If you are building a wind or solar project in India and selling across a state border, your business rests on one permission: connectivity to the inter-state transmission system. Until recently that permission had a name everyone knew: Long-Term Open Access (LTA), granted case by case, tied to specific transmission corridors. Since the GNA framework replaced it, the permission is called General Network Access, and almost nothing about how people talk about it has caught up. This note explains what actually changed, and more usefully what the filed numbers say about how the system behaves now.
The short version
Under LTA, you asked for a path: a corridor, a set of lines, a destination. Under GNA, you ask for access at a point: a named substation, a defined MW, and (for generators) a declared injection profile. You no longer need to identify the specific transmission the system will build to absorb you; that planning burden moved to the CTU. In exchange, you get access that is not tied to particular circuits, and a process that is supposed to be quicker and more standardised.
That is the theory. In practice, the part every developer cares about (how long the process takes, whether the dates hold) is only visible by reading the registers over time, which is what this platform does. The measured reality:
Filed to granted: 176 days median. Granted to effective: 1,064 days. End to end, a typical application lives 1,285 days in the pipeline — and the stated dates move, with a median slip of 123 days between register editions.
Nobody's project finance model has those numbers in it. They should, because they are the difference between "connectivity in year two" and "connectivity in year four."
What GNA actually gives you
A GNA grant is defined by four things: the substation, the MW, the direction (generation injecting, or consumption drawing), and the effective period. Note what is absent: a specific line, or a specific bay commitment in the grant itself. The physical slot, the bay, is coordinated separately, which is why a substation can show MW headroom on paper while having zero free bays in practice. When that happens, the headroom is theoretical for anyone wanting an independent connection. Both numbers are on this platform's substation pages, side by side, because either one alone will mislead you.
Generation and consumption GNA are the same instrument pointing in opposite directions, and the registers that list them are separate. That separation hides something the filings never join: at several nodes, consumption-side applications (green hydrogen electrolysers, bulk industrial loads, ports) are drawing down the same margins that generation is queueing to inject into. We call those two-sided collisions, and there were roughly 6 GW of claims sitting on them at the last edition. No public document states that overlap. It falls out of the join.
The lifecycle, measured
Every application moves through three observable states. It appears in the 4.x register as applied. A subset moves to the to-be-effective list as granted: approved, waiting for its effectivity date. And eventually it leaves that list and shows up in the effective book: connected. Each transition has a clock, and the clocks are not what the process documents imply:
The connectivity lifecycle, drawn to scale from the platform's archive. Medians; individual applications vary, and dated grants routinely move between editions.
- Filed → granted: 176 days median. This is the "fast" step, and it is genuinely fast when the node has margin and the application is clean.
- Granted → effective: 1,064 days median. The long pole. Bays, transmission schemes, construction, commissioning, all of it hides in this gap.
- Dates move. Comparing consecutive editions of the to-be-effective list, grant by grant, the median scheduled-date revision is 123 days later. Roughly a quarter of a year, silently, between one edition and the next.
One more measured fact worth carrying: applications backed by auction awards (a signed PPA behind them) close out at 2%, against 14% for everything else. Procurement discipline survives the queue. Paper does not.
What LTA still has to do with it
Legacy LTAs didn't vanish: existing access continues under grandfathering, and some conversion and augmentation processes still reference the old framework. If you hear "LTA" in a meeting in 2026, it is either a legacy right, a habit, or a consumer-side open-access discussion where state-level rules still use the older vocabulary. The operative question is always the same regardless of acronym: at which substation, how many MW, effective when, and what else is claiming that margin?
What this means for a new project
Work the clock backwards. If your target node's dated inflow crosses its remaining margin in 2027, and the filed-to-grant median is 176 days, the last comfortable filing month is roughly mid-2026, not 2027. That arithmetic (crossing year minus measured latency) is what this platform calls the file-by clock, and it is why we draw closure as a bracket rather than a date: the floor assumes nothing, the ceiling assumes stated dates hold, and the measured slippage says the truth usually lands between.
The regime changed so that access is easier to ask for. The physics of substations and the arithmetic of queues did not change at all. Both are still published, monthly, in PDFs, and both are now a search away: indiagridmonitor.com.