Record open Company files. Capital. Supply chain. Verified 3 Aug 2026
DholeraDigital
India's semiconductor build-out, tracked from the ground
ConfirmedRs 91,000 cr
Next windowQ4 2026
Pages on record75

Front PageIntelligenceThe node question: what Dholera will actuall...

Intelligence

The node question: what Dholera will actually make when it opens

A fab is defined by the process nodes it can run. On this specific point the public record now contains a direct contradiction, and almost nobody covering Dholera has noticed it.

What this page establishes

  1. What was promised in 2024
  2. What Bloomberg reported in July 2026
  3. Tata's response, quoted precisely
  4. The contradiction nobody has printed
  5. Why the distinction actually matters
90nm at open

Bloomberg reported on 17 July 2026 that Tata Electronics will open the Dholera fab mostly on the 90nm process node rather than the 28nm publicly promised in 2024, with commercial production in mid-2028. TrendForce independently carried the report on 20 July 2026.

Source: Bloomberg via Outlook Business, 17 Jul 2026; TrendForce, 20 Jul 2026. As of 27 July 2026.

What was promised in 2024

When the Union Cabinet approved the plant on 29 February 2024, the public framing was a 28nm-capable fab: Rs 91,000 crore, up to 50,000 wafer starts per month, and a node range spanning 28nm to 110nm DURABLE. Tata Electronics' own foundry page still lists that range, naming 28, 40, 55, 90 and 110nm and describing target products including power management integrated circuits and microcontrollers. The 28nm figure did most of the rhetorical work, because it is the number that signals a modern, credible foundry rather than a legacy line.

What Bloomberg reported in July 2026

On 17 July 2026 Bloomberg reported that the fab will open mostly on 90nm, not 28nm, and that commercial production is expected in mid-2028. TrendForce carried the same finding on 20 July 2026, attributing it to the Bloomberg report and naming ASML as the equipment supplier for the 300mmThe industry-standard large wafer size used by modern commercial fabs, including Dholera's. fab. Union Electronics and IT Minister Ashwini Vaishnaw stated the plant is expected to begin commercial production by mid-2028 REPORTED on the node detail, which rests on reporting rather than a company filing.

Tata's response, quoted precisely

A Tata Electronics spokesperson responded that the plan "has always been to start with 55nm and 90nm before introducing 28nm", and that 28nm "will be a key part of our offering". PSMC's spokesperson Eric Tang said technology transfers are "typically introduced gradually, starting with more mature nodes". Both statements are entirely consistent with normal industry practice, and the partnership file explains why a gradual transfer is the sane engineering choice for a first fab.

The contradiction nobody has printed

Here is the part that makes this more than a clarification. In the Tata Sons annual report for the year ended March 2025, chairman N Chandrasekaran wrote that "We have chosen to start our chip journey at the 28nm node", describing it as a deliberate starting point REPORTED on the quotation as carried in coverage of the Bloomberg report. That is not easily reconciled with a spokesperson saying the plan had always been to start at 55nm and 90nm.

Both statements cannot be describing the same plan. Either the starting node changed between the annual report and July 2026, which would be an ordinary and defensible engineering decision, or the public description of the plan has been inconsistent. This record does not assert which, because the available sources do not settle it. It states the contradiction plainly and lets it stand, which is more than the coverage has managed so far.

Why the distinction actually matters

Not for prestige. For customers. A 90nm line and a 28nm line serve overlapping but different markets: 90nm is deep legacy territory suited to power management, analog and certain microcontrollers, while 28nm reaches into more demanding embedded and connectivity applications and is the workhorse node much of the automotive and industrial world is standardising on. Opening at 90nm is a perfectly rational way to build yield discipline before attempting a harder process, and our mature-node economics page defends that logic. But it does change which customers the fab can serve in its first years, and therefore how quickly it can fill capacity.

What would settle it

  • A dated company statement specifying the node mix at production start, rather than a spokesperson's characterisation of intent.
  • Equipment orders, which reveal capability: no valued semiconductor equipment purchase orders for Dholera are public as of August 2026, and the ASML relationship remains an MoUMemorandum of understanding: a stated intent to engage, not a contract or delivered capital..
  • Customer qualification announcements, which reveal which nodes actually have demand behind them.

Until one of those arrives, the honest position is the one at the top of this page: the official specification says 28 to 110nm, the reporting says the line opens mostly at 90nm, and the company's own chairman wrote something different again. Our company file carries the running record, and the capital ledger keeps the confirmed money separate from the contested detail.

Cite this: "Bloomberg reported in July 2026 that Tata's Dholera fab will open mostly on 90nm, not the 28nm promised in 2024." Dholera Digital, 2026-08-03. https://dholera.digital
Sources and verification trail
  1. Primary-source research pass, 4 August 2026: company newsrooms, ASML, PIB, NICDC/DPIIT Delivery Monitoring Unit report of 31 May 2026, GSEM and CMO Gujarat.
  2. Tier-1 press named inline with dates: Bloomberg via Outlook Business, TrendForce, Economic Times, Business Standard.
  3. Negative findings are published as findings, not omitted.
  4. Method and correction policy: dholera.digital/editorial-standards/