Logistical issues led Elon Musk to redirect numerous Nvidia chips, initially intended for Tesla’s electric vehicles, to X Corp. On Tuesday, Musk explained that Tesla had nowhere to send the Nvidia chips, so they would have just remained in storage.
That’s a response to a CNBC report highlighting a memo from Nvidia stating that 12,000 of its top AI chips, originally meant for Tesla, were sent to X instead. Future shipments intended for X were later reassigned to Tesla.
Musk also announced that the Gigafactory in Texas is nearly complete and will house 50,000 H100 chips. He mentioned that about half of Tesla’s $10 billion AI-related spending this year will be for internal use, including the AI inference computer and Dojo supercomputer. He noted that Nvidia hardware accounts for two-thirds of the cost of building AI training superclusters, and Tesla plans to spend $3-4 billion on Nvidia hardware this year. Tesla is working on its own supercomputer to advance driverless-car technology, aiming to increase the number of active H100s from 35,000 to 85,000 by year-end.
Why does it matter?
The following situation has sparked criticism that Musk’s focus on AI and robotics might detract from Tesla’s core car business. Musk, who currently holds 13% of shares directly and about 21% with options, has requested 25% ownership to increase his influence. In January, he threatened to take his advanced technology ideas elsewhere if he isn’t granted more ownership.
Air Liquide has announced plans to invest $250 million in a new plant in Idaho aimed at supplying gases to US semiconductor manufacturer Micron Technology Inc. The plant is scheduled to be operational by the end of 2025.
The facility, strategically located to respond to Micron’s growing production needs, will produce ultra-high purity gases essential for semiconductor fabrication processes. Micron has started this year with the mass production of its high-bandwidth memory (HBM) semiconductors for use in Nvidia’s latest chip for AI.
These gases play a crucial role in ensuring the quality and reliability of semiconductor chips, integral to a wide range of technologies, including smartphones, computers, and automotive electronics. The investment is poised to ‘support the production of leading-edge memory chips, notably to meet the growing demand for computing capacities required by AI’, according to Matthieu Giard, Air Liquide’s CEO for the Americas.
Why does it matter?
The investment by Air Liquide in this new plant is crucial for several reasons. First, it enhances the reliability of the semiconductor supply chain, which has been under strain due to increased global demand and supply chain disruptions. By securing a local and reliable source of critical gases, Micron can better manage production timelines and mitigate risks associated with international logistics. The expansion also aligns with broader industry trends towards enhancing semiconductor manufacturing capabilities. As technology advances, particularly in areas such as AI, 5G networks, and autonomous vehicles, the demand for high-performance semiconductors continues to grow. Air Liquide’s investment ensures that Micron is well-positioned to meet this growing demand and maintain its competitive edge in the market.
Vanguard International Semiconductor and NXP Semiconductors are set to build a $7.8 billion chip manufacturing plant in Singapore to cater to the rising demand in the region. The joint venture between Taiwan-based Vanguard, partly owned by Taiwan Semiconductor Manufacturing Co. (TSMC), and Dutch chip maker NXP will establish a 300 mm semiconductor wafer manufacturing facility. Construction is scheduled to commence in the second half of 2024, with initial production expected by 2027. Vanguard will hold a 60% stake in the venture, while NXP will own the remaining 40%.
The new facility will produce chips for the automotive, industrial, consumer, and mobile markets. That’s part of a broader global trend, as countries and companies invest heavily to secure a leading position in the semiconductor industry. According to the consulting firm International Business Strategies, the industry is projected to surpass $1 trillion by the decade’s end.
Other nations are also ramping up their semiconductor investments. South Korea recently announced a $19 billion support package for its chip industry. China raised $48 billion for its largest national semiconductor fund, and Japan and the EU have committed significant government support to bolster their semiconductor sectors.
According to a draft of this year’s long-term economic policy plan, Japan is considering new legislation to support the commercial production of advanced semiconductors. The plan, set to be finalised around 21 June, aims to strengthen the chip supply chain by promoting domestic production, enhancing human resources, and boosting research and development in collaboration with international partners.
The draft emphasises the need for legislative measures to facilitate the mass production of next-generation semiconductors. Recently, the industry ministry highlighted the necessity of a new regulatory framework to support the chip foundry venture Rapidus, which aims to begin mass production of cutting-edge chips by 2027. Tokyo has already agreed to provide up to 920 billion yen ($5.94 billion) in subsidies for Rapidus, focusing on research and development.
However, additional support, such as government guarantees, will be required for mass production to attract investment. Rapidus, led by industry veterans, plans to produce these advanced chips on Hokkaido island in collaboration with IBM and research organisation Imec from Belgium, marking a significant step in Japan’s efforts to revive its semiconductor industry.
Intel has launched its sixth generation Xeon server processors and revealed its Gaudi 3 AI accelerator chips, aiming to regain its position in the data centre market. The new Xeon processors come in two versions: a high-performance model for complex AI computations and an efficiency model designed to reduce power consumption and server rack requirements significantly. Intel’s market share in the data centre segment has been slipping, with AMD gaining ground due to Intel’s previous manufacturing setbacks.
Intel’s Gaudi 3 AI chips are notably priced lower than competitors, with an eight-chip accelerator kit costing about $125,000 compared to over $300,000 for a similar Nvidia setup. Intel CEO Pat Gelsinger highlighted the cost-effectiveness of the Gaudi 3 chips during a presentation at the Computex trade fair, emphasising their competitive edge. The efficiency model of the Xeon chips is available immediately, while the performance model will be released in the third quarter.
In addition to server processors, Intel introduced its next-generation laptop chip, Lunar Lake, which promises a 40% reduction in power usage and enhanced AI processing capabilities. Set to ship in the third quarter, Lunar Lake is part of Intel’s broader strategy to dominate the AI PC market. Future advancements include the Arrow Lake chip next year and Panther Lake in 2025, showcasing Intel’s ongoing innovation and collaboration with TSMC to enhance its technology.
Advanced Micro Devices (AMD) unveiled its latest AI processors at the Computex technology trade show in Taipei on Monday, signalling its commitment to challenging Nvidia’s dominance in the AI semiconductor market. AMD CEO Lisa Su introduced the MI325X accelerator, set for release in late 2024, and outlined the company’s ambitious roadmap to develop new AI chips annually. The move aligns with Nvidia’s strategy, as both companies race to meet the soaring demand for advanced AI data centre chips essential for generative AI programs.
AMD is not only aiming to compete with Nvidia but also to surpass it with innovations like the MI350 series, expected in 2025, which promises a 35-fold improvement in AI inference performance over current models. The company also previewed the MI400 series, set for 2026, featuring a new architecture called ‘Next’. Su emphasised that AI is the company’s top priority, driving a focus on rapid product development to maintain a competitive edge in the market.
The shift towards an annual product cycle reflects the growing importance of AI capabilities in the tech industry. Investors who have been keenly following the AI chip market have seen AMD’s shares more than double since the start of 2023, though Nvidia’s shares have surged even more dramatically. AMD’s plans include AI chip sales projections of $4 billion for 2024, up $500 million from previous estimates, and introducing new central processor units (CPUs) and neural processing units (NPUs) for AI tasks in PCs.
Why does it matter?
As the PC market looks to rebound from a prolonged slump, AMD is banking on its advanced AI capabilities to drive growth. Major PC providers like HP and Lenovo are set to incorporate AMD’s AI chips in their devices, which already meet Microsoft’s Copilot+ PC requirements. This strategic focus on AI-enhanced hardware highlights AMD’s commitment to staying at the forefront of technological innovation and market demand.
The European Commission has given the green light to Italian state aid for semiconductor manufacturer STMicroelectronics to construct a €5 billion microchip plant in Sicily. This approval comes as Europe seeks to reduce dependence on Asian imports for critical manufacturing components. The plant in Catania will specialise in producing microchips that enhance energy efficiency in electric vehicles and will receive a direct grant of about €2 billion from Rome.
The move comes amidst heightened scrutiny of Europe’s reliance on Asian chip supplies due to pandemic-related disruptions and trade tensions with China. To address these concerns, the EU has introduced its Chips Act, aiming to attract chip manufacturers and secure vital components for hi-tech industries. European antitrust chief Margrethe Vestager emphasised the strategic importance of diversifying chip supply chains and reducing dependencies on single suppliers.
As major shareholders in STMicro, the governments of France and Italy are backing the project. The new plant, STMicro’s second facility in Sicily, will produce silicon carbide chips known for their energy efficiency. This initiative signals a shift towards self-sufficiency in semiconductor production and supports Europe’s digital and green transition objectives. With operations expected to reach full capacity by 2032, the plant aims to bolster regional security of chip supply and meet the growing demand from automakers like Tesla, BMW, and Renault.
According to Prime Minister Anwar Ibrahim, Malaysia aims to attract a minimum of 500 billion ringgit ($107 billion) in investments for its semiconductor industry. With Malaysia already accounting for 13% of global testing and packaging in the semiconductor sector, the country seeks to solidify its position as a global manufacturing hub. In recent years, it has successfully attracted multibillion-dollar investments from industry giants like Intel and Infineon.
Anwar highlighted that the targeted investment will focus on areas such as integrated circuit design, advanced packaging, and manufacturing equipment for semiconductor chips. Additionally, Malaysia aims to establish at least ten local companies specialising in semiconductor chip design and advanced packaging, with projected revenues ranging from $210 million to $1 billion. To support these ambitious goals, Malaysia plans to allocate $5.3 billion in fiscal support, with further details expected to be announced later.
The Prime Minister emphasised Malaysia’s capacity to diversify and ascend the value chain, emphasising a shift towards high-end manufacturing, semiconductor design, and advanced packaging. Although Anwar still needs to provide a specific timeline for achieving these targets, Malaysia has already begun taking concrete steps to bolster its semiconductor sector. Initiatives include plans to build Southeast Asia’s largest integrated circuit design park, offering incentives such as tax breaks and subsidies to attract global tech companies and investors. These efforts signify Malaysia’s strategic move towards high-value, front-end design work beyond traditional backend chip assembly and testing.
China’s domestic AI chipmakers are rapidly closing the gap on international leaders, according to Xu Bing, co-founder of SenseTime Group Inc. Despite the significant lag in computational power compared to the US, China possesses the talent and data necessary to advance in the AI field, Xu stated during an interview at the UBS Asian Investment Conference in Hong Kong. SenseTime, a leading AI company in China, faces challenges due to US sanctions that restrict access to advanced AI technology, such as Nvidia’s accelerators.
The US trade controls have spurred the development of domestic alternatives from companies like Huawei Technologies and Shanghai Biren Technology, both also affected by US restrictions. Xu emphasised that although Asia faces a considerable shortfall in computational resources, the region is abundant in talent and data. He noted that China’s AI chip industry is catching up quickly, with SenseTime collaborating with local semiconductor firms to enhance their computing capabilities.
While the exact gap between Chinese and US AI technology is uncertain, estimated between one to three years, Xu is optimistic that this disadvantage in computing power will be temporary. He believes that, over time, the disparity in computing resources will diminish, viewing computing power as a commodity China will eventually acquire in sufficient quantity. Notable Chinese companies making strides in AI chips include Moore Threads Intelligent Beijing Co., Huawei, and other key players like Baidu Inc. and Naura Technology Group Ltd, which have received government attention and support.
South Korea’s finance minister, Choi Sang-mok, announced plans to introduce detailed tax incentives to increase the value of listed companies as part of ongoing corporate reforms. These measures, which are part of the ‘Value-up Programme’, will be refined after gathering feedback from market participants in June and July. The government aims to balance fairness and effectiveness in the proposed tax benefits, initially introduced in February, which were seen as insufficient by the market.
In addition to these corporate reforms, Choi emphasised continued government support for the vital semiconductor industry in South Korea. Next month, detailed measures will be released to bolster the industry’s global competitiveness, building on a recently announced support policy package.
Choi also addressed broader economic issues, highlighting support for a three-way free trade agreement (FTA) with Japan and China following a trilateral summit where leaders expressed their commitment to accelerating negotiations. On the domestic front, he projected consumer inflation to stabilise in the mid-to-lower 2% range in the latter half of the year. He anticipated higher tax revenue in 2024 despite currently weaker-than-expected corporate tax income.
