The second day of COMPUTEX 2024 turned into a vivid illustration of how personal relationships and high‑stakes technology demonstrations intertwine in the global AI race. While the exhibition halls buzzed with the latest GPUs, AI software, and futuristic concepts, the real drama unfolded at the Foxconn booth where chairman Liu Yang‑wei had prepared an elaborate welcome for NVIDIA founder Jensen Huang. The anticipation was palpable not only because of the celebrity status of Huang but also because his visit signals a deepening partnership between the world’s leading GPU designer and Taiwan’s premier contract manufacturer. This encounter is more than a photo‑op; it reflects the strategic alignment needed to accelerate AI infrastructure deployment worldwide.
Liu’s decision to arrive early, accompanied by a cadre of senior executives from Foxconn’s affiliates, underscores the lengths to which Taiwanese tech conglomerates will go to secure face‑time with key allies. Despite Huang’s arrival being delayed by roughly seventy minutes—a delay caused by a packed schedule of meetings and site visits—the waiting team remained steadfast, refusing to leave their posts. This discipline highlights a cultural trait in Taiwan’s tech sector: respect for partners and a willingness to invest time now for potential long‑term gains. The patience paid off, as the eventual meeting set the stage for a series of symbolic gifts that carried both goodwill and strategic messaging.
The first offering was Sharp’s AI companion robot, Poketomo, a compact device designed to interact with users through natural language processing and emotional recognition algorithms. While still pre‑market in Taiwan, Poketomo exemplifies the growing consumer‑grade AI segment that aims to provide companionship, mental‑wellness support, and home automation assistance. By presenting this robot, Foxconn signaled its readiness to expand beyond hardware manufacturing into integrated AI‑enabled consumer products, a move that could diversify its revenue streams amid fluctuating demand for traditional electronics.
Jensen Huang’s reaction to Poketomo was immediate and enthusiastic; he clipped the device to his clothing and jokingly asked for a second unit for his two children. This light‑hearted moment, captured by numerous cameras and smartphones, quickly became viral content across social platforms, inadvertently giving Poketomo a publicity boost that no paid campaign could match. The episode illustrates how high‑profile endorsements can accelerate market awareness for emerging AI gadgets, especially when the endorser is a trusted visionary in the AI community.
Following the AI pet, Foxconn presented a more substantial token: a brand‑new electric vehicle named Cavira, produced by its subsidiary Foxtron (formerly Foxtron Vehicle Technologies) and slated for launch in July 2024 with a price tag close to NT$1.39 million. The Cavira is built on Foxconn’s MIH (Mobility in Harmony) open‑platform architecture, which seeks to standardize EV development much like the Android ecosystem did for smartphones. Gifting Huang a Cavira serves dual purposes: it showcases Taiwan’s capability to produce premium EVs and offers a tangible testbed for Foxconn’s vehicle‑software integration.
From a strategic perspective, the EV gift aligns with Foxconn’s broader ambition to become a key player in the mobility sector, leveraging its manufacturing scale and supply‑chain expertise. As global automakers grapple with semiconductor shortages and the shift to software‑defined vehicles, Foxconn’s ability to provide end‑to‑end solutions—from chassis to cockpit AI—could make it an attractive partner. Huang’s potential use of the Cavira during his Taiwan stay also offers Foxconn real‑world feedback and visibility among influential tech leaders.
The cultural note of offering traditional Taiwanese red‑bean cakes added a layer of hospitality that resonated with both local media and international observers. While seemingly trivial, such gestures reinforce the narrative of Taiwan as a welcoming hub for global tech leaders, blending cutting‑edge innovation with deep‑rooted cultural traditions. In the context of international business, these subtle touches can improve rapport and pave the way for smoother negotiations on complex projects like AI server deployments or joint EV ventures.
The true substance of the meeting, however, lay in the discussion of AI server manufacturing efficiency. Huang highlighted how the assembly time for NVIDIA’s new Vera Rubin compute trays has plummeted from roughly two hours under the Grace Blackwell architecture to just five minutes—a 95 % reduction. He credited Foxconn’s engineering teams for achieving this leap through process redesign, automation, and tight collaboration on component standardization. This advancement directly translates into lower capital expenditures for data‑center operators and faster time‑to‑market for AI services.
Foxconn’s role in achieving such automation levels is a testament to its investment in robotic assembly lines, AI‑driven quality inspection, and lean manufacturing principles honed over decades of consumer‑electronics production. By applying these competencies to high‑mix, low‑volume AI server builds, Foxconn is helping to solve one of the industry’s bottlenecks: the ability to scale AI infrastructure rapidly without sacrificing reliability. For cloud providers and enterprises, this means more predictable capacity planning and reduced risk of deployment delays.
During the tour of the Vera Rubin NVL72 system display, Huang signed the exhibition wall and left the enthusiastic remark “100% Automation! Amaze, Amaze, Amaze,” underscoring his approval of Foxconn’s manufacturing prowess. His comments also revealed a broader belief: AI will permeate everyday life far beyond the confines of data centers. He envisaged AI companions like Poketomo assisting with elder care, AI‑enhanced diagnostic tools in clinics, and humanoid robots supporting logistics and manufacturing—areas where Taiwan’s strength in precision hardware could be decisive.
Looking at the wider market, the AI hardware race is intensifying as cloud giants, sovereign AI initiatives, and startups compete for scarce compute resources. Taiwan’s position as a linchpin in the semiconductor and electronics supply chain gives it unique leverage, but also exposes it to geopolitical tensions and cyclical demand swings. Companies that can diversify—adding AI consumer products, EVs, and advanced manufacturing services—are better positioned to weather volatility while capturing upside from multiple growth vectors.
For investors and industry leaders, the key takeaways are clear: monitor partnerships that combine cutting‑edge chip design with world‑class manufacturing, as these collaborations drive the most significant performance‑per‑dollar gains in AI infrastructure. Consider exposure to firms that are expanding beyond pure play foundry models into integrated solutions such as AI companions, EVs, and automation‑focused manufacturing. Policymakers should continue to support Taiwan’s innovation ecosystem through incentives for R&D, talent retention, and strategic industry‑academia linkages, ensuring the island remains an attractive hub for the next wave of AI‑enabled technologies.