FPT: From Food Processing to a Billion-Dollar Software Exporter

FPT is the clearest example of a Vietnamese company that changed what it did three times and survived every change. It has been a food-technology firm, a computer importer, an internet provider, a software outsourcer and now an AI and chip-design aspirant, and at each step the leadership sold the previous identity to fund the next one. This article explains how the software business went from an embarrassment inside the group to its engine, why Japan mattered more than the United States, and what the company has to get right to reach its next target. It is part of the Vietnam Company Stories hub.

How did a food-processing company end up writing software?

FPT ended up in software because its founders were scientists who took whatever business the early market allowed, and computers were the first one that scaled. The 1988 name reflected an initial plan to process food for Soviet export that never materialised.

The company was established on 13 September 1988 as the Food Processing Technology Company, a state-affiliated venture under the Institute of Technology Research, by thirteen scientists led by Trương Gia Bình, a physicist trained in Moscow. Vietnam had just begun the Đổi Mới reforms; foreign trade was being liberalised and technically educated people were looking for commercial outlets. Food technology was a plausible first idea, and it went nowhere.

What worked was computers. In 1990 the company renamed itself the Corporation for Financing and Promoting Technology, keeping the FPT initials, and became a distributor for foreign hardware vendors, notably IBM, Compaq and later HP. Distribution gave it cash flow, a national sales network and the first generation of managers who understood commercial discipline. Through the 1990s FPT was, in substance, Vietnam’s largest computer dealer with a systems-integration arm attached.

The founders were never satisfied with that. Trương Gia Bình, whose story is told in the profile of FPT’s founder, saw India’s software exporters growing on the other side of the Bay of Bengal and asked why Vietnam, with its mathematics tradition and low wages, could not do the same. The answer, in 1999, was to try.

Why did the first years of FPT Software nearly fail?

FPT Software nearly failed because Vietnam had no reputation, no English-speaking engineering base and no sales presence in the markets it targeted. The unit lost money from 1999 to about 2002 and was kept alive by the group’s distribution profits and the chairman’s insistence.

The launch plan was modelled on India: sell offshore development to American and European clients at a fraction of onshore rates. FPT opened a small office in Bangalore to learn from Indian firms and sent salespeople to the United States. The results were poor. Prospective clients had never heard of Vietnam as a software location, the dot-com crash of 2000 and 2001 dried up discretionary spending, and FPT’s engineers, though technically strong, lacked the English and process maturity that clients expected.

Inside the group the unit was contentious. Distribution managers questioned why profits from selling servers were being spent on a venture with no revenue. Nguyễn Thành Nam, the software unit’s first leader and later group CEO, has described those years as a period of repeated near-closure. The board kept it going largely because the chairman was unwilling to concede the thesis.

The turning point was a change of target market. Around 2001 and 2002 FPT began winning small contracts from Japanese companies that were looking to reduce their reliance on Chinese vendors and were willing to work with a new supplier if it invested in Japanese-language capability. FPT did, hiring Japanese teachers, sending engineers to Tokyo and building a bridge-engineer model in which bilingual coordinators sat between Japanese clients and Vietnamese developers. That model became the company.

How does the Japan-first strategy actually work?

The Japan strategy works by embedding bilingual bridge engineers at the client, delivering through large offshore teams in Vietnam and accepting the long, relationship-driven sales cycles Japanese companies expect. FPT Japan is now one of the larger foreign IT vendors in the country by headcount.

FPT Japan was formally established in 2005 and has grown into a subsidiary with thousands of staff in Tokyo, Osaka, Nagoya and other cities, most of them Vietnamese engineers who have learned Japanese to a professional level. The company has run its own Japanese-language programmes for two decades and, through FPT University, now trains students in Japanese from the first year of an engineering degree. The supply of bilingual engineers is the real moat; it takes years to build and competitors cannot buy it.

The client base reads like a list of Japan’s manufacturing and financial establishment: automotive suppliers, electronics conglomerates, insurers and trading houses. Much of the work is unglamorous maintenance and modernisation of legacy systems, which is precisely the work Japanese firms struggle to staff domestically as their engineering workforce ages. FPT’s pitch is not innovation but reliable, lower-cost capacity that speaks the client’s language.

Japan still accounts for on the order of 40 percent of FPT Software’s overseas revenue, with the Americas, Europe and Asia-Pacific making up the rest. The concentration is a strength in a downturn, because Japanese clients rarely switch vendors, and a risk in a currency crisis, because the yen’s fall from 2022 onward cut the dollar value of Japanese contracts even as volume grew.

What happened when FPT crossed $1 billion in software exports?

When FPT Software’s revenue from foreign markets passed $1 billion in 2023, the company reset its ambition to $5 billion by 2030 and shifted its pitch from outsourcing to AI, automotive and semiconductor services. The milestone was as much about credibility as money.

The billion-dollar mark mattered because it placed FPT in the tier of mid-sized global IT services firms and made it visible to clients that would not previously have considered a Vietnamese vendor. Group revenue in 2024 was on the order of VND 63 trillion, about $2.5 billion, with pre-tax profit of roughly VND 11 trillion. Technology services, which include the export business and domestic systems integration, contribute more than half of revenue and a larger share of growth.

The company used the moment to buy capabilities it lacked. Acquisitions since 2014 include RWE’s IT unit in Slovakia, the American consultancy Intellinet, the Colorado engineering firm Cardinal Peak, the French consultancy AOSIS and the Japanese firm Next Advanced Communications, each adding domain expertise or client relationships in a specific market. None was large; the pattern is bolt-on rather than transformational, which suits a company that grows organically and is wary of integration risk.

The Nvidia partnership announced in April 2024, under which FPT committed roughly $200 million to build an AI factory using Nvidia hardware and to train engineers on its platform, was the most visible signal of the new direction. FPT also established a semiconductor design unit in 2022, positioning itself downstream of the assembly and test capacity described in the story of Intel’s Saigon plant. Whether Vietnam can move from packaging chips to designing them is an open question; FPT has decided to be the company that tries.

How do FPT’s other businesses support the software story?

FPT’s telecom, education and retail units supply cash, talent and a domestic customer base. FPT Telecom is a leading fibre broadband provider, FPT Education trains tens of thousands of engineers, and FPT Retail runs the fast-growing Long Châu pharmacy chain.

FPT Telecom, started in 1997 as one of Vietnam’s first internet providers, is now a fibre-broadband and pay-TV operator with a national footprint and steady margins. It competes with Viettel and VNPT for household connections, and while it will never match Viettel’s scale, it produces reliable cash and a data-centre business that supports the group’s cloud offerings.

FPT Education is the strategic piece. FPT University, launched in 2006, and the network of schools and vocational colleges around it enrol a student body in the low hundreds of thousands across all levels. The engineering programmes are designed around the group’s needs, including Japanese-language tracks and, more recently, AI and semiconductor curricula. Few IT services firms anywhere own their own pipeline of graduates; FPT does, and it is a large reason it can add thousands of engineers a year without wage inflation destroying margins.

FPT Retail was once a phone-shop chain competing with Mobile World, and it has since reinvented itself around Long Châu, a pharmacy chain that grew from a few dozen outlets to more than a thousand within a few years and is now the retail unit’s growth engine. The retail business is not core to the software story, but it demonstrates the group’s habit of reallocating capital toward whatever is working.

What does FPT’s history teach founders and investors?

The lesson is that a company can change its core business repeatedly if it keeps a cash-generating base, tolerates years of losses in the new venture and builds the talent pipeline itself. The corollary is that this only works with founder-level patience and control.

For founders, the transferable idea is sequencing. FPT never abandoned a profitable business to chase the next one; distribution funded software, software funded education and telecom, and all of them now fund AI and chips. Companies that pivot by cutting the old business too early lose the cash flow that makes the pivot survivable. FPT’s software unit lost money for three to four years and was tolerable only because the group had other income.

The second lesson is about markets. FPT succeeded in Japan, a market Indian firms found difficult, because it invested in a specific capability the market valued and competitors ignored. A challenger from a country with no reputation is better off dominating a niche that incumbents underserve than competing on the same terms in the largest market. The company’s American business, built later and partly through acquisition, has grown but has never had the depth of the Japanese franchise.

For investors, FPT is one of the few Vietnamese listed companies with a long record of disciplined capital allocation, transparent reporting and steady dividends, which is why it has been a core holding for foreign funds since its 2006 listing on the Ho Chi Minh Stock Exchange and why it trades at a premium to the market. The risk is that the premium assumes the AI and chip bets will succeed as the outsourcing bet did, and the base rates for that are not encouraging.

What could derail FPT’s $5 billion target?

The target could slip if generative AI compresses the volume of routine outsourcing work faster than FPT can move up the value chain, if the Japanese economy stagnates, or if the company overextends into semiconductors and AI infrastructure without the client demand to fill the capacity.

The AI question cuts both ways. FPT’s business is built on large teams doing testing, maintenance and modernisation, work that coding assistants are demonstrably making more efficient. If clients need fewer engineer-hours per project, an hourly-billing model shrinks unless prices rise or scope expands. FPT’s response has been to sell AI-enabled services and to train its workforce on the tools, but every outsourcer says that, and the pricing evidence is not yet in.

The semiconductor ambition carries execution risk of a different kind. Chip design requires senior talent that Vietnam does not yet produce in volume, and the same engineers are courted by product companies such as VNG and by foreign captives and that Taiwanese, Korean and American firms pay far more to retain. FPT Semiconductor has announced products and partnerships, but the unit is small, and the group’s credibility in the field rests on future delivery rather than past results. Investors who saw the software unit grow from nothing may extend it the benefit of the doubt; competitors will not.

Finally there is succession. Trương Gia Bình is in his late sixties and remains chairman and the company’s dominant strategic voice. The group has a deep bench of long-serving executives and has rotated its chief executive several times without disruption, which is reassuring, but the willingness to lose money for years on a founder’s conviction is harder to institutionalise than a process. The next pivot, if there is one, will test whether the culture outlives the founder.

Frequently Asked Questions