Foxconn and Sharp's Decade of Friction Yields an AI Server Pivot

A decade of missteps, then a pivot that actually makes sense

It has been ten years since Foxconn absorbed Sharp into its orbit, and for most of that time the relationship looked less like a marriage and more like an uneasy merger of two companies that spoke different business languages. Now, with Sharp announcing in September its entry into AI server production in partnership with Foxconn, there is a tangible sign that something has shifted. The partnership finally found a lane where both sides have complementary strengths.

The details matter. Foxconn will supply AI servers to Sharp starting in fiscal 2027. Sharp will sell them to enterprises, local governments, and data center operators, then layer in maintenance and operations support. By fiscal 2030, the company is targeting ¥2,500 billion in revenue from the business. That is a specific number and a specific timeline — unusual for a relationship that, until now, has produced very few of either.

Why the first nine years produced so little

To understand why this pivot is notable, you need to understand what went wrong before. Sharp entered Foxconn’s umbrella in 2016 to survive a crisis triggered by over-investment in LCD panels. The deal was dramatic: Foxconn committed roughly ¥600 billion, offered stronger conditions than the competing Japanese government-backed fund, and won the day. At the signing, Foxconn chairman Terry Gou was visibly triumphant.

But the two companies were structurally different in ways that made collaboration difficult. Sharp is a heritage brand that plans products, designs them, develops technology, and sells them under its own name. Foxconn is the world’s largest electronics manufacturing services provider, with roughly 40 percent global market share. Its strength is executing at scale on someone else’s blueprint, not inventing the blueprint.

When Foxconn’s Dai Zhengwu became Sharp’s president, he brought a cost-cutting regime that reorganized the R&D division into a profit-seeking unit, lowered approval thresholds from ¥100 million to under ¥3 million, and signaled that no department was off-limits. Researchers felt pressure to prioritize near-term earnings over long-term exploration. Some left. The culture clash was real and documented by former employees who spoke to the source material.

The EV that never launched

The early years produced one high-profile collaboration that never materialized: an electric vehicle project. Foxconn has long talked about moving into EVs, and Sharp’s involvement was supposed to be part of that vision. Years after the acquisition, nothing had been shipped. The EV attempt illustrated the core problem — Sharp had the brand and the consumer electronics DNA, but Foxconn’s manufacturing advantages did not naturally translate into a vehicle platform.

The SDP mistake

Perhaps the most damaging misstep came from a decision that still puzzles observers. In 2022, Sharp took full ownership of Sharp Display Products (SDP), the subsidiary that manufactured the large LCD panels which had been the original source of its financial crisis. The reasoning was that display demand would recover and that the Sakai factory could hold an advantage for panels destined for the Americas. Panel prices instead fell further. Sharp booked a net loss of approximately ¥260 billion in its 2022 fiscal year.

The move was opaque even by corporate standards. In February 2021, Sharp’s board had approved the sale of its stake in SDP. A month later, the sale was canceled at the buyer’s request. Rumors circulated that Foxconn wanted to integrate SDP with its Chinese factories and take the combined entity public. That never happened. The 2023 shareholders meeting turned hostile, with investors demanding accountability.

The language barrier and leadership whiplash

Communication across the Taiwan-Japan axis added another layer of friction. When Dai stepped down and Wu Bixin took over as president, internal documents switched to bilingual Japanese-English, increasing workload and complicating coordination with Taiwan-based teams. Under the next Japanese president, Ootsu Masahiro, operations reverted to Japanese and Taiwan involvement receded. Liu Yangwei, who became Foxconn’s chairman, began visiting Japan frequently, but his message to Sharp president Kawamura Tetsuji was blunt: tell us what you need, because we cannot guess.

That exchange, reported in the source material, captures the essence of a decade of misaligned expectations. Foxconn had acquired Sharp but did not have a coherent plan for how to use it beyond cost discipline. Sharp had a brand and engineering talent but was struggling to define what it could offer a manufacturing giant.

What the AI server play gets right

The current pivot works because it assigns each company a role that matches its actual strength. Foxconn supplies the servers, leveraging its procurement network, manufacturing scale, and EMS relationships. Sharp sells, maintains, and operates them in the Japanese domestic market, where it already has relationships with enterprises and municipalities. Neither side is pretending to do the other’s job.

This is not a generic partnership announcement. It is a concrete business with a defined supply arrangement, a sales timeline starting in fiscal 2027, and a revenue target for fiscal 2030. The fact that it took ten years to arrive here is a indictment of the early strategy, not a reason to dismiss what comes next.

What this means beyond Japan

For Western readers, the story is worth watching for two reasons. First, it illustrates a realignment in the global AI infrastructure supply chain. Foxconn already dominates server manufacturing at scale. Pairing that capacity with a Japanese brand that has distribution channels and customer trust in one of the world’s largest economies creates a channel that neither side could easily replicate alone. Second, it raises the possibility of electronics reshoring dynamics that have not received proportional attention outside Japan. A partnership between a Taiwanese EMS leader and a Japanese legacy brand, focused on AI hardware for the domestic and regional market, is a template that could matter beyond this single deal.

The test ahead

The challenge for Sharp is whether it can now produce the differentiated technology or products that make the partnership sticky rather than transactional. A former employee noted the point plainly: if Sharp can create something unique, Foxconn’s scale becomes an asset instead of a constraint. Sharp’s founder Hikokatsu Hayakawa famously told his engineers to build products others could not copy. The company’s own management philosophy warns against chasing size alone — a principle it abandoned when it bet heavily on LCD capacity a decade ago.

The AI server deal is not the end of the story. It is the first chapter where the two companies seem to understand each other. Whether it leads to sustained growth or another cycle of misplaced bets will depend on what Sharp builds and how quickly Foxconn delivers.

The next few years will tell.