For a decade, “India will make its own chips” was a policy announcement rather than a factory. In 2026, that changed. There is now silicon being packaged on Indian soil, a fabrication plant approaching first production, and a supply chain forming around both.
Here is what is genuinely operational, what is still under construction, and — importantly — what this does and does not mean for the price of your next phone.
What is actually running right now
Micron’s Sanand facility, Gujarat
Micron Technology’s Semiconductor Assembly, Test and Packaging (ATMP) facility in Sanand was inaugurated on February 28, 2026 — making it the first operational semiconductor facility of the current mission cycle.
Micron expects the plant to assemble and test tens of millions of chips during 2026, scaling to hundreds of millions in 2027.
An important clarification that often gets lost in coverage: this is packaging and testing, not fabrication. The silicon wafers arrive from elsewhere. Sanand cuts them into individual dies, packages them into usable chips, and tests them. It is a genuinely valuable and technically demanding part of the supply chain — but it is not “making chips from sand.”
Tata’s Dholera fab, Gujarat
This is the one that matters most, because it is actual wafer fabrication.
- Special Economic Zone notified: April 9, 2026
- Site size: 66 hectares
- Expected employment: around 21,000 people
- Technology partner: PSMC (Powerchip Semiconductor Manufacturing Corporation), Taiwan
- Capacity target: approximately 50,000 wafers
- Starting node: 28nm, later advancing to 22nm
- First silicon expected: late 2026 into 2027
Why 28nm is the smart choice, not a disappointment
The most common criticism of Dholera is that 28nm is “old technology” when TSMC is producing at 3nm and below. That criticism misunderstands the market.
Leading-edge nodes — 3nm, 5nm — are used for smartphone flagship processors and AI accelerators. That is a narrow, brutally competitive segment requiring extreme ultraviolet lithography equipment costing hundreds of millions of dollars per machine, with a supply queue measured in years.
28nm and 22nm, by contrast, power the overwhelming majority of chips in the world. Automotive controllers, power management units, display drivers, industrial sensors, IoT modules, appliance microcontrollers. This is high-volume, long-lifecycle, steady-demand silicon — and during the 2021 shortage, it was precisely these mature-node chips that halted car production lines worldwide, not the cutting-edge ones.
Starting at 28nm is what a serious entrant does. Attempting to leapfrog to 3nm as a first project would be an expensive way to fail.
What this means for chip prices (be realistic)
Here is where expectations need managing carefully.
Your next phone will not get cheaper because of Dholera. Smartphone application processors are made on leading-edge nodes that India will not produce for many years. The Snapdragon or Dimensity chip in a 2027 phone will still come from Taiwan or South Korea.
What domestic production realistically delivers:
- Supply chain resilience — fewer production halts when global shortages hit mature-node components
- Reduced import dependence in automotive and industrial electronics, where India has substantial manufacturing
- Skills and ecosystem development — arguably the most valuable long-term outcome, as 21,000 trained semiconductor workers create capability that compounds
- Strategic autonomy in defence, telecom and infrastructure electronics
Those are real and significant benefits. “Cheaper smartphones” is not among them, and anyone promising that is overselling.
The honest risks
Semiconductor fabrication is among the least forgiving manufacturing disciplines that exists. Three genuine risks deserve acknowledgement:
Yield ramp is hard. Building a fab is the easy part. Getting acceptable yields — the percentage of working dies per wafer — takes time even for experienced operators. Early yields are typically poor, and improving them is a slow, expensive grind. Timelines slipping is normal, not scandalous.
Talent depth. India has excellent chip design talent; a large share of global semiconductor design work already happens in Bengaluru and Hyderabad. Manufacturing process engineering is a different discipline, and that expertise pool is thinner. The PSMC partnership exists substantially to bridge this.
Sustained capital commitment. Fabs require continuous reinvestment to stay viable. A single plant is a start, not a finished industry.
How to read the news over the next 18 months
When you see semiconductor headlines about India, three questions separate substance from announcement:
- Is this fabrication or packaging? Both matter, but they are very different levels of technical achievement.
- Is this a signed agreement or operational production? MoUs are common; running lines are rare.
- What node, and for what end market? This determines whether it affects cars, appliances, or phones.
Advertise with TechReview.Trade
Reach India’s technology decision-makers
Sponsored analysis, thought-leadership placements, banner inventory and newsletter features. Media kit, traffic data and rate card shared on request.
+91 8368966439 · Typical reply within a few hours, Mon–Sat
Frequently asked questions
Has India made its own semiconductor chip yet?
Partly. Micron’s Sanand facility has been assembling, testing and packaging chips since February 2026. Full wafer fabrication at Tata’s Dholera plant is expected to produce first silicon from late 2026 into 2027.
What is the difference between a fab and an ATMP facility?
A fab (fabrication plant) creates the actual circuits on silicon wafers. An ATMP facility takes finished wafers and cuts, packages and tests them into usable chips. Fabrication is the more capital-intensive and technically demanding stage.
What chips will Tata’s Dholera plant make?
It begins at the 28nm node, later advancing to 22nm — mature nodes used for automotive controllers, power management, display drivers and industrial electronics, rather than smartphone flagship processors.
Will Indian-made chips make phones cheaper?
No, not directly. Smartphone processors use leading-edge nodes that India will not manufacture for years. The real benefits are supply-chain resilience, reduced import dependence in automotive and industrial electronics, and domestic skills development.
Who is Tata’s technology partner for the Dholera fab?
PSMC (Powerchip Semiconductor Manufacturing Corporation) of Taiwan, which brings the process engineering expertise required to operate a fab at commercially viable yields.
The bottom line
2026 is the year India’s semiconductor programme moved from announcements to actual output. That is a real milestone, and it deserves to be recognised as one.
It is also the beginning of a long, capital-hungry project whose most meaningful returns — a trained workforce, a supplier ecosystem, and genuine supply-chain independence in mature-node silicon — will take a decade to fully arrive. Both things are true at once, and the coverage that tells you only one of them is selling something.
Related reading: Google Pixel 11 India: Price and Specs · The AI Infrastructure War of 2026
