Foxconn Chairman Young Liu says there is no risk of overcapacity in the AI server market because demand from governments and enterprises remains strong. The company is actively expanding its U.S. manufacturing operations, despite higher costs, and has six factory buildings in Houston, Texas that produce AI servers for Nvidia. Liu also revealed that Foxconn is working on a new collaboration with OpenAI after meeting with CEO Sam Altman, though no formal announcement has been made yet.
Foxconn is aiming to capture 50 percent of the global AI server market within a couple of years. Its production capacity is fully booked for at least a year and a half, and AI infrastructure is expected to make up 60 to 65 percent of revenue in three to five years. Hon Hai Precision Industry, Foxconn's parent company, expects double-digit revenue growth next year driven by AI data center demand. The company also showcased a new Nvidia DSX POD modular data center and three new electric vehicle models.
In the AI model space, Artificial Analysis launched the HLE-Diamond benchmark, which tests models against 1,000 research questions. Google's Gemini 4 Argon topped the rankings with a score of 69.1 percent, followed by Claude Opus 5.5 at 67.9 percent and GPT-6 Astra at 63.8 percent. The benchmark found that reasoning performance is much higher than knowledge performance across most models, and open weights models still trail closed models by over 30 points. When choosing a frontier AI model for products, cost, error rates, and consistency matter more than benchmark scores alone. GPT-6.1 Sol is the cheapest at $0.72 per task, while Claude Opus 5.5 scores highest overall. Gemini 4 Argon has the lowest hallucination rate at 15 percent.
On the business and investment front, Nvidia's proposed financing arrangement could boost chip demand by helping fund data center projects without heavy upfront costs. Anthropic agreed to spend $45 billion renting computing capacity at the Nscale campus. However, Wall Street warns that beating earnings estimates no longer guarantees gains for AI stocks. Raymond James strategist Matt Orton said about one-quarter of AI firms that beat forecasts still saw shares fall over 5 percent. Cash generation is now a key concern, as seen with Amazon's negative $42 billion free cash flow, though Orton highlighted Microsoft for tighter financial discipline.
Hyundai Motor Group plans to build a $4.3 billion AI data center in Saemangeum, Korea, with construction starting in March 2027 and commercial operations planned for early 2030. The center will serve as a supercomputing hub for training autonomous driving algorithms. Meanwhile, forensic accountant Anthony Scilipoti is warning that current AI spending resembles the Nortel Networks era, comparing the boom to a party running out of alcohol. Shopify and Dexcom are positioned to benefit from AI, with Shopify tapping into agent-driven commerce and Dexcom using AI to improve its continuous glucose monitoring devices.
Key Takeaways
- Foxconn is collaborating with OpenAI and already manufactures AI servers for Nvidia at six Houston factory buildings, with plans to expand U.S. operations.
- Foxconn targets capturing 50 percent of the global AI server market within a couple of years, with production capacity fully booked for at least 18 months.
- Google's Gemini 4 Argon leads the new HLE-Diamond AI benchmark at 69.1 percent, followed by Claude Opus 5.5 at 67.9 percent and GPT-6 Astra at 63.8 percent.
- When selecting frontier AI models, cost and error rates matter more than benchmark scores; GPT-6.1 Sol is cheapest at $0.72 per task, while Gemini 4 Argon has the lowest hallucination rate at 15 percent.
- Nvidia's financing plan could boost chip demand, with Anthropic committing $45 billion in computing capacity rentals at the Nscale campus.
- Wall Street warns that AI stock earnings beats no longer guarantee gains, with about one-quarter of companies still declining after beating forecasts.
- Amazon reported negative $42 billion free cash flow, while Microsoft is highlighted for tighter financial discipline in the AI sector.
- Hyundai is building a $4.3 billion AI data center in Saemangeum, Korea, set to begin construction in March 2027 for autonomous driving algorithm training.
- Foxconn's parent company Hon Hai expects double-digit revenue growth next year, driven by strong AI data center product demand.
- Shopify and Dexcom are positioned for AI-driven growth, with Shopify focusing on agent-driven commerce and Dexcom improving glucose monitoring devices with AI.
Foxconn says AI demand strong with no overcapacity risk yet
Foxconn Chairman Young Liu says there is no risk of overcapacity because AI demand is still strong. Governments and enterprises are only in the early stages of adopting AI. He also discussed plans to grow Foxconn's U.S. manufacturing operations despite higher costs.
Foxconn working on new tie-up with OpenAI and denies AI bubble
Foxconn is working on a new collaboration with OpenAI. Chairman Young Liu met with OpenAI CEO Sam Altman and said an announcement will come once details are confirmed. Foxconn already has six factory buildings in Houston, Texas that make AI servers for Nvidia. The company also deployed more than 10 humanoid robots at its Houston plant and plans to expand that effort.
Foxconn targets 50 percent of global AI server market
Foxconn Chairman Young Liu says the company could capture 50 percent of the global AI server market within a couple of years. Its production capacity is fully booked for at least a year and a half. AI infrastructure is expected to make up 60 to 65 percent of revenue in three to five years. The company plans to expand U.S. manufacturing but will focus on high-margin products first.
Hon Hai expects double-digit revenue growth next year
Hon Hai Precision Industry expects revenue to grow by a double-digit percentage next year, driven by strong demand for AI data center products. Chairman Young Liu met with OpenAI CEO Sam Altman to discuss potential cooperation. Hon Hai plans to expand its six Houston factory buildings. The company also showcased a new Nvidia DSX POD modular data center and three new electric vehicle models.
Gemini 4 Argon tops new HLE-Diamond AI benchmark
Artificial Analysis launched HLE-Diamond, a benchmark with 1,000 research questions. Google's Gemini 4 Argon scored 69.1 percent to lead the rankings. Claude Opus 5.5 scored 67.9 percent and GPT-6 Astra scored 63.8 percent. The benchmark found that reasoning performance is much higher than knowledge performance across most models. Open weights models still trail closed models by over 30 points.
How to Pick the Best Frontier AI Model for Your Product
Choosing a frontier AI model depends on more than benchmark scores. Cost, error rates, and consistency matter more in agentic apps. GPT-6.1 Sol is cheapest at $0.72 per task, while Claude Opus 5.5 scores highest. Gemini 4 Argon has the lowest hallucination rate at 15 percent. No single model wins at everything.
Writer Produces 370 Books in 20 Months Using AI Tools
A writer completed about 370 books in 20 months with help from artificial intelligence. The author used AI tools for idea generation, drafting, and finalizing texts. This experiment shows how AI is reshaping the publishing industry. It also raises questions about content quality and intellectual property rights.
ThinkJev Matches Thinking Speed to Decision Complexity
ThinkJev uses one hybrid-thinking model to answer typed decisions fast or slow. System 1 answers directly while System 2 thinks first. At a 5 percent error target, Qwen3.5-27B answers 24.5 percent of decisions with a 3.9 percent realized error. The approach saves thinking tokens by choosing when to engage deeper reasoning.
Hyundai to Build $4.3 Billion AI Data Center in Saemangeum
Hyundai Motor Group will build a $4.3 billion AI data center in Saemangeum, Korea. Construction starts in March 2027 with commercial operations planned for early 2030. The center will serve as a supercomputing hub for training autonomous driving algorithms around the clock. Two special purpose companies will manage construction and GPU procurement.
Forensic Accountant Anthony Scilipoti Turns Focus to AI Industry
Anthony Scilipoti, a forensic accountant and CEO of Veritas Investment Research, is now targeting the AI industry. His team previously warned investors about overhyped stocks like Nortel Networks. He compares current AI spending to the Nortel era, noting overhyped exuberance and circular money flows. Scilipoti describes the AI boom as a party running out of alcohol.
Shopify and Dexcom poised for major growth from AI
Two stocks, Shopify and Dexcom, are positioned to benefit from the AI revolution. Shopify is tapping into agent-driven commerce, a growing but still small part of its business. Dexcom is using AI to improve its continuous glucose monitoring devices with personalized health tools. Both companies have large markets still untapped, offering long-term growth potential.
Kentucky man uses religious law to opt out of AI at work
David Durman, a University of Kentucky administrator, invoked Title VII to avoid using AI tools at work. He follows the Humanist faith and believes offloading decisions to AI conflicts with his beliefs. The university granted his accommodation request in September. His case highlights growing moral objections to AI use in the workplace.
Nvidia AI financing plan may boost chip demand, Wedbush says
Nvidia's proposed financing arrangement could support AI infrastructure projects and drive demand for its chips. Wedbush says the move may help fund data center projects without heavy upfront costs. Anthropic agreed to spend $45 billion renting computing capacity at the Nscale campus. The plan comes amid concerns about debt financing at other Nvidia-backed providers like CoreWeave and Firmus.
Wall Street warns AI stock earnings beats no longer guarantee gains
Wall Street is warning that beating earnings estimates is no longer enough for AI stocks to rise. Raymond James strategist Matt Orton said about one-quarter of AI firms that beat forecasts still saw shares fall over 5%. Cash generation is now a key concern, as seen with Amazon's negative $42 billion free cash flow. Orton highlighted Microsoft for tighter financial discipline and sees potential in health care stocks like Eli Lilly.
Sources
- No risk of overcapacity yet as large part of AI demand still unrealized: Foxconn Chairman
- Foxconn working on a new tie-up with OpenAI, says there is no AI bubble
- Foxconn hopes to own 50% of the global AI server market within a couple years
- Hon Hai expects revenue growth in double digits
- Gemini 4 Argon Tops New HLE-Diamond Benchmark On Closed-Ended Research Questions
- How to Choose the Right Frontier AI Model for Your Product
- From writing to millions: A writer completes 370 books in 20 months with the help of artificial intelligence
- ThinkJev: One Thought, Every Budget for Calibrated Typed Decisions
- Hyundai Motor Group to build $4.3 billion AI data center in Saemangeum
- Heavy hitter: Anthony Scilipoti’s next target? AI
- 2 Stocks That Could Quietly Become Big Winners in the Artificial Intelligence (AI) Revolution
- As bosses push workers to use AI, a Kentucky man found a way to opt out: a federal law protecting religious beliefs
- Nvidia’s AI financing push could sustain infrastructure investment and drive chip demand, Wedbush says
- Wall Street Delivers Stark Warning for AI Stocks as Earnings Beats Fail to Deliver
Comments
Please log in to post a comment.