# Hon Hai (Foxconn) (2317) Q2 2026 earnings

Results filed 2026-08-12 for the quarter ended 2026-06-30. Source page: https://earnings.opennash.com/companies/hon-hai/2026-q2/
Press release and documents: https://image.honhai.com/upload/202608/law_talk/Hon_Hai_2Q26_Results_English_20260812_3711.pdf
> Summaries and transcripts are machine-generated and can contain errors.

## In brief

Foxconn's cloud and networking business passed half of revenue for the first time, as next-generation AI racks near shipment and custom ASIC server projects widen its customer base.

AI infrastructure is now Foxconn's largest business: Cloud and Networking Products exceeded 50% of revenue for the first time, helped by AI rack ramps and strong switch demand. Customers are widening beyond North American cloud providers to Neocloud, sovereign AI and large enterprises. ASIC server projects, co-packaged optics (CPO) switches and general servers added to demand. Consumer electronics enters its seasonal peak in the third quarter, while computing products face tight memory supply.
Management wants to move from contract manufacturer to AI platform supplier. That means selling full systems, modular data centers and software platforms, alongside EVs, robotics and semiconductors. Next-generation Vera Rubin racks begin shipping in the fourth quarter and become the main product next year, with a 50% share target. The main risk is execution and supply: the ramp depends on chip packaging capacity, and capital spending is set to rise more than 30% while working capital is already heavily used.

## Key numbers

| Metric | Value | Change | Basis |
| --- | --- | --- | --- |
| Revenue | NT$2.53 trillion | up 41% from a year ago | GAAP |
| Net profit after tax (attributable to parent) | NT$60.0 billion | up 35% from a year ago | GAAP |
| Diluted EPS | 4.24 | up from 3.16 a year ago | GAAP |
| Gross margin | 6.12% | down 0.21 ppts from a year ago | GAAP |
| Operating profit | NT$94.8 billion | up 68% from a year ago | GAAP |
| Operating margin | 3.75% | up 0.6 ppts from a year ago | GAAP |
| Operating cash flow (first half, net outflow) | NT$69.1 billion | net outflow versus an inflow in the first half of 2025 | GAAP |
| Free cash flow (first half, net outflow) | NT$150.0 billion |  | non-GAAP |

## Segments and operating metrics

| Name | Value | Change |
| --- | --- | --- |
| Cloud and Networking Products (share of Q2 revenue) | 51% |  |
| Smart Consumer Electronics (share of Q2 revenue) | 29 % |  |
| Computing Products (share of Q2 revenue) | 15 % |  |
| Components and Other Products (share of Q2 revenue) | 5% |  |

## Outlook

| Metric | Period | Outlook |
| --- | --- | --- |
| Capital expenditure growth | 2026 | more than 30% YoY |
| Full-year operating margin | 2026 | exceed last year’s level of 3.2% |

## Key takeaways

1. **Vera Rubin racks ship in Q4, lead 2027** Mass production preparation for next-generation AI racks starts in the third quarter, with shipments in the fourth quarter and the platform becoming the main product next year. Foxconn says it is still moving toward a 50% share, but market volume depends on chip packaging (CoWoS) capacity allocation.
2. **ASIC servers become a second growth leg** New ASIC customers and projects mean Foxconn expects a significantly higher ASIC share this year, targeting above 40%. Most ASIC projects use consignment, so reported revenue understates their scale.
3. **Customer base widens beyond North American cloud providers** Management says Neocloud, sovereign AI and large enterprise customers are now part of the mix, and orders for next year are trending up. New cloud customers should begin contributing before the end of this year.
4. **CPO optical switches begin shipping into AI racks** Co-packaged optics switches have begun shipping, mainly for scale-out links between racks, and 800G-and-above switch revenue is still expected to double for the full year. Foxconn sees optical networking as its next AI growth trend and a profit-mix improver.
5. **EVs, robotics and AI platforms move toward deliveries** Cavira EV deliveries began in July, a Mitsubishi Motors vehicle starts delivering in the second half, and humanoid robots entered US factories. Genesis (manufacturing) and CityGPT (smart city) passed field validation as Foxconn pushes toward being an AI platform supplier.

## Analyst Q&A

- [0:00:00] Anne Lee, BofA Securities: What products does Foxconn supply for TPU projects, and is general server and CPU demand outlook better than three months ago and for next year? Answer: Foxconn declined to name customers but said ASIC share will rise significantly, and general servers should grow quarter by quarter with better visibility, with industry CPU shipments seen up about 16% in 2026 and 23% next year.
- [0:00:00] Wayne Liu, Neuberger Berman: What are the scale and outlook for networking products including CPO, and has order visibility and Neocloud demand improved? Answer: Order visibility is improving with next year's orders trending up, new cloud customers should contribute before year-end, and CPO switches have begun shipping as optical networking becomes a growth trend.
- [0:00:00] Kaylin Tsai, Macquarie Securities: Could GPU servers also move to consignment, and how large is ASIC in AI server revenue and share? Answer: Consignment may extend to GPU projects case by case, ASIC was slightly above 10% of AI server revenue in the first half, and Foxconn targets above 40% ASIC share.
- [0:00:00] Pei-Ying Hsieh, TVBS News: When will Vera Rubin ramp, at what scale, and can Foxconn reach close to 50% share? Answer: Shipments start in the fourth quarter and it becomes the main product next year, with Foxconn still moving toward a 50% share, subject to chip packaging capacity.
- [0:00:00] Angela Hsiang, BofA Securities: What is the schedule and margin impact of standalone LPU and Vera CPU racks, and how do the AI platform efforts add competitive value? Answer: LPU and Vera CPU racks enter production in the third quarter with larger shipments next year but limited revenue impact given lower prices, while the smart platforms raise internal efficiency and can be sold externally.
- [0:00:00] Melody Chang, EBC Financial News: What is the cooperation and positioning with SHARP for the next ten years? Answer: Foxconn remains firmly supportive and sees joint opportunities in AI, robotics, EVs and smart living, citing SHARP's first-generation AI server developed with Foxconn.

## Questions

**What is driving Foxconn's AI server growth in the second half of 2026?** Next-generation AI racks begin mass production in the third quarter, with AI rack shipments expected to grow high double digits from the second quarter. Demand for GB-series products is expected to continue at least into next year.
**How is Foxconn positioning in custom ASIC AI servers?** Management cited new ASIC customers and projects and targets an ASIC market share above 40%. ASIC was slightly above 10% of AI server revenue in the first half, and most ASIC projects use consignment, which understates reported revenue.
**Who is buying Foxconn's AI infrastructure besides the big US cloud providers?** Foxconn says demand now includes Neocloud providers, AI model developers, sovereign AI projects and large enterprises. It expects new cloud customers to begin contributing before the end of this year.
**What is Foxconn doing in networking and optical products?** CPO optical switches have begun shipping, mainly for scale-out between AI racks, and revenue from 800G-and-above switches is expected to double for the full year. Foxconn is also investing in optical modules, advanced packaging, cables and connectors.
**What are the main risks to Foxconn's AI plans?** Management points to chip supply, since next year's rack volume depends on CoWoS capacity allocation, and to heavy working capital and capital spending growth of more than 30% this year. Product mix shifts toward consumer electronics in the second half also affect margins.

## Slides and investor documents

- [Hon Hai 2Q26 Results presentation (English)](https://image.honhai.com/upload/202608/law_talk/Hon_Hai_2Q26_Results_English_20260812_3711.pdf) (Slide deck)

## Transcript

The company's own transcript: https://image.honhai.com/upload/202608/law_talk/Hon_Hai_2Q26_Results_Transcript_English_20260818_1815.pdf


**Speaker** [0:00:00]
TRANSCRIPT 2317.TW – Q2 2026 Hon Hai Precision Industry Co., Ltd Investor Conference Call EVENT DATE/TIME: August 12, 2026 Taipei time 3:00 P.M. (New York: August 12, 3:00 A.M. / London: August 12, 8:00 A.M.)
Investor Conference Call on FY26 Second Quarter Financial Results
Corporate Participants
Michael Chiang Hon Hai Technology Group – Rotating CEO David Huang Hon Hai Technology Group – CFO James Wu Hon Hai Technology Group – Spokesperson Kristen Fang Hon Hai Technology Group – Senior IR Manager
Conference Call Participants
Anne Lee Nomura Securities – Analyst Wayne Liu Neuberger Berman – Analyst Kaylin Tsai Macquarie Securities - Analyst Pei-Ying Hsieh TVBS News - Reporter
Angela Hsiang KGI Securities – Analyst Melody Chang EBC Financial News - Reporter
Presentation

**Kristen Fang (Investor Relations)** [0:00:00]
Hello to all the investors and media. This is Kristen. Welcome to Hon Hai’s (“Foxconn”) Second Quarter 2026 Investor Conference Call. Joining us today are Rotating CEO Michael Chiang, CFO David Huang, and Spokesperson James Wu. The conference call is scheduled for one hour, starting with our presentation, followed by Q&A session.
As usual, please carefully read the safe harbor notice on the next page before we start the meeting.
We will now proceed to the first session, the presentation. This time, we will cover four topics, including the performance review of the second quarter of 2026, the business outlook for the third quarter and full year, major business developments, and recent major events.
Before moving into the financial figures, we would first like to share the key highlights of this quarter.
First, operating profit for both the second quarter and the first half increased by more than 60% YoY, significantly higher than revenue growth. This shows that Foxconn’s core profitability is improving rapidly.
Second, looking ahead to the third quarter, we expect significant QoQ growth and strong YoY growth. For the full year, we maintain our view of strong growth.
Third, what we provide is not a single product, but a complete AI system ranging from components, modules, and server racks to modular datacenters and supercomputing centers.
Fourth, our customers’ AI capital expenditure continues to increase. We will actively expand global capacity to meet customer demand and expect sustained growth next year.
I will now invite CFO David to walk through the financial performance for the second quarter.

**David Huang (CFO)** [0:00:00]
Thank you, Kristen. Hello everyone. I am David Huang, CFO of Hon Hai Technology Group. I am going to start with some highlights on Foxconn's financial results for the second quarter of 2026.
Firstly, please refer to page 6 of the presentation for the income statement for the second quarter of 2026.
The revenue for 2Q26 was NT$2.53 trillion, representing 41% YoY growth, and setting a new record high for the same period in the company’s history. Compared with the first quarter, revenue also grew 19%, reflecting the continued growth of our major businesses.
In terms of profitability, gross margin for the second quarter was 6.12%, down 0.21 ppts YoY, mainly due to changes in product mix. However, gross profit still increased by 36% YoY.
Operating margin was 3.75%, up 0.6 ppts YoY, mainly benefiting from higher gross profit driven by revenue growth, as well as a lower opex ratio.
Operating profit reached NT$94.8 billion, up 68% YoY. The growth rate was significantly higher than revenue growth, mainly because gross profit increased by 36%, while expenses increased by only 5%. As a result, operating profit growth was very strong and reached a new record high for the same period.
Next, looking at non-operating items, this quarter recorded a gain of NT$63 million, down approximately NT$8.5 billion from the same period last year, mainly due to fund investments and net interest expense.
Fund investment gains decreased by NT$5.9 billion compared with the same period last year, mainly because some of our investee companies went public last year, when their share prices increased significantly.
Net interest expense increased by NT$2.7 billion, mainly because the expansion of the AI business and overall operating scale drove higher funding requirements. Despite this, operating profit this quarter increased significantly compared with the same period last year.
The effective tax rate was 26.2%, up 2.2 ppts YoY, mainly due to profit growth at subsidiaries, which led to an estimated NT$2.6 billion increase in income tax provisions for
earnings repatriation.
Second-quarter net profit after tax was NT$60.0 billion, up 35% YoY and up 20% QoQ, setting a new record high for the same period.
EPS was NT$4.27, an increase of NT$1.08 from the same period last year, representing 34% growth, and up 20% QoQ. This also set a record high for the same period.
Overall, second-quarter revenue continued to grow, and profit growth this quarter was even faster than revenue growth. Core operating performance remained strong.
Next, please turn to page 7 for the income statement for the first half of the year.
Revenue for the first half of 2026 was NT$4.65 trillion, up 35% YoY. Operating profit was NT$170.5 billion, up 65% YoY. Net profit after tax was NT$109.9 billion, up 27% YoY. EPS was NT$7.84, up 26% YoY.
Among these, operating profit, net profit, and EPS all reached new record highs for the same period.
Looking at the overall performance for the first half, operating profit growth was significantly higher than revenue growth. This shows that as operating scale expanded, the company’s profitability also improved.
For the first half of the year, gross margin was 6.15%, down 0.08 ppts YoY, but gross profit still increased by 33%. At the same time, operating margin increased to 3.67%, up 0.67 ppts YoY. In other words, although changes in product mix caused gross margin to decline slightly, operating margin continued to improve through revenue scale growth and improved expense efficiency.
In terms of capital efficiency, ROE for the first half of the year was 6.21%, up 0.73 ppts YoY. This reflects that while the company’s operating scale is growing rapidly, core profitability and capital efficiency are also continuing to improve.
Overall, operating scale, operating profit, and earnings per share all maintained growth in the first half. Core business profit growth was higher than revenue growth, and overall operating performance remained solid.
Next, please turn to page 8 of the presentation for the balance sheet.
As of the end of June 2026, cash and time deposits on the books totaled NT$1.52 trillion. Net cash was NT$219.8 billion, down NT$23.8 billion from the same period last year, mainly reflecting the continued expansion of operating scale, which drove higher working capital requirements.
In terms of working capital management, the cash conversion cycle for the second quarter was 42 days, an improvement of 6 days from the same period last year.
Excluding the impact of accounts receivable management tools, the cash conversion cycle was 52 days, still an improvement of 2 days compared with the same period last year. This shows that, amid rapid growth in operating scale, the company’s working capital management efficiency continued to improve.
The debt ratio for the period was 62%, remaining broadly stable. The financial structure remained solid.
Finally, please turn to page 9 of the presentation for the cash flow statement.
In the first half of 2026, as operating scale grew rapidly, working capital requirements, such as accounts receivable and inventory, also increased. As a result, operating cash flow was a net outflow of NT$69.1 billion.
Free cash flow for the first half was a net outflow of NT$150.0 billion, mainly due to increased working capital investment and the impact of capital expenditure.
In terms of capital expenditure, CAPEX was NT$80.9 billion in the first half, up 5% YoY, in line with expectations.
Overall, business grew rapidly in the first half of this year, driving higher working capital and capital expenditure requirements. Although cash flow saw a phased outflow, working capital efficiency continued to improve, profitability and operating scale continued to grow, and the financial structure remained solid.
This concludes the brief review of the financial statements for the second quarter of 2026. Next, I will hand the floor over to Rotating CEO Michael Chiang.

**Michael Chiang (CEO)** [0:00:00]
Thank you, David. Hello to all investors. I am Michael Chiang, Rotating CEO. I am very pleased to meet everyone again today and to share Foxconn’s second-quarter performance and future outlook.
First, let us review our second-quarter operations.
Our second-quarter revenue reached a new record high for the same period, and performance also exceeded the outlook we provided at the investor conference in May. Revenue achieved strong growth both QoQ and YoY. The performance across four major product segments was also better than expected.
Among them, Cloud and Networking Products achieved strong QoQ growth. Since the second quarter is traditionally the off-season for ICT products, the revenue contribution from Cloud and Networking Products increased further, exceeding 50% for the first time.
From an overall profitability perspective, second-quarter operating profit and net profit after tax both reached new record highs for the same period.
In particular, second-quarter operating profit reached NT$94.8 billion. Not only were the QoQ and YoY growth rates very high, but this also marked the highest single-quarter operating profit in the company’s history. This represents a significant improvement in our core business profitability.
For me, growth is not only about expanding our scale. More importantly, it is about whether that growth can be converted into better profitability. This is also the direction I have continued to promote since taking over as Rotating CEO: to grow our scale, while also improving profitability.
Next, I would like to explain our outlook for the third quarter. In the third quarter, AI operations will continue to grow. In addition, ICT products will enter the peak season in the second half of the year. Therefore, we expect significant QoQ growth and strong YoY growth.
For Cloud and Networking Products, AI demand continues to grow at a rapid pace. CSP customers are increasing capital expenditure, and our AI racks continue to ramp up. We expect this segment to deliver strong growth both QoQ and YoY.
For Smart Consumer Electronics, as the segment has entered its traditional peak season, we expect significant growth in the third quarter compared with both the previous quarter and the same period last year.
For Computing Products, as the peak demand period for new products’ launch has been behind us, and due to tight memory supply, we expect a slight decline in the third quarter compared with the second quarter. However, as price increases for some components have driven shipment prices higher, we expect significant growth compared with the same period last year.
As for Components and Other Products, due to shipment growth in components related to major businesses, including camera modules and connectors, we expect significant YoY growth.
In terms of the full-year outlook, we maintain our view of strong growth, but visibility is now even better than it was in May. The reason remains very strong demand for AI servers.
Here, I would also like to share with everyone Foxconn’s qualitative transformation in the AI era. Since 2024, AI server revenue has continued to increase, gradually becoming the Group’s most important growth engine, with revenue maintaining double-digit YoY growth every quarter. Moreover, this momentum has accelerated significantly over the past year.
In addition, AI servers are infrastructure-related products, so seasonal volatility is relatively smaller. As a result, revenue performance throughout the year is more stable, offsetting the seasonal fluctuations of the consumer electronics industry. The previous pattern of “weakness in May and June” is no longer evident.
The July revenue we just announced exceeded NT$900 billion for the first time in a single month, setting a new record high for monthly revenue before the peak season has even begun. This fully demonstrates the explosive power of AI demand, as well as our core position in the AI supply chain.
During the previous investor conference, I shared several financial highlights with you. Over the past quarter, we have seen these indicators continue to move in a positive direction.
Looking at the first half, operating profit increased by 65%, outperforming revenue and gross profit growth. This reflects that the scaling of AI and the improvement in automation efficiency have significantly strengthened the profitability of our core business. Due to our
business model, we may not be able to control the direction of gross margin rate, but we will definitely continue to pursue the operating objective of maximizing earnings.
The Group’s ROE in the first half has already reached 6.21%, representing a clear improvement from the same period last year and moving close to our full-year target of 12%. Going forward, as the AI business continues to grow and the benefits of vertical integration gradually materialize, we believe the Group’s overall profitability and capital efficiency will continue to improve. We believe ROE still has further room to move upward.
As for operating margin, it reached 3.67% in the first half, an increase of 0.67 ppts compared with the same period last year. We are confident that the full-year operating margin this year will exceed last year’s level of 3.2%.
Next, I would like to share the cash flow situation.
Adjusted EBITDA in the first half increased by 52% YoY, up NT$77.6 billion from the first half of last year. This was far higher than the combined increase in capital expenditure and interest expense, demonstrating that Foxconn’s cash flow is becoming increasingly healthy.
The outlook for next year is also very positive. We are seeing extremely strong customer demand for AI. We expect Foxconn’s capital expenditure to continue increasing, mainly to expand key capacity ahead of demand, including servers, racks, liquid cooling, and testing, as well as to strengthen global manufacturing and automation deployment.
This includes capacity expansion in locations such as Taiwan, the United States, Mexico, and Vietnam. In response to localized demand from U.S. customers, we will expand our R&D and manufacturing capabilities in Texas, Wisconsin, Ohio, and California.
We can be optimistic that these capital expenditures will generate very significant benefits for revenue, profitability, and cash flow in the future.
Next, I will continue to explain our major business developments.
These strong financial results and robust growth outlook did not come from nowhere. They are supported by a solid foundation. Everyone may be curious about what the key to Foxconn’s continued operating momentum is, given the rapidly changing external environment. I would like to use this pyramid chart to explain the Group’s core operating model in greater depth.
At the bottom layer is the global deployment and scale benefits that we have built over the long term. Five to six years ago, Foxconn already believed that regional manufacturing would become a trend, so we established localized production capacity on a large scale. The number of our global sites has nearly doubled. In recent years, changes in the geopolitical environment have led countries to place greater emphasis on localized supply chains. This has exactly matched Foxconn’s earlier expectations. Today, Foxconn’s presence across more than 240 sites in 24 countries allows us to respond proactively, allocate capacity more flexibly, leverage the value of Local for Local, and respond to customer needs.
The middle layer is the vertical integration and high degree of automation that we have cultivated over the long term. Foxconn’s advantage lies in our high in-house production ratio for key components. This allows us to maintain good operating efficiency and yield even when facing high-value and highly complex AI products. More importantly, we can meet customers’ demand for one-stop services while also demonstrating our profitability.
In addition, the automation level of our core manufacturing processes is very high. Taking our largest-scale processes, such as CNC and SMT as examples, our automation ratio is already close to 100%. These are extremely important for yield and output in large-scale mass production. This is also one of our leading advantages.
Based on these capabilities, the top layer is R&D innovation and a high-end segment focus. In recent years, we have continued to increase R&D investment in new businesses. This year, R&D spending in 3+3+3 and the Hon Hai Research Institute is close to NT$20 billion, up 40% YoY.
With the support of scale, vertical integration, and manufacturing capabilities, we are able to jointly develop with tier-1 customers, focus on forward-looking and proprietary technologies, increase the share of high-value-added products, and drive Foxconn’s transformation toward becoming a technology manufacturing platform service provider.
These three levels, from scale as the foundation, through efficiency and integration, to R&D-driven high-end value, form the key operating model that enables Foxconn to capture the enormous global demand for AI at this moment and continue leading through the transition between technology generations.
With the model described above, together with our scale and long-term customer
partnerships, we are able to secure our position as a primary supplier to tier-1 AI customers.
Moreover, what customers value most is not simply server hardware, but one-stop system delivery capability. This is exactly Foxconn’s core advantage. Facing customers’ requirements for datacenter construction timelines, we can not only provide AI servers and racks, but also extend the value chain to modular datacenters, supercomputing centers, and the overall infrastructure systems required to build sovereign AI.
In the past, we focused on major North American CSPs, but we have now gradually expanded into Neocloud, sovereign AI, and large enterprise customers. Our market share is expanding.
The reason Foxconn can become customers’ first choice lies, I believe, in our ability to build partnerships from zero, our design capability from zero to one, and our global mass- production capability from one to one hundred. We can provide customers with industry- leading Time to Market, Time to Cost, and Time to Technology, helping them deploy AI computing power at the fastest possible speed.
In recent years, major North American CSPs have successively raised their capital expenditure budgets. This year, the total is expected to approach US$800 billion, and it is expected to soon cross the US$1 trillion threshold. This unprecedented wave of global cloud investment is Foxconn’s most direct growth driver over the next several years.
As the world’s largest AI server provider, whether for GPU or ASIC chip solutions, we will certainly actively expand capacity and help customers implement these enormous capital investments. At present, AI server demand remains extremely strong.
Looking ahead to the third quarter, with the launch of new platforms, we expect AI rack shipments to show high double-digit growth compared with the second quarter, and to drive continued rapid growth in AI server revenue.
Moreover, looking ahead to the full year of 2026, we maintain our expectation that AI rack shipments will grow by multiple times. Also, because we are a major supplier to customers for GPU products, we are gradually becoming a major partner in these customers’ ASIC projects as well. Therefore, we will also see substantial growth in ASIC this year.
In addition, in high-performance networking products, we maintain the target we proposed
last time. Revenue from 800G and above switches is still expected to double for the full year, and CPO optical switches will also begin mass production shipments in the third quarter according to plan. The significance of AI to Foxconn is not only about server racks. It is also the core driver of our transformation and upgrade.
Together with the three future industries and three core technologies that we have continued to invest, this will form a platform-based ecosystem with strong technological barriers and high software-hardware integration. Our long-term goal is very clear: to further transform from the world’s most important technology manufacturer into an AI platform supplier.
Next, I will explain the latest developments in 3+3+3.
First, in EVs, our CDMS model continues to deliver results, and our overseas market expansion is also progressing steadily. The vehicle model developed in cooperation with Mitsubishi Motors of Japan will begin deliveries in the second half of the year. Our cooperation with Poland EMP, targeting EV projects in Europe, has also begun moving forward. In addition, the EV charging equipment plant jointly established with Saudi Arabia is also expected to begin production in the fourth quarter. In the domestic market, the new MODEL C variant, Cavira, began deliveries in July. Our new Kaohsiung plant has also begun producing electric buses and electric bus batteries.
In Digital Health, we are focusing on three directions: smart healthcare scenarios, platform integration, and global commercialization, in order to accelerate the implementation of solutions.
In robotics, in the second quarter, we introduced humanoid and collaborative robots into our factories in the United States. This is a very important milestone. Going forward, through the integration of Physical AI, AI vision, force-control technology, and digital twin technologies, we will accelerate the upgrade of smart assembly, promote manufacturing from automation toward autonomy, and at the same time expand business opportunities in high-value-added components and embodied intelligence platforms.
In terms of progress across the three core technologies, first, in AI technology, our deployment is extending from infrastructure all the way to platforms and models. In AIDC infrastructure, we are integrating computing power, datacenters, and energy management to promote the standardization and scaling of AIDC. In sovereign AI, we are also continuing to promote implementation across Europe, Africa, and Asia-Pacific markets. Our proprietary
model, FoxBrain, continues to improve its model architecture and training efficiency, while also deepening localized knowledge to enhance the capabilities of our proprietary model.
Second, in semiconductors, we are advancing our capabilities across capacity, key chips, and design services at the same time. Order visibility for the 8-inch wafer fab in Japan already extends into the fourth quarter.
Proprietary technologies including silicon carbide, automotive MCU, and AI power chips are also progressing according to plan.
In next-generation communications, we continue to deploy low-earth-orbit satellites and space communications technologies, while providing complete manufacturing services from ground equipment to satellite bodies.
AI computing equipment is currently our hardware core. The three smart platforms, including Smart Manufacturing, Smart EV, and Smart City, are important areas of deployments for us to promote AI software-hardware integration and platform value creation.
In Smart Manufacturing, the Genesis platform has completed field validation and is gradually being introduced into our major sites. Through rapid deployment of AI agents and cross-site replication, we will continue to improve production efficiency and smart manufacturing capabilities.
In Smart EV, we continue to advance electronic and electrical architecture, smart cockpit, and vehicle-cloud data platforms to help customers accelerate development and mass production.
In Smart City, CityGPT has completed benchmark field validation, and we have begun promoting deployment across cities and overseas markets.
Through software-hardware integration across the three smart platforms, we will continue to move the Group toward the goal of becoming a globally leading AI platform provider.
Finally, I would like to summarize the key points for everyone. The three major industries of ICT, AI, and EV are driving the Group’s transformation and upgrade. Foxconn has already entered a new growth trajectory and has become a growth company with significantly improving profitability. We will also strictly maintain financial discipline and enhance
shareholder equity. Foxconn’s three-layer operating model provides systematic solutions, not single hardware products.
Together with our five core values of speed, quality, engineering services, flexibility, and cost, we are definitely the preferred partner for AI customers in GPU, ASIC, and various related hardware.
As customer capital expenditures increase, Foxconn’s market share is also rising. Therefore, we will continue to expand capital expenditure, expand global capacity, and extend the value chain. We will continue to capture every customer and every business opportunity.
Next, I will invite James to walk everyone through our recent major events.

**James Wu (Executive)** [0:00:00]
Thank you, Michael. Next, I will walk everyone through recent major events.
Firstly, the shareholders’ meeting approved a NT$100.9 billion cash dividend proposal, with a cash dividend of NT$7.2 per share. This also marks the seventh consecutive year in which our cash dividend payout ratio has exceeded 50%.
In terms of strategic cooperations, we are working with Intel to jointly promote next- generation AI platforms. We are also working with Schneider to promote next-generation AI datacenter and green energy solutions. In addition, our packaging and testing plant with Thales and Radiall has also officially broken ground in France.
In terms of exhibitions, at this year’s COMPUTEX, we showcased the highly anticipated next-generation platform, the Vera Rubin rack, CPO high-performance networking technology, and generative AI applications across the three smart platforms.
In addition, we participated in VivaTech in France for the first time, showcasing our deployment in AI and EV to Europe.
At the Taiwan Expo in Japan, we showcased two electric vehicles, MODEL A and MODEL B, as well as AI modular datacenters, attracting many Japanese partners.
Recently, the Hon Hai Research Institute also worked with National Yang Ming Chiao Tung University to develop ultra-large-capacity silicon photonics technology, laying a key technological milestone for transmission in supercomputing and communications.
In terms of R&D capabilities, in the first half of this year, we published a total of 48 conference papers and 48 journal papers, and obtained 16 patents. These demonstrate that Foxconn has extremely solid proprietary R&D capabilities in key frontier technologies.
Recently, we have also received many international recognitions. In the Extel rankings, we received the Most Honored Company award, along with seven other major awards. In the Fortune Global 500, we rose by five places to No. 23.
In addition to the company’s overall performance, our colleagues also delivered outstanding achievements. Our Chief Campus Operation Officer, Kathy Yang, was selected as one of the world’s Most Powerful Women in Business. Our Chief Human Resources Officer, Emily Hsia, is also recognized as HR Asia Annual Best HR Officer. Our US CHRO, Winnie Tu, also received recognition from the Manufacturing Institute STEP Awards.
In the field of R&D, our director Chang Ching-Ray received the AUMS Award, and Dr. Hung Yu-Heng from the Semiconductor Research Institute was selected as an Outstanding Young Engineer.
Next, in terms of ESG, we expanded global third-party audits. We have cumulatively completed 85 RBA-VAP audits and obtained six platinum-level certifications, the highest ranking of the program. In green energy, we established a renewable energy partnership with Brookfield in Vietnam. We have also published our Supplier Responsibility Report for three consecutive years, working together with the supply chain to implement our sustainability commitments.
In talent development, this year’s Hon Hai Technology Award received a very strong response. We attracted more than 200 outstanding academic works from Taiwan and overseas. In Vietnam, we also held a competition to cultivate local host talent.
In terms of employee relations, in Slovakia, we held Volunteer Week and a summer family day. In the Czech Republic, we also held Technology Day. In Taiwan, we also held a family day to strengthen cohesion among employees and their families.
Finally, I am very pleased to share with everyone that the 2026 Hon Hai Tech Day will take place on October 30 and 31 at TaiNEX 1, Nangang Exhibition Center. This time, we are once again expanding the scale, covering large exhibition areas on both the first and fourth floors. Friday, October 30, will be an invitation-only guest day. Saturday, October 31, will officially open to the general public for free admission.
In the fourth-floor exhibition area, we will fully present the latest technological achievements from deep integration of the three smart platforms and AI, and will also launch our latest products. In the first-floor exhibition area, we will create the Family Pavilion for the first time, focusing on technology education. It will include five themed exhibition areas, including sustainability, future mobility, and Smart City, and will provide more than 10 rich career experience activities, allowing technology and education to take root from an early age.
This is not only Foxconn’s annual event for software-hardware integration, moving from manufacturing toward an AI platform, but also a technology carnival that parents and children can enjoy together. On behalf of Foxconn , I sincerely invite all of you to participate in this major event.
That concludes today’s presentation. Thank you.
Questions and Answers

**Kristen Fang (Investor Relations)** [0:00:00]
Thank you, James. Next, we will move to the Q&A session. We’ll go through questions that were raised in advance of today’s call and answer those first. After that, we will open the floor to any questions.
Question 1: The first question is for Rotating CEO Michael Chiang, regarding the outlook for AI servers. With the launch of the new-generation AI rack products, how does the company view the impact of product transition on AI rack shipments in the second half of the year? What is your view on the overall competitive landscape? Also, under the trend toward product standardization, will this allow more competitors to enter the market? Thank you.

**Michael Chiang (CEO)** [0:00:00]
Thank you, Kristen. We maintain our plan for the new-generation AI racks to begin mass production in the third quarter. As validation is gradually completed, and as supply chain and capacity preparations come into place, shipment volume in the following quarters will gradually increase. Overall demand remains very strong.
Currently, demand for the GB platform remains stable. We expect demand for GB-series products to continue at least into next year. Our expectation that overall AI rack shipments will grow quarter by quarter remains unchanged.
As we enter the new-generation AI platform, we believe large customers will concentrate toward large suppliers with scaled delivery capabilities. Product standardization does not mean that entry barriers are lower. Instead, it will further highlight the advantages of suppliers that possess scaled delivery capability and complete vertical integration capability.
Scaled delivery capability depends on the degree of automation, and this is one of Foxconn’s important advantages. This includes our experience in building automated equipment in the ICT industry. Starting from the new generation, we have achieved 100% automation in Compute Tray.
What I want to emphasize is that in the AI hardware industry, what Foxconn provides is the
entire system, not a single product. Compared with most peers, which only provide part of the AI server value chain, such as Level 6 motherboards, Level 10 assembly, or Level 11 system assembly, we provide a complete one-stop service: from the very beginning, starting with Level 1 components, to GPU modules, motherboards, Switch Tray, Compute Tray, rack assembly, and even connections between racks.
Inside the rack, we also integrate our self-manufactured components, including high-value- added areas such as liquid cooling, power management, connectors, busbars, mechanical components, and high-speed transmission. The in-house production ratio of components has reached more than 50%.
Moreover, AI racks are highly complex. They involve high-power design, cooling systems, system integration, testing and validation, as well as customer customization requirements. All of these require long-accumulated engineering capabilities and supply chain management experience. The entry barriers will only become higher and higher.
Through our design capability from zero to one, and our global mass production capability from one to one hundred, we have become customers’ preferred partner. We expect our market share to further increase in the new-generation platforms. Thank you.

**Kristen Fang (Investor Relations)** [0:00:00]
Question 2: The second question is related to funding planning, and is for the CFO. Over the past few years, the company has continued to expand capital expenditure. As the AI business grows further, it appears that the investment scale will increase further this year. Could the CFO share how you view the impact of these capital expenditures on profitability and cash flow returns? Also, how will the company fund future increases in capital expenditure and working capital requirements? Will there be any need for equity fundraising?

**David Huang (CFO)** [0:00:00]
Thank you, Kristen. I would like to first explain this from the perspective of capital expenditure and investment returns.
Over the past few years, as the AI business has grown, global manufacturing deployment has expanded, and automation demand has increased, we have continued to increase capital expenditure. In 2024, capital expenditure was approximately NT$136.3 billion, an increase of approximately NT$21.6 billion YoY. In 2025, it further increased to approximately NT$173.8 billion, an increase of NT$37.4 billion YoY.
Based on current operating results, these previous investments have gradually been reflected in revenue scale and profitability. In 2025, revenue reached NT$8.1 trillion, an increase of NT$1.24 trillion YoY. Operating profit reached NT$259.2 billion, an increase of NT$58.6 billion YoY.
Capital expenditure in the first half of this year was approximately NT$80.9 billion, an increase of NT$3.7 billion YoY. During the same period, operating profit reached NT$170.5 billion, an increase of NT$67.4 billion YoY.
In other words, operating profit grew faster than revenue, showing that previous capital investments are gradually being converted into economies of scale and profitability.
From the perspective of cash generation, EBITDA last year was approximately NT$357.1 billion, up NT$68.0 billion YoY. In the first half of this year, EBITDA reached NT$227.7 billion, up NT$77.6 billion YoY. This shows that the cash generation capability of our core operations continues to improve.
Of course, in the first half of this year, the rapid growth of the AI business and overall operating scale also drove an increase in working capital requirements, including accounts receivable and inventory. Therefore, free cash flow had a phased outflow.
However, from the perspective of overall profitability and cash generation capability, previous investments have gradually begun to contribute, and also support the company’s continued investment expansion and execution of its cash dividend policy.
Looking at the full year, in response to demand for AI server racks and regional manufacturing, we expect full-year capital expenditure to increase by more than 30% YoY. This will mainly be used to expand key capacity ahead of demand, including servers, racks, liquid cooling, and testing, while also strengthening our global manufacturing automation deployment.
As Rotating CEO Michael explained earlier, these capital expenditures are mainly based on customer demand and business growth. Therefore, we believe the related investments will continue to generate contributions in the future.
Going forward, the company’s funding requirements can mainly be divided into two parts. The first is capital expenditure, and the second is working capital requirements that increase along with business growth.
In terms of funding arrangements, we will maintain both stability and flexibility. The basic principle is to match short-term funding with short-term needs, and long-term funding with long-term needs.
In terms of selecting financial tools, we will comprehensively consider factors including cost, which is very important, as well as maturity and external market conditions, in order to maintain the company’s financial flexibility.
Short-term working capital will mainly be supported through working capital management and bank credit lines. Medium- to long-term capital expenditure requirements will be funded through different capital market tools to raise long-term funding.
Based on current market conditions, Foxconn’s credit profile allows us to obtain relatively competitive debt financing costs. We will use this as one of our main funding sources.
As for equity fundraising, at this stage, the company does not have plans to support business growth through a cash capital increase.
In terms of capital structure management, we will continue to maintain an investment-grade credit rating, and manage our financial structure using indicators such as ROC above 12% and net debt divided by EBITDA below 1.5 times.
Our principle is to maintain a solid capital structure and sufficient financial flexibility while supporting the growth of AI and the overall business.
So, to summarize, previous capital investments have gradually been reflected in revenue, operating profit, and cash generation capability.
Although future CAPEX and working capital requirements will continue to increase, at this stage the company has sufficient internal cash generation capability and financing capacity to support growth.
In addition, the company currently has cash and time deposits totaling NT$1.5 trillion, with net cash of NT$219.8 billion. Under this situation, we also maintain an S&P A- credit rating with a positive outlook.
With our current solid financial structure and abundant liquidity, we have a strong financial resilience and funding foundation to support the company’s continued growth, investment expansion, and response to market changes.

**Kristen Fang (Investor Relations)** [0:00:00]
Thank you, David. For those who want to ask questions, you may now click the "raise hand" button. After the next question, we will open the floor for questions from investors and media. This is the final question.
Question 3: The final question is for Michael, regarding margins. Considering changes in product mix, rising component and rack costs, and the smartphone peak season in the second half, how does the company view operating margin performance in the second half? In particular, as AI rack unit prices become higher and higher, can the company still maintain operating margin above 3%? Thank you.

**Michael Chiang (CEO)** [0:00:00]
Thank you, Kristen.
As the specifications of new-generation AI servers continue to upgrade, system complexity, component value, and rack costs will all increase.
At the same time, as the second half of the year enters the peak season for the consumer electronics industry, changes in product mix will also have some impact on gross margin and operating margin.
However, as a co-design and co-development partner for customers’ AI servers, we work with customers from the early stages of product development to introduce the concepts of Design for Automation and Design for Manufacturing.
For new-generation AI racks, together with the introduction of new components, we have already designed them toward the direction of fully automated production lines.
In other words, we have applied Foxconn’s experience in mass production capability, and introduced our strengths in standardized modules from the very beginning, significantly improving product manufacturability and production efficiency.
At the same time, we are also developing automated equipment simultaneously. This ensures that from the early stage of customers’ new product mass production, we can further improve production efficiency, product quality, and mass production capability through standardized, modularized, and automated design, together with automation equipment developed internally by Foxconn.
On the one hand, we can meet customers’ shipment targets. On the other hand, we can also ensure Foxconn’s leading advantage and continue to expand market share.
As AI rack specifications upgrade, the value per rack also increases. We will flexibly adopt either the buy and sell or consignment transaction model based on the characteristics of different customers and projects. Among these, ASIC AI server projects with a higher degree of customization are currently mostly based on the consignment model.
This not only reduces the working capital occupation from high-unit-price components, but also helps improve asset utilization efficiency, further optimizing our operating structure and financial quality. Although prices, cost structures, and product mix will change across different generations of AI products, we remain very confident in our operating efficiency and cost management.
Operating margin in the first half of this year reached 3.67%, up 0.67 ppts from the same period last year. Therefore, we expect this year’s performance to exceed last year’s 3.2% level, and our medium- to long-term target of maintaining operating margin above 3% remains unchanged.

**Kristen Fang (Investor Relations)** [0:00:00]
Thank you, Michael. Now, we will move to Q-and-A session for online investors as well as media. English questions are also welcome. Please state your questions completely in one turn. Each participant may ask up to two questions. Thank you.
The first question comes from analyst Anne from Nomura Securities. Please go ahead.

**Anne Lee (BofA Securities)** [0:00:00]
Anne Lee Nomura Securities- Analyst Question: My first question is about the company successfully entering the supply chain for TPU products this year. Could the company share the main product categories and services currently supplied, as well as what related services you expect to further pursue in the future? That,is my first question.
My second question is about Agentic AI driving CPU demand. How does the company view the outlook for the general server market? For this year, the company’s guidance is double- digit growth. But I would like to understand whether what you are seeing now is more optimistic compared with what you saw three months ago. Also, could you comment on the growth outlook for general servers next year? Those are my two questions. Thank you.

**Kristen Fang (Investor Relations)** [0:00:00]
Thank you. For the first question, we will ask Michael to respond.

**Michael Chiang (CEO)** [0:00:00]
Thank you, Anne, for the question. Regarding details of individual customers and individual products, it is not very convenient for us to comment here.
However, what we can share is that this year, we have indeed made some new progress in ASIC projects, including the introduction of new customers and new project opportunities. Therefore, in addition to having a very high share in the GPU market, our market share in ASIC will also increase significantly this year.
We have accumulated quite comprehensive product and manufacturing experience in the AI server field, and we will also have more comprehensive participation in customers’ future generations of products.
Foxconn has strong advantages in AI server system design and manufacturing capabilities, and we also provide dual solutions for both GPU and ASIC. Therefore, we are very confident in the development of ASIC projects, and we will continue to grow along with strong customer demand.
We will also take on different levels of roles based on the needs and characteristics of different projects, across key areas including modules, cooling systems, compute trays, and rack integration, in order to help customers accelerate mass production ramp-up.
In addition, our experience in the design and development of system-level rack architecture is also a capability that customers value highly. This experience can help customers develop more competitive ASIC servers, especially for future integration requirements in liquid cooling and high-power systems. Therefore, we believe this will be an important advantage for deepening cooperation with customers and expanding into more projects going forward.

**Executive** [0:00:00]
I will answer the second question. Since the beginning of this year, expectations for CPU demand have continued to expand. The main reason behind this is the development of AI.
As AI applications gradually increase, in addition to demand for AI training and inference, this will also drive more computing demand for general servers. We estimate that the overall general server market will still have solid growth momentum over the next few years.
Looking at this year, general servers are also expected to grow quarter by quarter. Growth in the second half compared with the first half will be more than double-digit. Demand visibility is indeed somewhat better than it was three months ago.
Based on industry estimates, shipments of CPUs for general servers are expected to increase by around 16% YoY in 2026. If we look at next year, this may further increase to 23%, showing that overall demand momentum continues to strengthen.
As we are one of the world’s major server ODMs, we have strong and long-term cooperation with major CSPs and brand customers. We expect Foxconn’s growth in general servers and CPU-based server applications to have the opportunity to outperform the overall industry.

**Kristen Fang (Investor Relations)** [0:00:00]
Next question comes from Wayne from Neuberger, please.
Wayne Liu Neuberger Berman - Analyst Question: I have two questions for the company. The first is regarding networking. For areas such as switches, transceiver modules, and components, could you share the current revenue scale and margin level within the company? In particular, regarding future market growth deployment, as well as guidance for revenue, profit, and market share. Especially for CPO and NPO, you just mentioned CPO switches. Is the company currently involved
only in these areas, or are there any quantitative targets that you can provide for our reference? That is the first part.
My second question is that U.S. CSPs have continued to raise their CAPEX outlook this year. Aside from these figures, has the company’s order visibility also improved at the same time? In addition to the several large CSP customers, have you recently seen more orders or demand from Neocloud customers? How does Foxconn view the demand outlook from Neocloud customers? Those are my two questions. Thank you.
Thank you. We will first ask Michael to answer the question regarding new customers.

**Michael Chiang (CEO)** [0:00:00]
Regarding your question, as CSP investment continues to expand, our order visibility is also improving at the same time. At present, we can already clearly see that orders next year will continue to move upward.
From a demand perspective, the business opportunities in AI datacenters are very large. The massive demand for AI computing power is mainly driven by four major categories: large language model companies, CSPs and Neocloud, governments, and enterprises. Presently, model companies, CSPs, and Neocloud are the main sources of demand.
Demand contribution from governments and enterprises is still relatively small at this stage. However, governments and enterprises around the world all know that they must keep up with the AI trend. They are just still exploring specific use cases. This represents a future demand driver with very strong potential.
In addition to large CSPs, we are also seeing increasing demand from more new cloud customers and AI model development companies. As we introduced in the presentation earlier, companies investing in large language model development hope to advance AI from today’s narrow AI to artificial general intelligence, and even to superintelligence.
However, simply entering the stage of artificial general intelligence requires processing more than one trillion parameters, which greatly increases the demand for computing power. Therefore, this type of customer has already become an important source of growth in current AI infrastructure demand.
Going forward, in addition to continuing to deepen cooperation with major CSP customers, we will also actively capture the incremental opportunities brought by new types of customers.
As AI applications gradually extend from model training to inference applications, enterprise AI services, sovereign cloud, and other areas, demand from new cloud customers is expected to continue growing. We should soon begin seeing contributions from new cloud customers before the end of this year.

**Executive** [0:00:00]
I will answer the second question regarding networking. In the first half of this year, Foxconn’s switch products benefited from simultaneous increases in both volume and ASP, and our networking product revenue achieved strong growth.
As AI rack computing power increases, demand for network bandwidth and high-speed interconnects has also increased. This has driven the continued upgrade of switch specifications.
We believe this will further drive the overall networking product segment to gradually move from traditional high-speed copper interconnects into optical connectivity and higher-end CPO technologies.
We believe optical communications and networking will become the next important growth trend for AI infrastructure. The progress of high-end switch products is currently in line with our internal expectations. Our CPO switches have already begun shipping. At present, they are mainly used in scale-out, meaning horizontal expansion between AI racks. Going forward, as platform generations and system architectures upgrade, they will also gradually expand into scale-up applications.
In addition to CPO switches, we are also continuing to strengthen our deployment in key components such as optical modules, optical-electrical integration, advanced packaging, cables, and connectors. We are also investing in modules related to high-speed transmission and power management.
By using vertical integration and increaseing the in-house production ratio of key components, we believe there is still further room to increase value-add, whether at the single-machine level or at the overall system level. As product complexity increases, it will also bring better profitability.
The market estimates that optical communications will maintain extremely rapid growth over the next few years, and the total market size may reach more than US$100 billion. Therefore, we will gradually expand the scale of CPO switches, optical transceivers, and related products. Looking at the long term, networking products will also become an important source of improved profit contribution for Foxconn.

**Kristen Fang (Investor Relations)** [0:00:00]
Next question comes from Kaylin from Macquarie. Kaylin, please.
Kaylin Tsai Macquarie Securities - Analyst Question: I have one larger question. This year, a number of ODMs have begun adopting the consignment model for some AI server projects. Foxconn has also mentioned that it has adopted this model in ASIC server projects. Going forward, is there also an opportunity for our GPU servers to adopt consignment? How does the company view the future changes in the proportion between the consignment and buy and sell transaction models? In addition, after adopting the consignment model, what percentage of first-half AI server revenue is currently accounted for by ASIC servers? As new projects begin mass production in the second half, how does the company view future revenue growth and market share for the ASIC business? Thank you.
Regarding the transaction model, we will ask the CFO to respond.

**David Huang (CFO)** [0:00:00]
Thank you, Kaylin. At present, consignment is mainly applied to ASIC projects.
As Rotating CEO Michael mentioned earlier, in the future, whether in ASIC or GPU projects, it is possible to adopt the consignment model. This mainly depends on customer requirements, project characteristics, and the cooperation model between both parties.
At present, the proportion of ASIC projects adopting consignment is relatively high, mainly because different customers have greater differences in application scenarios, system architectures, and customization requirements.
For Foxconn, consignment can reduce component inventory and capital occupation, while also reducing the risk of inventory price fluctuations. This helps improve asset turnover efficiency and overall financial quality.
As for the ratio between consignment and buy and sell, we will not set a fixed transition timetable or ratio. Instead, we will evaluate each customer and project on a case-by-case basis.
As ASIC projects continue to increase, and as some customers adjust their supply chain management models, we expect the share of consignment to gradually increase. This is also in line with the development trend of the AI server market. According to market estimates, ASIC’s shipment share within AI servers will gradually increase from around 28% this year to close to 40% by 2030.
Last year, around 10% of our AI server revenue came from ASIC servers. Looking at the first half of this year, ASIC’s revenue share was slightly higher than the full-year level last year. However, ASIC projects are currently mainly based on the consignment model. Therefore, looking only at the revenue recognized in the financial statements does not fully reflect the actual business scale.
Looking ahead to the second half, as new products begin shipping and the number of projects continues to increase, we expect ASIC shipment scale and profit contribution to continue improving. As for market share, our target for ASIC is also above 40%.
In addition to our existing product development and mass production experience, as mentioned earlier, we will also leverage our vertical integration advantages to participate more comprehensively in ASIC projects and provide complete services according to customer needs. We also maintain a very positive view on ASIC growth over the next few years.

**Kristen Fang (Investor Relations)** [0:00:00]
Next question comes from Pei-Ying from TVBS, please.
Pei-Ying Hsieh TVBS News - Reporter Question: Hello Michael, James, and CFO. This is Pei-Ying from TVBS. I would like to ask about the market’s high level of interest in Vera Rubin shipments and market share. When does Foxconn currently expect Vera Rubin to ramp up? What will the approximate scale be? Is it possible to secure close to 50% market share? Thank you.
For this question, we will ask Michael to respond.

**Michael Chiang (CEO)** [0:00:00]
Thank you, Pei-Ying, for the question. In terms of product progress, we maintain our plan for the new-generation AI racks to begin mass production preparation in the third quarter and begin shipments in the fourth quarter.
As product validation is gradually completed, and as the supply chain and capacity gradually come into place, we expect shipment volume in the following quarters to continue increasing. Next year, this will become our main product.
The automation level of new-generation AI racks has increased significantly. Together with improvements in our vertical integration and system delivery capabilities, this has helped us secure new customers and new projects, driving further improvement in market share. At present, we are still moving toward the target of 50%.
As for the overall market rack volume next year, we believe the key will still depend on how much CoWoS capacity allocation chip customers are able to obtain. The market currently expects CoWoS capacity to grow by more than 50% next year. However, how much of that can ultimately be converted into new-generation AI racks will still depend on chip supply conditions.
At present, the customer demand and project visibility we are seeing are both quite positive. We will continue to closely monitor the overall supply chain and datacenter build-out progress, while continuing to expand global AI capacity and prepare for mass production and shipments. Thank you.

**Kristen Fang (Investor Relations)** [0:00:00]
Next question comes from Angela from KGI Securities please.
Angela Hsiang KGI Securities - Analyst Question: Hello Michael, James, and CFO. I have two main questions. First, earlier we mentioned the progress of GPU racks or Vera Rubin racks. But in addition to GPU IT racks, could you also discuss some standalone racks that were previously mentioned, such as LPU racks or Vera CPU racks? Could you share the approximate mass production schedule for these products, how we should view their volume, and how their gross margin might compare with the company average? That is my first question.
The second question is more about some of the company’s key advantages. You mentioned that we are aiming to focus more on becoming an AI platform supplier. I think this positioning is certainly different from many other ODMs. In the presentation just now, you also mentioned the company’s cooperation with some international companies in AI datacenter infrastructure, such as Intel and Schneider, as well as cooperation in sovereign AI platforms. The company has also developed its own AI model and done optimization such as Foxbrain.
Could you further share how these deployments help our competitiveness? For example, in terms of profitability in our financial statements, or in AI production and manufacturing, will these add value to orders and lead to more visible contributions later on? Or where exactly will these enhance the advantages of our own business? These are my two questions. Thank you.

**Michael Chiang (CEO)** [0:00:00]
Thank you, Angela, for the question regarding standalone racks.
For customers’ next-generation CPU, LPU, and switch racks, we are currently carrying out capacity allocation and early-stage mass production preparation.
Based on the current progress, LPU and Vera CPU related products are expected to enter production in the third quarter, and gradually ramp up in the fourth quarter.
The larger-scale shipments are expected to take place next year.
Because the ASP of these products is relatively lower than GPU AI racks, even as shipment volume gradually increases next year, their impact on overall revenue and gross profit will be relatively limited compared with GPU racks.
However, the significance of these products is that no matter what platform products customers need, we can provide them.
For most projects, we also help customers achieve design and development from zero to one, as well as large-scale mass production from one to one hundred, becoming their most important partner.
This is my response to your question. Thank you.

**Executive** [0:00:00]
Let me briefly answer the part related to the AI platform.
Taiwanese companies basically focus more on AI hardware. However, we believe that for AI to ultimately generate greater value, there must be more focus on AI software and AI applications.
This is also an area where Foxconn is very focused.
For example, we have the three smart platforms. In Smart Manufacturing, I mentioned Genesis earlier, which is mainly about how to improve internal efficiency.
For Smart EV and Smart City, these can also become business models that we sell externally.
So, to answer your question, on the one hand, we can improve internal efficiency to enhance profitability. On the other hand, we can turn these projects into external sales, generating revenue and contribution.

**Kristen Fang (Investor Relations)** [0:00:00]
Due to time constraints, we will open the floor for one final question. Next question comes from Melody from EBC Financial News, please.
Melody Chang EBC Financial News - Reporter Question: Hello, Michael and CFO. This is Melody from EBC Financial News. I would like to ask a question related to SHARP. This year marks the tenth anniversary of Foxconn’s investment in SHARP. Could you share the cooperation and positioning between the two sides for the next ten years? Thank you.

**Michael Chiang (CEO)** [0:00:00]
Indeed, today is a very special day.
On this day in 2016, Foxconn completed its investment in SHARP. So today is an important milestone marking the tenth anniversary of Foxconn and SHARP joining hands.
We have always believed that SHARP is not only a globally renowned brand, but also an important asset of Japan. It is also a company that represents Japan’s deep spirit of innovation and craftsmanship culture.
Ten years ago, Foxconn believed in SHARP. Over the past ten years, we have gone through reform and transformation together.
Today, we remain very firm in our support for SHARP.
Now, Foxconn has firmly established itself as the world’s largest ICT manufacturer. In addition, at a time when AI is developing rapidly, we are also the world’s largest AI server provider, and we are expanding into new business areas through 3+3+3.
These areas are all resources that SHARP can make good use of.
Over the next ten years, we look forward to working together with SHARP to create new growth opportunities in areas such as AI, robotics, electric vehicles, and smart living.
Our goal has always been the same: to help SHARP once again become a brand that makes Japan proud and that the world looks forward to.
Last week, SHARP also officially announced its first-generation latest AI server. This product is the result of cooperation between SHARP and Foxconn, and is the best demonstration of the two companies jointly developing AI business opportunities.
We believe we will see more and more similar cooperation cases in the future.

**Kristen Fang (Investor Relations)** [0:00:00]
The current time is 4:05 p.m. We will conclude the investor conference call here. Thank you all for your participation. If you have any additional questions afterward, please feel free to contact our IR team. Thank you, everyone. Goodbye.
Disclaimer This English version is a non-verbatim translation of Mandarin transcript for reference only.
Hon Hai Precision Industry Co., Ltd. (the “Company”) reserves the right to make changes to content, or other information in this document without obligation to notify any person of such changes.
Email : ir@foxconn.com | Website : https://www.honhai.com/en-us/
In the conference call upon which this transcript is based, the Company may make projections or other
forward-looking statements regarding a variety of items. Such forward-looking statements are based upon
current expectations and involve risks and uncertainties. Actual results may differ materially from those stated
in any forward-looking statement based on a number of important factors and risks. Although the Company
may indicate and believe that the assumptions underlying the forward -looking statements are reasonable, any
of the assumptions could prove inaccurate or incorrect and, therefore, there can be no assurance that the
results contemplated in the forward-looking statements will be realized.
THE INFORMATION CONTAINED IN THIS TRANSCRIPT IS A TEXTUAL REPRESENTATION OF THE COMPANY'S
CONFERENCE CALL AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE
MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBS TANCE OF THE
CONFERENCE CALLS. IN NO WAY DOES THE COMPANY ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT
OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED IN ANY TRANSCRIPT.
Email : ir@foxconn.com | Website : https://www.honhai.com/en-us/
