Loading summary
A
In July 2025, intel announced that they will be gradually closing their assembly and test site in Costa Rica. End of an era. Intel has been in Costa rica for almost 30 years. That a and T factory was their only Latin American manufacturing site. In 2000, nearly 40% of Costa Rica's exports were intel microprocessors. They were a chip export giant. Kinda. In today's video, let us look back at Intel's tenure in Costa Rica. I've not mentioned Costa Rica before, so I think we are up for an introduction. Costa Rica sits between Nicaragua and Panama on the isthmus connecting north and South America. To the east we have the Caribbean Sea which is part of the Atlantic Ocean. And to the west there is of course the Pacific Ocean. Size wise, the country is a bit larger than Denmark and about 42% larger than Taiwan. The interior is mountainous with large ranges in the region's southeast. Rivers flow down from those mountains to the plains. Today the population is about 5 million people. Most of them live in the country's urbanized Central Valley, which include the capital San Jose. Costa Rica is Latin America's oldest democracy. In the 1500s the country was a backwater colony in Spain's American empire. Lacking silver, gold or indigenous slaves, the resident colonists were poor and most farmed their own land. Costa Rica gained its independence from Spain in the 1800s. Going into the first half of the 20th century, the country was a small, elite dominated republic that mostly exported coffee and bananas. In 1948, election disputes led to a brief civil war won by a coffee farmer turned political activist named Jose Figueras Ferrer. Ferrer passed a new constitution, nationalized the banks and quite famously disbanded the military. For the next few decades, the Costa Rican economy continued to focus on agriculture and state led spending. But after a debt Crisis in the 1980s brought on by excessive foreign currency borrowing, the government felt reform was necessary. They passed policies that were more market oriented, set up free trade zones for tax free import export of goods, and sought for major foreign corporate investors. Foreign direct investment started to grow from about 60 million or so a year in the mid-1980s and to a quarter of a billion a decade later. Companies like Motorola started building assembly plants there to make phones. In 1995, intel was firing on all cylinders, generating over $16 billion in revenue and $3.6 billion in net income. The booming PC industry was driving immense demand growth for CPUs and since intel allowed no one else to manufacture their CPUs they they had to build many new fabs to meet this demand. In addition to their US fabs in California, Arizona, New Mexico and Oregon. They also had factories in Malaysia, Israel, Ireland, the Philippines, and the People's Republic of China. All were growing. With sales continuing to surge, intel needed yet more capacity. This, combined with their desire to continue their global expansion, led the decision to build a new assembly and test factory in either Latin America or Asia. Assembly and test refers to a series of steps done after the wafer is fabricated. First, we need to cut up the wafers to get 300 to 500 microprocessor dies in them. After that, each chip is tested. A sophisticated machine called a tester sends digital signals into the chip and measures the signals that the chip sends back. If they match spec, then it passes. If it passes and the chip is packaged into a plastic or ceramic housing for protection from the surrounding environment, after one final set of tests, the chip is ready to go. Assembly and test plants are not as pricey as front end wafer fabs, but nevertheless still represent a significant investment. If a wafer fab then cost about 1 to $1.5 billion, then the assembly and test factory might cost anywhere from 100 to $300 million. Unlike with Waferfab, assembly and test remains rather labor intensive. Labor takes about 25 to 30% of the total operating cost, making it a key factor in determining the plant's economic viability. So where should intel put this assembly and test factory? Early on in the process, intel decided that it should expand to a new country. Their current overseas sites in Malaysia, Ireland and so on were already at capacity limits. Intel also did not want more than 30% of their asset base in any single country. The site selection team began with a list of 13 countries in Asia and Latin America. Each country was chosen for having some mix of a pro business attitude, stable political regime, decent logistics and manufacturing lead time, fast permitting and good human capital. In terms of Asia, the list had the prc, India, Indonesia, Korea, Singapore, Thailand and Taiwan. Then in Latin America, there was Argentina, Brazil, Chile, Costa Rica, Mexico and Puerto Rico. At this point, Costa Rica was not seen as a major contender. They were only being considered in the first place. Almost as an afterthought. The job of attracting FDI is handled by a non profit, non political organization called the Costa Rican Investment Promotion Agency or Cinda, founded in 1982. Throughout the decade they lobbied for mostly textile firms. But by the late 1980s, Costa Rican labor costs had gotten too high to be competitive in that area. So in 1993, the agency shifted to lobbying electronics firms like Intel. After two years of nothing happening. Intel finally responded in late 1995, inviting Cynda to a meeting at their Santa Clara headquarters. A few months later, Cynda sent over a detailed investment package which got them onto the list. After selecting their basic list, intel officials did a first round review, searching for any potential deal breakers in their important criteria. This cut things down to seven countries. Indonesia, Thailand, Brazil, Argentina, Chile, Costa Rica and Mexico. Then, because intel already had substantial facilities in Southeast Asia, Indonesia and Thailand were dropped from the list. Argentina was also removed for undisclosed reasons. So in the end, just the four Latin American countries made it to the shortlist. Now it was time to do on site visits. For Intel's first trip in April 1996, Sinda arranged meetings with banks and accounting firms to discuss legal and financial transparency, then with existing electronics manufacturers to discuss actual on the ground experiences. Sinda also arranged a meeting with Costa Rica's then president, Jose Maria Figueras. The president took a keen interest in a potential intel investment and assigned senior government officials to quickly respond to requests. At the end of the visit, the four Intel VPs were honest about their biggest worry. Former Foreign Trade Minister Jose Rossi recalled at the end of the visit, they diplomatically told us that while one day they might have something in Costa Rica, their project was too big. It would be like putting a whale in a swimming pool. Can Costa Rica's infrastructure capacity, power, air freight and such handle an intel estimates found that a fully laden assembly and testing plant would use 30% of the whole country's energy capacity. In the same line, were there enough educated technical workers. Costa Rica's education system was rated as quite excellent, but there were gaps in English, physics and chemistry. It wasn't semiconductor ready. Nevertheless, intel let Costa Rica stay in the race and slowly other countries got dropped. Chile fell out because they had little experience in electronics manufacturing and the flights to and from there were not ideal. And Brazil lost out due to its investment climate. The government was getting more pro business and things were promising. However, interest rates remained high and the bureaucracy heavy. So in the end that left just Costa Rica and Mexico. Mexico offered a better financial package, lower energy prices and free land. But the country, having recently exited a currency crisis, suffered high inflation. And perhaps most damningly, Mexico had mandatory union rules. Mexican officials then offered to waive those rules just for intel, which apparently gave the company the heebie jeebies. Not that Costa Rica didn't also offer benefits. The government billed a water facility for Intel's exclusive use expedited permit approval, granted an eight year exemption on taxes with another 50% discount for the four years afterwards and provided electricity at cost. But they said what was offered was no different from what was offered to other electronics companies of similar size anyway. In November 13, 1996, intel announced they had chosen Costa Rica, feeling most comfortable with the country's political stability, legal institutions, and economic openness and transparency. Construction on the $300 million plant, called A6T6, began in late April 1997. President Figueres attended the groundbreaking ceremony, the plant's first site. General Manager Cesar Quezon gave remarks, saying, to appreciate the future is to remember the event taking place today, as today marks the beginning of a new era in Costa Rica. The plant was completed on time and began operating in November 1997. By 1999, intel had invested between 300 and $500 million into the 50 hectare facility. The factory sits about 15 kilometers west of San Jose and was a clone copy of its sister facilities in Malaysia and the People's Republic of China. Intel's sheer size materially changed several headline macroeconomic numbers. Costa Rica's total 1998 GDP was $13.6 billion, so the central bank attributed over half of 1999 GDP growth and 40% of total exports that year to intel alone. ASICS T6 initially assembled and tested about 20% of Intel's Xeon server processors, as well as its desktop Celeron and Pentium Pro microprocessors. What? No itaniums? In 2000, intel added an engineering services group to Costa rica, hiring about 100 specialized engineers to provide internal engineering support as well as software development. Then in 2003, intel announced a major facility expansion, investing $110 million to add a third facility, called CR3, to assemble and test chipsets. These are chips on the motherboard that manage the flow of data between the CPU and its external components. They serve as the bridge through which the CPU communicates with the rest of the system. In terms of job creation, a 6T6 first phase directly created over 2,200 jobs, reaching this number by mid 1999. CR3 added 600 more jobs. The vast majority of the workers were local. Considering that the country's population at the time was 3.5 million and its total workforce was about 1.3 million, that is pretty significant. Half of those workers were operators, a quarter were technicians, and the rest professionals. Costa Rican workers received training at Intel's plant in Arizona for their tasks so that they can be as good as as any other intel worker. Intel said that it paid wages higher than average for the Costa Rican manufacturing sector, about 615 per month compared to $406 per month. But intel also hired for more qualified, higher earning workers, so this shouldn't be surprising. A major perk mention regarded intel suppliers moving to Costa Rica. Intel had about 200 suppliers of varying size, with about 60 marked as being very important. It was initially estimated that these companies moving would bring over $500 million of investment and maybe create a surrounding cluster. There were a few wins. Intel's influence pushed FedEx, UPS and other air freight companies to initiate regular cargo flights with the United States. Obviously, wafers and chips have to be shipped in and out regularly. Unfortunately, it turns out that many of Intel's suppliers hesitated to come. Those who did only opening small service centers with few employees. Generally, they were hesitant to invest deeply into the country without another big Anker customer present. Did intel replace those foreign suppliers with local ones? With regards to the assembly test, factory localization was limited. For one thing, the primary inputs for assembly and test are silicon wafers, which had to be imported. With regards to other things, there were other issues. Potential local suppliers were often small or medium sized businesses, so they struggled to meet Intel's high manufacturing demands. For instance, the demand for 24. 7 response time. Thusly, the top three things that Intel's factory bought from the local ecosystem were electricity, maintenance services and nitrogen. Perhaps the biggest intangible benefit was intel putting Costa Rica on the map as a viable site for investment. A 1999 survey found that 72% of respondents now considered Costa Rica a viable investment candidate thanks to Intel's decision. It was basically a big stamp of approval, and intel encouraged Cynda to go out and find other big fish. Cynda tried to take advantage, launching a big campaign. This eventually brought Procter and Gamble and Abbott Laboratories. Both multinationals, set up major administrative hubs in Costa Rica and remain there to this day. Unfortunately, the Asian financial crisis and the crash of the telecom bubble killed intriguing FDI opportunities with the hard drive maker Seagate and telecom equipment maker Lucent. After over a decade of operation, intel decided to invest yet more into Costa Rica. In 2011, intel opened a new engineering design center, hiring 300 Costa Rican engineers and technicians for R and D of advanced hardware and chipsets. By now, the company had invested in $900 million total. But then in the early 2010s, the whole PC market saw major downturns in demand, driven most likely by the red hot boom in smartphones and tablets. A downturn that the terrible Windows 8 could not reverse. With PC shipments cratering, intel decided to pivot resources to Mobile chips and the Internet of things. In January 2014, they said they would lay off 5% of their workers as part of this pivot, targeting various offices across the United States, Spain and Malaysia. Thusly, intel later announced that that they will close the Costa Rica assembly and test facility, laying off 1,500 workers by the end of that year. The work would be consolidated to existing sites in China and Southeast Asia. The company said this had nothing to do with the productivity of Costa Rican workers, adding that it would keep and even expand the area's engineering, design services and R and D centers. By then, these offices employed about 1,200 people in total. In the early 2000 and tens, microprocessors remained Costa Rica's single biggest export item, about $2.4 billion worth, or 20% of 2013 total exports. Those will eventually go to zero. Fortunately, Costa Rica had since developed a strong cluster of medical devices manufacturing exports, along with its agricultural staple exports of beef and bananas. They also retained strong tourism and business services sectors, so the loss of semiconductors fortunately wasn't a crippling one. Intel and CINDA work together to help the 1,500 laid off workers find new jobs, opening a website to aggregate resumes. Considering that manufacturing turnover is usually pretty high, it seems most workers quickly found new jobs. One possible upside of the intel factory's closure was how it raised the proportion of goods intel bought from domestic suppliers, some of which can be deemed higher value, for example software development as well as R and D services like consulting and training. So you can say that the closure represented some kind of a healthy ish transition. Then in 2019, the COVID pandemic hit. The world locked down and almost everyone except the essential personnel had to go online. This meant that everyone had to go out and buy a PC and suddenly everyone got on zoom calls. Now intel needed all the PC and server chip capacity that it can get. They never left the Costa Rica campus, so the facility was apparently sitting there. I also reckon intel was concerned about having so much of its supply chain across the Pacific in Southeast Asia and China. Moving some of that to Latin America can achieve near shoring shorter freight times while retaining somewhat cheaper labor costs. So six years after the Costa Rica assembly and test plant first closed, intel said it would invest about $350 million to reopen the facility. Iana Rojas, general manager of Intel Costa Rica, said the existing infrastructure synergy with the test operations that already exist on site, the talent, the free trade zone regime and legal environment gave Intel a favorable option to start its assembly capacity in Costa Rica. The reopening will create 200 new jobs and operations will kick off in the second half of 2021. Cinder reported that the first exports started heading out in September. As the chip shortage or so called chip crisis developed, policymakers desired a more global and secure semiconductor supply chain. So the US government explored potential partnerships with Costa Rica as a potential chip manufacturing hub. In July 2023, the US identified Costa Rica as one of the six partner countries to receive funds from its International Technology, Security and Innovation Fund, a fund established by the US CHIPS act to boost domestic semiconductor manufacturing and research. Soon after Intel, Costa rica announced another $1.2 billion in investments over the next two years to upgrade infrastructure and technical engineering. That was in 2023. I think there was a lot of optimism in Costa Rica as late as October 2024 that it can benefit from what felt like a sustainable, near shoring and de risking drive to build more semiconductor factories, maybe even a fab in the country. But a lot can change in two years and unfortunately things turned in a negative way. Intel since experienced massive downturns due to Covid hangovers, an overextended expansion drive and AI GPUs taking spend share away from CPUs. For a while they teetered on the brink after Lipp Bhutan, an American born in Malaysia, took over as CEO. He appeared to have decided that ANT was better off consolidated in Southeast Asia after all, or on American soil. If the closure goes through after 2026, intel will no longer have any assembly and test facilities in Latin America. It's a little disappointing considering how things on the manufacturing side seem so promising in the early 2000s. But Intel's other Costa Rican operations will continue their R and D center for testing and validating chip hardware and software, plus their corporate HR Administrative Support Services center. Between those two, intel employs over 2,000 Costa Ricans and those are great jobs. And who knows, there's a big CPU sales surge thanks to the AI agents Boom. Maybe intel will need the capacity again and postpone the factory's closure once more. After 30 years in Costa Rica, Intel's working relationship there remains super strong. Alright everyone, that's it for tonight. Thanks for watching. Subscribe to the Channel, Sign up for the Patreon and I'll see you guys next time.
Podcast Summary: Asianometry — “Intel’s 30 Years in Costa Rica”
Host: Jon Y
Date: May 21, 2026
This episode traces Intel’s near three-decade presence in Costa Rica: from the assembly and test (A&T) factory’s founding in the 1990s, through its massive impact on the Costa Rican economy and society, repeated downsizings, closure and reopening, to the final announcement in 2025 that Intel will again close its only Latin American manufacturing site. Host Jon Y explores how this partnership helped redefine Costa Rica as a technology hub, analyzes the push and pull factors behind Intel’s location decisions, and reflects on the broader implications for both Intel’s business and Costa Rican development.
[00:02 – 06:45]
“If a wafer fab then cost about 1 to $1.5 billion, then the assembly and test factory might cost anywhere from 100 to $300 million. Assembly and test remains rather labor intensive. Labor takes about 25 to 30% of the total operating cost.” — Jon Y [05:00]
[06:45 – 13:20]
“...their project was too big. It would be like putting a whale in a swimming pool. Can Costa Rica’s infrastructure...handle an Intel factory?” — Jon Y, on initial apprehension from Intel execs [10:30]
[13:20 – 19:20]
“It was basically a big stamp of approval, and Intel encouraged Cynda to go out and find other big fish.” — Jon Y [19:00]
[19:20 – 23:00]
[23:00 – 30:00]
“In January 2014, they said they would lay off 5% of their workers...thusly, Intel later announced that they will close the Costa Rica assembly and test facility, laying off 1,500 workers by the end of that year.” — Jon Y [28:10]
[30:00 – 38:20]
“There was a lot of optimism in Costa Rica as late as October 2024 that it can benefit from what felt like a sustainable, near-shoring and de-risking drive to build more semiconductor factories, maybe even a fab in the country.” — Jon Y [37:45]
[38:20 – 41:30]
“After 30 years in Costa Rica, Intel’s working relationship there remains super strong...And who knows, there’s a big CPU sales surge thanks to the AI agents boom. Maybe Intel will need the capacity again and postpone the factory’s closure once more.” — Jon Y [41:30]
On Costa Rica’s Initial Suitability:
“Their project was too big. It would be like putting a whale in a swimming pool.” — Jon Y [10:30]
On Intel’s Macro Impact:
“Costa Rica’s total 1998 GDP was $13.6 billion, so the central bank attributed over half of 1999 GDP growth and 40% of total exports that year to intel alone.” — Jon Y [16:35]
On Local Impact vs. Hopes for a Cluster:
“Those who did [come] only opening small service centers with few employees. Generally, they were hesitant to invest deeply into the country without another big anchor customer present.” — Jon Y [18:43]
On the Enduring Relationship:
“Intel’s other Costa Rican operations will continue their R&D center for testing and validating chip hardware and software, plus their corporate HR Administrative Support Services center. Between those two, intel employs over 2,000 Costa Ricans and those are great jobs.” — Jon Y [41:20]
Jon Y closes by reflecting on the broader arc of Intel in Costa Rica—a story of how a small country's focus, agility, and openness won over a global giant and changed its own economic destiny. Even as manufacturing winds down, the knowledge and service sectors Intel catalyzed continue to anchor high-quality jobs and stable bilateral relations.
“After 30 years in Costa Rica, Intel’s working relationship there remains super strong...Maybe Intel will need the capacity again and postpone the factory’s closure once more.” — Jon Y [41:30]
For anyone interested in economic development, semiconductor supply chains, or how national policy can change a country’s destiny, this episode is rich with insight and context.