Musk promises continued production increases at Shanghai Gigafactory, praises Jensen Huang remotely as the "Michael Jordan of tech," and unveils Terafab chipmaking timeline.

date
20:08 19/11/2026
avatar
GMT Eight
During this trip to Shenzhen, Musk made two statements, pointing respectively to the Shanghai Gigafactory that already underpins Tesla's deliveries and to an advanced chipmaking plan that has yet to be proven. On November 19, when meeting with Ren Hongbin, chairman of the China Council for the Promotion of International Trade, he called himself a "super fan of China" and promised to continue increasing production in China.
The world's richest person and CEO of SpaceX and Tesla, Inc., Musk, said in an interview on Thursday that Tesla, Inc., a leader in AI autonomous driving, Siasun Robot & Automation, and electric vehicle manufacturing, will continue to increase production in China, once again signaling that the company will continue to deepen its manufacturing operations in the world's largest auto market. In another remote interview on Thursday, Musk also disclosed a series of latest heavyweight frontier technology roadmap outlooks and core production capacity plans for Musk's business empire. In an interview with Chinese media, Musk called himself a "super fan" of China and lavished praise on China's architecture, natural scenery, and railway system. Media cited him as saying, "China is cool. I always encourage people from all over the world to come to China and see these amazing places for themselves. There are stunning cities, magnificent natural scenery, and world wonders like the Terracotta Warriors." Musk also said in the interview, "The more China and the United States cooperate, the more both sides benefit." Musk arrived in Shenzhen on November 18 with a U.S. delegation to participate in activities related to APEC Leaders' Week 2026. Chinese media said this was the first publicly reported visit by Musk to Shenzhen. After arriving, Musk wrote on social platform X, "Can the Thucydides Trap be avoided? I hope so, but the most interesting outcome is precisely the most likely one." The "Thucydides Trap" refers to the situation in which, when a rising power challenges an established power, competition between the two intensifies, GEO Group Inc political tensions rise accordingly, and conflict may ensue. A publicly circulated video also captured Musk's interaction with his young son who accompanied him to the meeting. Musk asked his son, "Do you like China?" The child answered, "Yes." Musk then joked that his son's toys are all printed with "Made in China," which made the child especially look forward to visiting China. Musk's Shenzhen trip: promises continued production increases at Shanghai Gigafactory! Praises Jensen Huang remotely Musk's two statements during this trip to Shenzhen point respectively to the Shanghai Gigafactory, which already supports Tesla, Inc.'s deliveries, and to an advanced chip manufacturing plan that still needs to be validated. In the first 10 months of this year, the Shanghai plant delivered more than 810,000 vehicles, up 21.4% year on year. The increase came mainly from overseas: exports rose 101.3% year on year to 421,000 vehicles, accounting for nearly 52% of the plant's deliveries, while Tesla, Inc.'s retail sales in the Chinese market fell 15% during the same period. Shortly before Musk's interview with Chinese media, Musk remotely joined the TomorrowX summit in Austin from Shenzhen and praised Jensen Huang as the "Michael Jordan" among tech entrepreneurs. He also stressed that this praise does not conflict with his Terafab chipmaking ambition. According to Musk's estimate, xAI currently adds about 0.5 gigawatts of computing infrastructure per month as measured by power generation capacity, equivalent to about 220,000 NVIDIA Corporation GB300 units and supporting systems. Terafab targets longer-term supply capacity, with plans to produce the first test chip around June 2027, and if all goes smoothly, first-phase mass production is expected by the end of 2028. For investors focused on AI semiconductors and Musk's business empire, the profit potential of the Shanghai production increase depends on export orders and per-vehicle profit, while whether the actual capacity of the "Terafab" super chip plant led by Musk can change the future chip procurement landscape depends on mass production yield and cost after the test chip. On November 19, when Musk met with Ren Hongbin, chairman of the China Council for the Promotion of International Trade, in Shenzhen, he called himself a "super fan" of China and said Tesla, Inc. will continue to increase production in China. What best explains his decision is the order structure of the Shanghai plant. In the first 10 months of this year, the plant delivered a total of 810,889 Model 3 and Model Y vehicles, up 21.4% year on year. Of these, exports totaled 420,984 vehicles, doubling year on year, while Tesla, Inc.'s retail sales in the Chinese market were 389,905 vehicles, down 15% year on year. An even more significant figure is that exports as a share of the Shanghai plant's wholesale sales rose to 51.9% from 31.3% in the same period last year, with overseas orders having become the main driver of delivery growth. The Shanghai plant currently has annual capacity of more than 950,000 vehicles, with a local parts procurement rate of 98%. Whether increased production can boost automotive business profit depends on whether overseas orders can continue to dilute fixed costs per vehicle and how much profit margin export freight and regional pricing can leave for Tesla, Inc. Musk did not announce a new production target or expansion timetable this time. In U.S. premarket trading that day, Tesla, Inc.'s stock price at one point rebounded about 2%, and driven by the Anthropic IPO frenzy, AI computing and semiconductor stocks rebounded more strongly. Anthropic priced at $83 per share, with a public offering size of about $81.34 billion, and including concurrent private placements, the total transaction size was about $107.34 billion, surpassing SpaceX to become the largest IPO in history. German brokerage grey-market quotes reached as high as $105, a premium of about 26% over the issue price. As of publication, the triple-leveraged semiconductor ETF (SOXL.US) surged more than 10% in U.S. premarket trading, with its leverage target corresponding to the single-day performance of the NYSE Semiconductor Index; Nasdaq 100 futures rose 2.15%, NVIDIA Corporation rose 1.7% premarket, Intel Corporation, AMD, Micron, and Marvell Technology, Inc. rose between 3.2% and 4.6% premarket, and AI cloud service providers CoreWeave and Nebius both rose more than 5%. The synchronized strength from chips to cloud service providers reflects the market's reassessment of the revenue trajectory of AI applications after large-scale financing by the most frontier AI large-model developers, and of the ability of AI infrastructure-related players that massively purchase and rent computing power to expand substantially. From Claude's computing bills to Optimus humanoid Siasun Robot & Automation's hands, unpacking Musk's Shenzhen There is already a priced computing contract between Anthropic and Musk. Anthropic's third-quarter revenue grew about 47% quarter on quarter to $16.956 billion, while computing and infrastructure costs during the same period reached $9.863 billion. Its prospectus discloses that the company has multiple NVIDIA Corporation GPU computing agreements with SpaceX's AI subsidiary, and after all expected capacity is delivered in December 2026, the total monthly fee will be about $3.1 billion. Each agreement allows early termination with 90 days' written notice after the initial non-cancellable period ends. This partly explains why Anthropic's listing can move chip and AI computing-related hot rental stocks as well as tech giants such as Tesla that benefit from the AI boom, and it also clarifies Tesla, Inc.'s position in this. The computing rental revenue belongs to SpaceX's business, while Tesla, Inc. is directly involved in the power supply segment. At the TomorrowX summit, Musk revealed that the team uses Tesla, Inc. Megapack energy storage systems to smooth load fluctuations during training on large GPU clusters, allowing power generation equipment to still supply power stably when training tasks suddenly increase or decrease load. xAI is seeking both more electricity and more training per kilowatt-hour. When Musk connected from Shenzhen to Austin, he said that measured by power generation capacity, xAI currently adds about 0.5 gigawatts of computing infrastructure per month, equivalent to about 220,000 NVIDIA Corporation GB300 units plus supporting CPUs, networking, and cooling facilities, with the goal of raising the monthly increase to more than 1 gigawatt by the end of 2027. The larger the cluster, the higher the cost of GPUs waiting for data from other nodes. To this end, engineers formerly responsible for Starlink and Starship software participated in rewriting training, reinforcement learning, and inference software, mainly using C and some C++, to reduce communication waiting and software overhead. Grok 4.8, which has entered the reinforcement learning stage, focuses on the other end, namely correcting errors in training questions and scoring, so that newly added computing power truly improves the model's reliability in handling user tasks. Musk is building a chip plant while calling Jensen Huang the "Michael Jordan" among tech entrepreneurs. The two correspond to different investment horizons. For now, xAI's cluster expansion still uses the NVIDIA Corporation platform on a large scale, and Musk believes that losing Jensen Huang's leadership would be a major blow to NVIDIA Corporation. More long term, he hopes to gain more manufacturing capability through Terafab, with the goal of having the research fab produce the first test chip around June 2027, and if all goes smoothly, first-phase mass production will be achieved by the end of 2028. Whether the first chip can work tests whether the design and process can run through. Moving from trial production to stable supply also depends on wafer yield, line throughput, and the cost per qualified chip. Starship recovery brings orbital computing one step closer, and the next threshold is low-cost reflight. In its 15th test flight, SpaceX for the first time caught the Starship upper stage returning from orbit with a tower. Musk estimates that if full, rapid reuse is achieved, the cost of reaching space could fall to one hundredth of the traditional level. His envisioned Starmind computing satellites would be powered by photovoltaic arrays, cooled by radiator panels, and connected to Starlink via laser links. SpaceX is targeting the manufacture of space photovoltaic modules with total nameplate power of 200 gigawatts per year. Turning this manufacturing goal into rentable in-orbit computing power also depends on launch frequency, payload per launch, and the satellites' ability to continuously dissipate heat. Anthropic has publicly expressed interest in jointly exploring multi-gigawatt orbital computing, while what the two sides have already signed for Claude's use is the ground-based Colossus 1 data center. When it comes to Siasun Robot & Automation and brain-computer interfaces, commercial progress depends on what people or machines can independently accomplish. Musk expects that Tesla, Inc. may demonstrate Optimus's real work capability within one or two months, and after about 12 to 18 months it will be able to wash, dry, and put away dishes. Wet and slippery dishes require fingers to adjust grip in real time, while tightening screws in the battery compartment requires precise coordination between the wrist and fingertips, which is why he listed the hands and forearms as the biggest engineering difficulty. He is more optimistic that continuous-operation positions such as mining and smelting will generate economic returns first, because higher equipment utilization helps dilute purchase costs. He also disclosed that more than 30 people have received Neuralink implants, and that Telepathy aims to help paralyzed people operate computers or robotic arms with their thoughts. The next product, Blindsight, is intended to help blind people gain visual perception by stimulating the visual cortex, and remains a research and development goal.