Chartered Semiconductor: How Singapore Lost a Champion and Won an Industry

Singapore built a national foundry, could not make it profitable, sold it, and ended up with more semiconductor capacity than before. The Chartered Semiconductor story is the clearest illustration of how a country can lose a champion and win an industry. This case study is part of the semiconductors and advanced manufacturing pillar of the Singapore Company Stories hub.

Why did Singapore build a national foundry?

Chartered Semiconductor was founded in 1987 to move Singapore up the semiconductor value chain from assembly and test into wafer fabrication, capturing higher value activity and building domestic technical capability.

The foundry model, manufacturing chips designed by other companies, was still new. Taiwan had pioneered it and demonstrated that a country without chip design capability could participate in the industry as a manufacturing partner.

The state provided capital through Singapore Technologies and later Temasek, in a pattern identical to the government-linked company structure described in the GLC case study: build a commercial company, appoint a professional board, require a return.

Why did Chartered struggle?

Foundry economics reward scale and technology leadership. Chartered ranked third globally but was far behind the leader in both, forcing it to invest at the leading edge without the volume to amortise that investment across enough wafers.

The industry structure is punishing for the number three. The leader captures the most advanced, highest-margin work; the low-cost follower serves mature nodes profitably. A mid-position player faces the leader’s technology costs without the leader’s pricing power.

Chartered spent heavily on advanced process development, formed technology alliances to share costs, and still could not close the gap. Its financial results were volatile, with several loss-making years across the semiconductor cycle.

Why did Temasek sell it?

After sustained losses and a capital requirement that would only grow, Temasek accepted an offer in 2009 from an Abu Dhabi state investment vehicle that was assembling a foundry business with the scale to compete.

The buyer’s logic was consolidation. Combining Chartered’s capacity with the manufacturing operations spun out of a major American chipmaker created a foundry with sufficient scale to invest at the leading edge, which neither had alone.

The seller’s logic mirrored the Neptune Orient Lines decision seven years later: a business that cannot reach a defensible position should be sold to an owner who can get it there, regardless of national symbolism.

What happened after the sale?

The combined entity operated as GlobalFoundries, which eventually abandoned leading-edge development and repositioned around specialty and mature nodes, precisely the segment where its Singapore operations were strongest.

That repositioning turned out to be commercially astute. Demand for mature-node chips in automotive, industrial, communications and power applications proved large, durable and less capital-intensive than chasing the leading edge.

Singapore became the company’s largest manufacturing site, receiving multi-billion dollar expansion investment that added substantial wafer capacity. The country ended up with more fab capacity, more employment and more supply chain depth than the national champion had ever delivered.

What does GlobalFoundries do in Singapore today?

GlobalFoundries operates fabs producing specialty and mature-node semiconductors for automotive, internet of things, industrial, communications and secure device applications, having added significant capacity through a major expansion.

The specialty focus is deliberate differentiation. These chips do not need the smallest transistors; they need reliability, long product lifecycles, specific analogue or radio frequency characteristics, and supply security that automotive and industrial customers require for a decade or more.

That customer profile is also less cyclical than consumer electronics. Automotive and industrial buyers commit to long qualification cycles and do not switch suppliers casually, which produces more predictable utilisation than the consumer-driven segments of the industry.

What is the lesson from Chartered’s story?

The lesson is that national capability and national ownership are different objectives, and confusing them is expensive. Singapore wanted semiconductor manufacturing capability; it assumed that required owning a foundry, and it did not.

Selling the company transferred the losses and the capital burden to an owner better placed to carry them, while the fabs, the workers, the suppliers and the ecosystem stayed. The country retained everything that mattered economically and shed what did not.

This distinction matters far beyond semiconductors. Governments and corporate groups routinely conflate strategic importance with the need for ownership, and end up funding losses to preserve a flag rather than a capability. Related decisions across the Singapore Company Stories hub show the same principle applied repeatedly.

How does the foundry business model work?

A foundry manufactures chips designed by other companies, selling manufacturing capacity rather than products. Customers range from small design houses to large fabless semiconductor firms that no longer own factories.

The model separated design from manufacturing and made the semiconductor industry vastly more competitive, because a startup could design a chip without building a billion-dollar factory. That separation created most of the modern chip industry.

For the foundry, the economics depend on utilisation. Fabs have enormous fixed costs and must run near capacity to be profitable, which is why foundries compete aggressively for volume and why downturns are so painful.

What are specialty and mature nodes?

Mature nodes are older process technologies that remain in high demand for automotive, industrial, power management, analogue and communications chips that do not require the smallest transistors. Specialty processes add particular electrical characteristics.

These chips represent a very large share of total units shipped globally, even though they attract far less attention than leading-edge processors. A modern car contains hundreds of them; almost none are leading edge.

The business is attractive because equipment is already depreciated, processes are qualified, and customers commit to long lifecycles. It is less attractive when everyone builds mature-node capacity at once, which is the current concern.

Did Singapore lose anything by selling?

It lost national ownership, the head office decision-making that comes with it, and the option to direct the company’s strategy. It retained the fabs, the employment, the supplier ecosystem and the tax base.

Head office loss is not trivial. Corporate headquarters concentrate high-value functions including finance, strategy, legal and senior engineering, and their departure removes career paths that a manufacturing site alone does not provide.

On balance the trade was favourable, because the alternative was continued state funding of losses in a business with no path to a defensible position. But acknowledging what was lost makes the analysis honest rather than merely triumphant.

What is the current competitive landscape in foundry?

The foundry market is highly concentrated at the leading edge and more fragmented in mature and specialty nodes, where several operators compete alongside integrated device manufacturers selling spare capacity.

Chinese mature-node capacity expansion has been particularly significant, adding substantial supply in segments where established players had expected stable pricing.

The result is a bifurcated industry: extraordinary profitability at the leading edge, and increasingly competitive conditions everywhere else, which pressures exactly the segment Singapore’s fabs occupy.

How should governments think about national champions?

The Chartered experience suggests a national champion is justified only where ownership itself provides something capability alone cannot, such as guaranteed supply in a conflict or control over technology that would otherwise be denied.

Where the objective is employment, skills, tax base and ecosystem, foreign ownership delivers those outcomes as well or better, without the state carrying the losses.

The harder question is what happens in a genuine crisis, when a foreign owner’s home government restricts what its Singapore facility may produce or ship. That is a real risk and it is the strongest argument for domestic ownership in strategic sectors.

What does this mean for other countries building chip capability?

The practical implication is to be explicit about the objective. If the goal is jobs, skills and supply access, attracting foreign fabs achieves it faster and cheaper than building a national champion from scratch.

If the goal is genuine technological sovereignty, ownership matters, but the cost is enormous and the probability of reaching a competitive position is low without a very large domestic market.

Countries that pursued the second objective while budgeting for the first have generally produced expensive, subscale operations that never reached viability, which is the failure mode the Chartered story narrowly avoided by exiting.

How did employees and the local ecosystem fare?

Manufacturing employment continued and subsequently expanded under the new owner, and the supplier ecosystem that had grown around the fabs remained in place and grew alongside the additional capacity.

Senior corporate roles were the main loss, since strategic decision-making moved to the parent company’s headquarters, which reduced the range of career paths available locally.

The engineers and technicians, who represent the accumulated national capability, largely stayed in the industry, which is the outcome that mattered most for the country’s long-term position.

What would have happened without the sale?

The counterfactual is continued state funding of a subscale foundry through several more capital cycles, with escalating investment requirements and no realistic path to closing the gap with the industry leader.

The opportunity cost is what makes the case decisive. Capital committed to defending a losing position is capital unavailable for the investments that subsequently built Singapore’s position in equipment, packaging and specialty manufacturing.

That reallocation argument is the strongest defence of any divestment decision, and it is the reasoning applied consistently across the exits documented in the Singapore Company Stories hub.

What is the geopolitical dimension today?

Semiconductor manufacturing has become an instrument of national security policy, with export controls, investment screening and subsidy programmes all deployed to shape where capacity sits and who can access it.

A foundry’s ownership nationality now affects which customers it can serve and which equipment it can buy, considerations that simply did not exist when Chartered was sold.

That change strengthens the argument for domestic ownership in ways the 2009 analysis could not anticipate, and it is a fair caveat to any confident retrospective judgement about the decision.

Could Singapore build another national champion?

It could, but the conditions would have to be unusual: a segment where ownership confers access that partnership cannot, a defensible technical niche, and a realistic path to scale without a large domestic market.

Advanced packaging, specialty materials and semiconductor equipment components are the plausible candidates, since all reward engineering depth rather than raw capital scale.

The more likely path remains attracting and anchoring foreign operations while building domestic capability in the supplier layer, which delivers most of the economic benefit with far less state capital at risk.

Frequently Asked Questions