Industrial Ecosystems 3: How Penang and Singapore Built a Chip Economy
How a bankrupt port and a new city-state each built a chip economy over fifty years, told through the people who did it, and what Jagiroad, Assam can copy.
0 of 11 sections read
A little after nine in the morning on 1 May 1975, a fire started in the mould room of Intel’s assembly plant in Penang, Malaysia. A faulty light was the likely culprit. By the time the fire was out, the factory that had been Intel’s first production site anywhere outside the United States was gone.
Ken Thompson, the plant manager, described what he found in Intel’s own company history. “The only thing standing was the cafeteria,” he said. Everything else had burned to the ground. The damage came to about 2.5 million US dollars, more than Intel had first put into the site.
What happened over the following months explains why Intel stayed. The roughly thousand Malaysian workers were kept on full pay. Production carried on from five borrowed locations, including the factories of other manufacturers during their off-shifts. The plant was rebuilt in under a year, and was running again in early 1976. Gene Flath, who ran manufacturing through the crisis, called it “a masterful piece of recovery.”
If you read the first report in this series, you will know something similar happened in Japan. Japan’s supplier base passed the same kind of test in 1997, when a fire destroyed an Aisin brake-parts plant and 200 small firms brought production back in days.
The lesson is the same in both places. A strong cluster is one that can take a hard blow and recover.
One line from the Penang rebuild, from a Malaysian supervisor named Pin Yong Lai, became something close to a motto for the whole cluster.
”We do not ask for expanded responsibilities. We earn them.”
Pin Yong Lai, Intel Penang, during the 1975 rebuild. Quoted in Intel’s company history.
A company does not keep a thousand workers on full pay for a burned-down factory in a country it plans to leave. Intel stayed because, three years in, Penang had already shown it was worth staying for. Assam needs to understand this as a chip-packaging plant rises at Jagiroad.
A foreign plant is not a gift. It is an audition, and the audition runs for decades.
This report traces two answers to one question: how does a place earn a lasting place in the global chip industry?
Penang, the bankrupt island in this fire story, rented itself to foreign chip plants and slowly grew suppliers, engineers and listed firms of its own.
Singapore, the city-state about 600 kilometres south-east, took the opposite route and spent state money to become a part-owner of the industry itself. Both bets worked.
We want to highlight the Penang story through three people: the doctor who made the first decision, the American engineers who took the first risk, and the local engineer who left his job to build the machines his employer used to import.
The Singapore half is mostly about that one state agency that outlasted every government that set it up (we will talk about the Singapore ecosystem a bit later in this report).
Where this fits
This is the third report in Nitisagar’s series on how industrial ecosystems get built. The first covered Japan and the vertical keiretsu. The second covered Rajkot, a cluster that grew with no giant on top.
This one runs from Penang down the peninsula to Singapore. It is the case closest to Jagiroad, where a whole economy has to form around a foreign anchor. It builds on an earlier Nitisagar report on Penang and India’s OSAT packaging clusters, which mapped the plants as they stand today. The next report goes to Pune.
A doctor named Lim Chong Eu bet a bankrupt island on an industry that did not yet exist in Malaysia. That decision is why Intel came. It starts six years before the fire, at the moment the island lost its living.
For more than a century and a half, George Town had been a free port. Ships traded tin, rubber and spices through it without paying duty, and the port was the economy. In 1969 the federal government in Kuala Lumpur took that free-port status away. Unemployment climbed to 16.4%, more than double the national rate. Young people began leaving for the mainland and for Singapore, as the ISEAS and Penang Institute history From Free Port to Modern Economy records.
Lim, an opposition politician as well as a doctor, won the state in that same bitter year. He did not have a plan for a chip industry, because nobody did in 1969. So he commissioned a study.
The Nathan Report of 1970 gave him a clear answer. Stop trying to rebuild the port. Build factories for export instead. Start with simple, labour-heavy assembly, and move up to higher-value work later. Use the one thing the island still had in quantity: a young and reasonably educated workforce.
Lim took that advice and made a specific bet. Of all the industries the island could chase, he judged that electronics assembly could soak up thousands of semi-skilled young people the fastest. He pushed the federal government for the law to make it possible. In 1972 Malaysia’s first free-trade zone opened at Bayan Lepas, on the flat southern end of the island.
The deal was honest about what each side wanted. A foreign chip company got duty-free parts, a long tax holiday, serviced land it could build on at once, and a cheap, literate workforce it could train to microscope precision. Penang got jobs by the thousand, on an island that had just lost all of them. Lim was not selling a subsidy cheque. He was selling a zone that worked from day one, and a promise that trained workers would keep coming.
That mix, a zone that works on arrival plus a real promise on skilled labour, is almost exactly what Assam is trying to put together around Jagiroad in 2026.
Within two years of Intel landing in 1972, eight foreign electronics firms were running lines in Penang. It looked like a chip economy. It was not one yet.
Intel was the first big tenant, and the story of how it chose the island is almost comic. In early 1972 a consultant talked Andy Grove, then a senior figure at Intel, into looking at Penang. The site on offer was five acres of paddy field. By that November Intel had put about 1.6 million US dollars into building its first assembly plant outside the United States there. It called the plant A1. The ground was so wet that, by Malaysian press accounts, the car carrying Intel’s visitors got stuck in the monsoon mud on the way out. At first the plant did one job: assemble a single product, the 1103 memory chip. By 1975 it employed around a thousand people. That is why the fire that year hit so hard.
Penang folklore named the eight pioneers the Eight Samurai. The exact list changes in the retelling, but it always includes the firms below.
| Firm | Home | What it did in Penang |
|---|---|---|
| Intel | United States | Assembly and test of memory and logic chips, from 1972 |
| National Semiconductor | United States | Chip assembly and test |
| Advanced Micro Devices | United States | Chip assembly and test |
| Hewlett-Packard | United States | Components, later instruments and, in time, the machine-vision work that trained a generation of local engineers |
| Osram | Germany | Optoelectronics and lighting components |
| Litronix, Clarion, Hitachi | United States, Japan | Optoelectronics, car audio and electronics assembly |
The jobs came, as Lim had bet they would, and unemployment fell. For the first time, a generation of Penang women took formal factory work, on the wire-bonding and test lines the back end of the chip industry runs on. For a decade this looked like a clear success.
But it was still only foreign factories assembling imported parts, and a factory that imports everything can leave as easily as it arrived. What Penang did to de-risk itself over the next fifteen years is why this story is so relevant to Assam.
A foreign plant gives you wages. Left alone, it gives you nothing else, and Jagiroad is standing on exactly that edge.
South-East Asian industrial policy has a half-mocking name for the kind of factory that arrives, hires people to assemble imported parts, and puts down no roots: a screwdriver plant (if you are around Indian defence Twitter, you would know it as screwdriver-giri).
A screwdriver plant looks like real industry. It has a gate, a car park, and thousands of badges clocking in each morning. What it does not have is a single local supplier that learns anything from it. The design, the parts, the machines and the know-how all arrive in a shipping container and leave in another one. The local economy gets wages and little else. And the plant can move to a cheaper country the day its tax holiday ends.
For its first decade and more, much of Penang’s electronics base was exactly this, and the gap is measurable. A survey by the Malaysian Investment Development Authority found that in 1981, local suppliers provided about 1% of what the island’s chip firms used. By 1986, after a deliberate push, several Penang plants reported local content above 6%. The World Bank’s account of the industry notes that these links did form, but the move to higher-value work stayed slow for years. The scholar Rajah Rasiah has shown that most firms in Malaysia stayed stuck in assembly and test for a long time.
Those figures are a critical benchmark against Jagiroad, because the parallel is exact. On day one, the chip-packaging plant coming to Assam will import almost everything it uses:
- substrates, lead frames and bonding wire
- moulding compound and specialty chemicals
- every piece of assembly and test equipment
Nitisagar’s Jagiroad packaging-process component map, part of our wider semiconductor supply-chain map, sets out part by part how long that import list is. Left alone, the plant will employ thousands of people in Morigaon and buy almost nothing inside Assam. That is the 1981 Penang outcome in 2026 clothes.
The danger with a big anchor plant is not that it fails. It is that it succeeds in a narrow way, for decades, handing out jobs while leaving no supplier base behind it.
Penang climbed from 1% to a deep local industry over about forty years. The question for Assam is whether anyone will do the patient, dull work the climb takes.
Penang did not climb out of the screwdriver trap by accident. Two deliberate moves did most of the work, and both are ordinary.
The first was matchmaking, and it started with the Chief Minister himself. Lim personally introduced the island’s small metal-working workshops to the big foreign plants, so that an Intel or a National Semiconductor had a local firm to call for a bracket, a jig or a sheet-metal part. Penang’s industry histories treat this hand-to-hand work as the seed of local sourcing. It is not glamorous and it does not photograph well. But it is how the first 1% became 6%, one named part, one local firm, one audit at a time.
The second move came in 1989, and it is the most copyable institution in this entire series.
The Penang Skills Development Centre was set up by the firms, not the government. The big foreign companies and the larger local ones pooled money and equipment, lent their own engineers as trainers, and ran a shared centre that trained technicians for the whole cluster to one standard. The state and federal governments backed it, accredited it and helped fund some programmes. But the firms ran it and wrote the syllabus, because the firms knew which skills the lines would need next.
That order of control is the whole trick. A government training institute teaches what its lecturers happen to know. An industry-run one teaches what the hiring managers need next quarter, because the hiring managers hold the pen. PSDC has since trained hundreds of thousands of people and become the model that other states in the region try to copy.
Figure 1 · Fifty years, two near-misses, and the institutions that answered each one
Green marks the two moments a shallow assembly base could have ossified, and the move that pulled it up each time.
Dates from the ISEAS and Penang Institute history, the Bayan Lepas free-zone and PSDC records, MIDA local-content survey data, and company histories.
India is already testing a version of the PSDC model, and the clearest case came out of SemiconIndia. In April 2024 Lam Research signed a tripartite MoU with the India Semiconductor Mission and the Indian Institute of Science to put its Semiverse virtual-fabrication software into Indian classrooms. The firm donates about 241 crore rupees in licences and sets the curriculum, the mission pays for the infrastructure, IISc trains the trainers, and the ten-year aim is 60,000 engineers. That is the PSDC order of control reproduced almost exactly: the company that will hire decides what gets taught.
The Penang lesson for Jagiroad is to pull the same arrangement down to the anchor plant and its local suppliers, not to stand up a state polytechnic teaching a course the plant cannot use. Nitisagar’s semiconductor careers guide maps the roles that cluster will actually hire for, from line operators to process engineers.
Skills and introductions lift a cluster off the floor. A cluster has truly climbed only when it starts producing its own companies. The clearest Penang example is a man who once worked inside one of the foreign anchors.
Chu Jenn Weng grew up in Penang and studied at Universiti Sains Malaysia. A used film camera, handed to him during his studies, grew into a fascination with how a machine could be taught to see. An internship showed him a camera guiding a robot on a production line, and he found the field he wanted. He joined Hewlett-Packard’s Malaysian arm, and through the 1990s he led a small in-house machine-vision team. He was a local engineer trained to a world standard inside a foreign plant. That is the whole point of letting the foreign plant in.
In 1998 he and a university friend, Steven Siaw Kok Tong, left their jobs. Chu told The Edge the decision plainly.
”We knew the opportunity had arrived and we decided to quit our high-paying jobs.”
Chu Jenn Weng, co-founder of ViTrox, to The Edge Malaysia.
They set up ViTrox in 2000 with 20,000 ringgit of seed money, a very small sum to start a technology company. The early work was as unglamorous as any first-generation founder would recognise. “We used the initial capital to purchase parts and cameras,” Chu recalled, “then, we started to write software.” They “worked day and night, even during the weekend.” He admits that in those first years, “frankly, I didn’t know where all these ideas were going to lead me.”
Now, ViTrox makes automated optical inspection and test equipment, the machines that scan chips and circuit boards for defects. In 2004 it launched its first 3D vision inspection system for chip components. Today it is listed on the Malaysian exchange and sells to the same kind of global chip firms Chu once worked for, and he has appeared on the country’s rich list.
ViTrox is not the only local company. Penang now grows its own group of listed chip companies, including the assembly-and-test firm Inari Amertron and the equipment makers Pentamaster and Globetronics. They are the local depth that fifty years of introductions, training and vendor work built.
Figure 2 · The ladder Penang climbed
From assembling other people’s chips to designing and selling its own machines. Each rung took years, and a specific lever moved the cluster up it.
Composite view, drawn from the Penang cluster histories and company records. The rungs carry no precise dates; the climb ran from the 1970s to today.
Chu’s own words about his ambition are the thesis of this whole report.
”I’d want to build a great global technology company in Malaysia by Malaysians.”
Chu Jenn Weng, co-founder of ViTrox, to The Edge Malaysia.
He could only say that sentence because, thirty years earlier, a doctor let a foreign plant into a paddy field and then did the slow work of making sure a local engineer could be trained inside it. The anchor is the start of the story. Chu is the end of it.
Singapore made the opposite bet. Where Penang rented itself to foreign capital, Singapore used state money to buy its way into owning the most capital-heavy half of the chip industry.
It did not start there. In the late 1960s and 1970s Singapore too was a cheap-labour site for foreign electronics assembly, in almost the same place as Penang. The way it moved off that floor is the useful contrast.
The semiconductor industry splits roughly into two halves. The back end is assembly, test and packaging, the labour-and-precision work that Penang bet on. The front end is wafer fabrication, the capital-heavy business of printing circuits onto silicon in a cleanroom that costs billions. Penang chased the back end under foreign flags. Singapore decided to buy its way into the front end using the state’s own balance sheet.
Figure 3 · Two bets across the causeway
The same starting point in the early 1970s, two opposite halves of the chip industry. Penang took the back end under foreign anchors. Singapore took the front end with state equity.
Composite view, from InvestPenang and Singapore EDB material, the Bayan Lepas free-zone history, and the PSDC and Chartered Semiconductor records.
The institution that made this possible was the Economic Development Board, set up in 1961 and run from 1986 to 2001 by Philip Yeo. By later accounts, Yeo’s method was to find frustrated engineering executives inside the big American chip firms and offer them what a founder cannot easily pull together alone: land, power, water, tax holidays, a purpose-built plant, and state money for staff.
In return, they built a fab on Singapore soil. In 1987 Singapore went further and created Chartered Semiconductor Manufacturing, a wafer-fab company controlled by the state-linked Singapore Technologies group. The state was not offering an incentive to a foreign fab. The state was the fab.
The clearest example came in September 1998. Philips, TSMC and the EDB’s investment arm announced a joint venture to build a roughly 1.2 billion US dollar wafer fab in Singapore, later known as Systems on Silicon Manufacturing. The quotes from that day show who was in the room. Philip Yeo, as EDB chairman, said he was “very pleased with the announcement of this new joint venture between two leading semiconductor companies.” Morris Chang, the founder of TSMC, said his firm was “pleased to be building on our already strong relationship with Philips Semiconductors.” Arthur van der Poel of Philips put the stakes in one line: “Timing is critical with a major investment like this.”
Chartered listed on the Nasdaq and competed for years as one of the world’s three largest contract chipmakers, which is close to the hardest thing to pull off in semiconductors. Then, in 2009, Singapore sold it. Abu Dhabi’s sovereign investor, ATIC, bought Chartered at a firm value of about 3.9 billion US dollars, as FinanceAsia reported, and merged it into GlobalFoundries, the venture ATIC had spun out of AMD earlier that year. Singapore kept the fab, the jobs and the trained workforce, and let the brand go.
The result, two decades on, sets the ceiling on what state money bought. The island holds around twenty wafer fabs today, the densest cluster of fabs anywhere on earth, run by GlobalFoundries, Micron, UMC, TSMC, STMicroelectronics, Soitec and Infineon.
Not one of them is owned by a Singaporean company. Singapore has no chip champion of its own and no major chip-design house, and its Ministry of Trade and Industry still puts the city-state at about a fifth of global chip-equipment output and 5% of wafer-fab capacity. Writing in Swarajya, Raghavan S Rao calls this a rentier model, and does not mean it as an insult.
”The wafers are made in Singapore. The companies making them are based elsewhere. A rentier strategy, played beautifully by a country that knew its own ceiling.”
Raghavan S Rao, in Swarajya, on Singapore’s semiconductor playbook.
The lesson for Assam cuts two ways. State money can buy fabs, equipment share and good jobs, as it did for Singapore. But it does not reliably buy a home-owned champion, and Assam has nothing like the money to try in any case. What both places did build, underneath their opposite bets, is the thing worth copying.
The part of Singapore that does not travel to Assam
Assam cannot take equity in a wafer fab. It does not have the fiscal reserves, and it should not borrow to pretend otherwise. The chequebook does not transfer. What does transfer is the institution behind it: a technocratic agency, staffed by experts, that picked its targets, negotiated hard, and was left to run across political cycles without being overturned at every election. Singapore’s lesson for Assam is the EDB, not the fab.
Hold Singapore’s EDB beside Penang’s PSDC and the same asset appears under two different strategies. Each place built an institution that firms could trust to still be there, and still be competent, in ten years. Neither the back-end bet nor the front-end bet would have worked without it.
The two bets ended up in very different places, and the numbers make the case for the long timeline.
Malaysia now handles roughly 13% of the world’s chip assembly, test and packaging, the back-end work Penang chose in 1972. Most of it is in and around Penang. Malaysia is also the source of about a fifth of all United States chip imports, which came to around 16 billion US dollars in 2024 alone, by data cited by InvestPenang and reported in The Diplomat.
Singapore turned its front-end bet into the highest concentration of wafer fabs anywhere and about a fifth of global chip-equipment output, with new fab investment still arriving in 2026.
The newest wave shows the same pattern forming again. In a September 2026 analysis for the ASEAN+3 Macroeconomic Research Office, the economists Kian Heng Peh and Wee Chian Koh note that Asia’s AI-hardware supply chain is highly concentrated, a point a widely shared chart of ASEAN+3’s place in that value chain makes at a glance. Malaysia’s gains from the AI boom, they write, have come mainly from “the hardware surrounding the accelerator” rather than the accelerator itself. Advanced packaging, they argue, is the country’s next test. That is the same climb up the value chain, told again for the AI decade.
The chips themselves are arriving in huge numbers. Epoch AI, which tracks where the world’s AI chips physically arrive, estimates that from 2022 to mid-2026 the United States took about 72% of all AI compute and China about 10%. Among individual countries, Malaysia ranks third, ahead of Japan, India and every single European country. Malaysia is not making those chips. It is hosting them, which is the depth question again in a new form.
The data centres tell the same story. Across the causeway in Johor, the Malaysian group YTL, controlled by the Yeoh family, finished the country’s first Nvidia-powered AI data centre at Kulai in late 2025. It runs on Nvidia’s Blackwell processors and hosts YTL’s own AI cloud, part of a 10 billion ringgit investment. The campus is Malaysian-owned, which is real progress on the ownership question Singapore never answered. But the chips inside it, and the design they run on, are still controlled abroad. Fifty years after the first foreign plant in the paddy field, the depth question has not closed. It has only moved up the stack.
Two things about these figures matter for Assam.
- They took fifty years, not five. None of these numbers existed in 1985. First came 16.4% unemployment and the empty port, then a decade of shallow assembly at 1% local content, then the slow build of skills, suppliers and local firms. A share of a world industry came last.
- Both opposite bets worked. There is no single correct model. Penang went deep in the back end under foreign flags and grew firms like ViTrox underneath; Singapore went deep in the front end with state money and became a part-owner of the industry.
What the two shared was never a strategy. It was the willingness to hold one steady for decades, through an institution that did not blow away with the government of the day.
Start with the honest clock, restated from the first two reports in this series.
Penang took about fifty years to go from a bankrupt port to a chip economy with its own listed companies. The first fifteen of those years were shallow assembly that could have led nowhere. A government that promises the same depth in five years is not building depth. It is cutting a ribbon.
However, the process transfers. We can isolate seven levers that run through the Penang and Singapore stories, and every one is within a state government’s reach.
| # | Lever | What Penang or Singapore did | What Assam can do now |
|---|---|---|---|
| 1 | An enclave that works on day one | Bayan Lepas: duty-free, tax-light, serviced land built before the tenant arrived. | Make the Jagiroad zone genuinely turnkey. Power, water, effluent and customs live before suppliers move in. |
| 2 | Industry-run skills centre | PSDC, 1989, founded and governed by the firms, state as backer. | Build an industry-run skills centre with the anchor and its suppliers holding the pen on curriculum, as the ISM-Lam-IISc model already does. |
| 3 | Technocratic investment agency | EDB and InvestPenang: expert, stable, left alone across political cycles. | Give AIIDC or a dedicated cell the mandate, the specialists and the continuity to outlast an election. |
| 4 | Broker the introductions | Lim personally matched local metal shops to Intel and its peers. | Run a formal vendor-development programme off the Jagiroad component map, part by part. |
| 5 | Ratcheted, not unconditional, incentives | Benefits earned by raising local content and moving up the value chain. | Graduate Assam’s incentives on local-content and capability milestones, not just on arrival. |
| 6 | Pay for the first qualification | State money covered the audits and introductions that got local firms onto approved-vendor lists. | Ring-fence funds for the first supplier audit and certification, not only for buyer roadshows. |
| 7 | Make the anchor pull, by contract | Deals structured so the plant raised local capability, not just local headcount. | Attach supplier-development and payment-term commitments to the Jagiroad anchor’s agreements. |
The one lever to build first
If a government can do only one thing from this report, it is lever 2. An industry-run skills centre, governed by the anchor and its suppliers, funded partly by them, accredited and backed by the state. PSDC is the cleanest, cheapest and most copyable institution in this entire series, and the Lam-ISM-IISc MoU shows an Indian version of it already running. Build that one well at Jagiroad and the other six have somewhere to stand.
Now the honest part. Four things do not transfer, and Assam should stop wishing for them.
- Fiscal firepower. Singapore could take equity in a wafer fab because it held sovereign reserves. Assam cannot. Its bet is the back end and the supplier base, the Penang half of this story, not the Singapore half.
- Timing. Intel arrived in 1972, when there were almost no other low-cost offshore options. Assam enters a crowded field in 2026, against a Penang that is now deep and experienced.
- Geography. Penang and Singapore sit on the Strait of Malacca, the busiest shipping lane in the world. Assam sits behind the Siliguri corridor, and no industrial policy rewrites the container cost that follows.
- A cross-border labour market. Penang, Johor and Singapore share workers and vendors across the causeway in a way no single Indian state can reproduce inside its own borders.
Next in this series
Industrial Ecosystems 4 goes to Pune and the Pimpri-Chinchwad belt, and how a state land agency, a cluster of engineering colleges, and anchors like Telco and Bajaj built one of India’s deepest auto-component supplier bases over sixty years. Then South Korea, the Netherlands and Germany, before a closing report that turns the whole comparison into an Assam roadmap with dates.
As in the Rajkot report, the gap for Assam is not a missing scheme. It is a government that will pick up the right tool for the right job, and keep picking it up for a decade.
| # | Lever | Indian tool on the shelf | What the state has to do | Rough cost |
|---|---|---|---|---|
| 1 | Turnkey enclave | AIIDC’s estate mandate, plus central semiconductor-mission infrastructure support | Deliver power, water, effluent and customs at Jagiroad ahead of the suppliers, not on request. | Hundreds of crore, front-loaded |
| 2 | Industry-run skills centre | NSDC and the state skill mission, plus the ISM-Lam-IISc skilling model | Let the anchor and its suppliers govern it and write the curriculum. State accredits and co-funds. | Tens of crore to set up, shared with firms |
| 3 | Technocratic agency | State administrative action; a strengthened AIIDC | Staff it with specialists, give it continuity, protect it from the election cycle. | Low fiscal, mostly political will |
| 4 | Vendor introductions | MSE-CDP, plus the Jagiroad component map | Run a formal, part-by-part matchmaking and mentoring programme off that map. | Rs 10-25 crore per common facility, co-funded |
| 5 | Ratcheted incentives | Assam’s own IIPA 2019 incentive design | Tie continued benefit to local-content and capability milestones. A design task, not a fiscal one. | Zero extra fiscal; a redesign of the existing stack |
| 6 | First-audit funding | TIES and ZED, plus Assam’s 75% certification refund | Ring-fence money for the first supplier audit and vendor qualification. | Rs 2-5 crore a year at cluster scale |
| 7 | MoU discipline | The state’s own land-and-power negotiating position | Attach supplier-development and TReDS payment terms to the anchor’s agreements. | Zero fiscal. The Japan lesson, imported again. |
The honest promise that ends this report, restated from the first two in the series, is both modest and real.
Assam will not become Penang, and it certainly will not become Singapore. But over a generation, using tools already on the Indian policy shelf, it can build a back-end supplier base with real depth. The test is simple.
The day the Jagiroad plant needs a part, a jig, a calibration or a trained technician, the answer should be found in the Brahmaputra valley, not flown in from abroad.
The schemes above are mapped against the state’s own incentive stack in Nitisagar’s Assam industrial policy explorer.
Fifty years ago a doctor in a bankrupt port made that same wager on a paddy field. It is not a fast way to build an economy. It is the only one in this series that has ever worked.
Penang, the origin and the climb:
- Institute of Southeast Asian Studies and Penang Institute, From Free Port to Modern Economy: Economic Development and Social Change in Penang, 1969 to 1990, the authoritative account of the free-port loss, the Nathan Report, and the electronics pivot.
- Prema-chandra Athukorala, The Tale of the Penang Export Hub, Australian National University, on how the cluster formed and upgraded.
- Rajah Rasiah’s work on technological upgrading in Malaysian electronics, and the World Bank’s chapter on Malaysia’s electrical and electronics sector, for the local-content survey figures and the limits of functional upgrading.
- Economic History of Malaya, Penang’s industrialisation and economic transformation, 1960s to 1980s.
The people and firms quoted:
- Intel, the Penang plant history in its company archive, for the 1972 founding, the 1975 fire and the recovery, and the quotes from Ken Thompson, Gene Flath and Pin Yong Lai.
- The Edge Malaysia, Realising ViTrox’s vision, for Chu Jenn Weng’s account of leaving Hewlett-Packard and founding ViTrox, and every quote attributed to him here.
- TSMC, the 1998 Philips-TSMC-EDB joint-venture announcement, for the quotes from Philip Yeo, Morris Chang and Arthur van der Poel.
The industry today:
- Singapore Ministry of Trade and Industry, the 2024 speech on the semiconductor industry, for the equipment-output and wafer-fab-capacity shares.
- Malaysian Investment Development Authority data on the roughly 13% global assembly-test-packaging share, and InvestPenang on the United States import share.
- Epoch AI, the AI chip imports explorer, for Malaysia’s rank among individual countries for AI-chip compute arrivals.
- ASEAN+3 Macroeconomic Research Office, Advanced Packaging is Malaysia’s Next AI Test, by Kian Heng Peh and Wee Chian Koh, September 2026, on where Malaysia’s AI-hardware gains actually come from.
- FinanceAsia, the 3.9 billion US dollar ATIC buyout of Chartered Semiconductor, 2009, and Raghavan S Rao, Singapore’s semiconductor playbook is a rentier’s playbook, Swarajya, for the fab-density and ownership figures.
- W.Media, YTL Power completes its Nvidia-powered AI data centre in Johor, 2025, on the Kulai AI campus.
- The South First, Lam Research, the India Semiconductor Mission and IISc sign an MoU to upskill the semiconductor workforce, April 2024, the clearest Indian example of the industry-run skilling model.
- Charts on X that summarise the two datasets above: ASEAN+3’s place in the AI hardware value chain and Malaysia’s third-place rank in AI-chip compute arrivals.
Related Nitisagar work:
- Industrial Ecosystems 1: How Japan Built Its Supplier Base.
- Industrial Ecosystems 2: How Rajkot Built Its Foundry Cluster.
- The earlier report on Penang and India’s OSAT packaging clusters.
- The Jagiroad packaging-process and MSME component map.
- The two portals behind this series: the semiconductor supply-chain map and the semiconductor careers guide.
- The India semiconductor fact sheet and the Assam industrial policy explorer.
The moves that matter, before they cost you
Reports like this one are how we go deep on the shifts that decide which projects work in Northeast India. Every new report, briefing, and analysis in your inbox — original reporting, no syndicated fluff.
We'll map your project against IIPA 2019 and Assam's incentive stack, complimentary, no commitment.
Get a Free Pre-Assessment