-MW/441-MWh German battery with no offtake terms alongside it. Concrete can be poured without a customer, and storage carries the merchant risk most exposed to the grid's own delays: a battery that cannot connect earns nothing.

Cloudberry closed its purchase of Orron's Nordic platform, a transaction in which a platform changes hands as the transition premium moves downstream to grid rights and contracted revenue. Buying an existing portfolio with connections and PPAs already in place is buying queue positions that have matured, which is why the deal matters more than its disclosed terms suggest.

Value is migrating, and developers who still build merchant capacity without an anchor tenant are not solving the grid problem; they are betting that someone else will pay to connect. Every Tata or DRC miner that self-supplies removes the deepest-pocketed buyer from that bet, and Cape Town's PPA shows the public sector learning the same lesson.

The merchant queue is left with fewer anchor tenants and more project risk at precisely the moment the queue itself has become the scarce input. Projects still paying to wait are buying last decade's scarcity at this decade's prices.

The next test is whether the New Zealand FID finds an offtaker before the Louisiana campus fills a data center lease. Neither project needs more capital; both need the same signature.