Nvidia has added around 200 employees in China this year, boosting its research capabilities and focusing on autonomous driving technologies, Bloomberg News reports. The company now employs nearly 600 people in Beijing, with a recently opened office in the Zhongguancun tech hub, according to sources familiar with the matter.
Despite its expansion, Nvidia faces headwinds in China, including an ongoing investigation into alleged violations of the country’s anti-monopoly laws. This probe is widely interpreted as a response to US restrictions on China’s chip sector. Nvidia declined to comment on the situation.
China contributed about 17% of Nvidia’s revenue in the year ending January, down from 26% two years earlier. Globally, Nvidia employs around 29,600 people across 36 countries, as detailed in its February 2024 filing.
Apple is teaming up with Broadcom to develop its first server chip dedicated to artificial intelligence processing. Expected to enter mass production by 2026, the chip, code-named Baltra, represents Apple’s latest venture into custom hardware aimed at reducing reliance on Nvidia’s high-cost and supply-constrained processors.
The chip will use Taiwan Semiconductor Manufacturing Co.’s N3P process, highlighting Apple’s commitment to furthering manufacturing technology. Broadcom, already benefiting from AI-driven growth, saw its shares climb 5% following the announcement.
This partnership builds on Apple’s recent successes in chip growth, including its M-series processors for Mac devices. It also echoes a broader trend among tech giants like Google, which have pursued custom AI chips to enhance their services. With the global market for custom chips projected to reach $45 billion by 2028, Apple’s move furthers its role as a leader in the AI tech industry.
STMicroelectronics, a major European semiconductor firm, has introduced the STM32N6 series, its first microcontrollers designed for edge AI and machine learning. The new release aims to enhance applications in consumer and industrial electronics.
The STM32N6 series supports local image and audio processing, eliminating the need for larger computers or data centres. Devices such as cars, factories, and wearables are expected to benefit from faster and more efficient data handling.
Edge AI operates on principles similar to generative AI, but with reduced computational demands tailored to specific local tasks. By processing data on-site, the technology reduces electricity usage and avoids the delays associated with sending data to remote centres.
STMicroelectronics’ innovation underscores the growing importance of energy-efficient AI solutions in modern electronics. The development reflects a shift towards integrating AI more seamlessly into everyday devices.
Morris Chang, the founder of Taiwan Semiconductor Manufacturing Co (TSMC), has criticised Intel’s former CEO Pat Gelsinger for focusing on becoming a contract chipmaker rather than prioritising AI. Speaking at an event to launch his autobiography, Chang suggested that Intel’s strategy under Gelsinger may have lacked direction, noting the company’s struggles in regaining leadership in chip manufacturing.
Gelsinger’s departure earlier this month followed frustrations over slow progress on Intel’s costly foundry initiative. Chang observed that finding a new CEO and strategy would present significant challenges for the company. He also recalled Intel’s historical missteps, such as rejecting an invitation to invest in TSMC during its early years, only to become a key customer later.
Intel’s foundry ambitions led to strained relationships with partners, including TSMC. Chang described Gelsinger as ‘a bit rude’ during his tenure, with reports of lost contracts compounding the company’s struggles. Intel has not commented on these remarks.
Alphabet’s stock jumped 5% on Tuesday after the company unveiled Willow, a groundbreaking chip that addresses a major hurdle in quantum computing. Introduced on Monday, Willow solved a problem in five minutes that would take classical computers longer than the age of the universe. This breakthrough brings quantum computing closer to practical applications in science, medicine, and finance.
Quantum computers, powered by qubits, promise incredible speed but face challenges with error rates that grow with scale. Google’s Willow chip reduces errors exponentially by stringing qubits together in a way that allows for real-time error correction. This innovation marks a significant step toward making quantum computing reliable and commercially viable.
While the technology is still in its infancy, experts believe Willow could drive breakthroughs across industries. Alphabet’s shares, up 25% this year, outpaced market expectations, with investors optimistic about the potential impact of quantum computing on the company’s future. Other tech leaders like Microsoft and Quantinuum are also advancing in the race to commercialise quantum systems.
The US Department of Commerce has finalised a $6.1B subsidy for Micron Technology, marking one of the largest awards under the CHIPS and Science Act. The funds will support the construction of new semiconductor plants in New York and Idaho, with expectations to create at least 20,000 jobs by the end of the decade.
An additional $275M investment has been earmarked for expanding Micron’s Virginia facility, which focuses on producing chips for critical sectors like defense, automotive, and networking. The White House emphasised that this effort aims to strengthen national security by reducing reliance on foreign suppliers, particularly from China and Taiwan.
The Biden administration has made semiconductor independence a priority, providing similar funding to companies like Intel and Taiwan Semiconductor Manufacturing Co. These initiatives aim to secure the US position in global chip production, a vital industry for technological and economic stability.
China has launched an anti-monopoly investigation into US chipmaker Nvidia, citing possible violations linked to its acquisition of Mellanox Technologies in 2020. The move is widely seen as retaliation against recent US restrictions on China’s semiconductor sector, including curbs on advanced chip exports.
Nvidia, which dominates China’s AI chip market, is accused of breaching conditions tied to the Mellanox deal, such as fair trade requirements and restrictions on bundling products. The probe comes amid heightened trade tensions, with Beijing encouraging local companies to avoid US chips and Washington tightening export controls on critical semiconductor technologies.
Analysts suggest the investigation is unlikely to significantly impact Nvidia in the near term, as its most advanced chips are already restricted from sale in China. However, the company faces growing competition from domestic firms like Huawei. China accounted for 17% of Nvidia’s revenue last year, down from 26% two years prior, as US-China tech conflicts continue to reshape the global semiconductor landscape.
Apple plans to launch its own cellular modem chips next year, starting with the iPhone SE, replacing components from Qualcomm. This marks a significant step towards reducing reliance on the chipmaker, whose modems have powered iPhones for years. Apple aims to phase out Qualcomm’s technology entirely by 2027.
The move follows Apple’s $1 billion acquisition of Intel‘s modem unit in 2019 and its integration of modem design into its chip development team. Additionally, Apple has signed a multi-billion-dollar deal with Broadcom for 5G components.
Qualcomm has acknowledged Apple’s transition but is leveraging advancements in laptops and AI-powered data centres to offset expected revenue declines. Investors and suppliers like Skyworks Solutions and Qorvo are closely monitoring the impact of Apple’s strategy on their business.
A Rotterdam court is set to hold a pretrial hearing on Monday concerning a former Russian employee of ASML accused of stealing intellectual property from the Dutch semiconductor equipment maker. The suspect, a 43-year-old Russian national, allegedly profited by selling company manuals, including those of ASML’s Mapper subsidiary, to Russian buyers, according to Dutch media reports.
ASML, which acquired Mapper in 2019, confirmed its awareness of the case and said it had filed a formal complaint, declining further comment during ongoing legal proceedings. The suspect is reportedly in custody, though details of the arrest remain unclear.
Mapper, a Dutch firm focused on developing E-beam lithography technology, was integrated into ASML following its 2019 bankruptcy. While Mapper’s product did not succeed, its engineers joined ASML’s chip-measuring business, helping to bolster the company’s capabilities. This acquisition eased concerns about sensitive technology falling into foreign hands, a priority for both the Dutch government and the US military.
Taiwan Semiconductor Manufacturing Co (TSMC) is negotiating with Nvidia to produce its Blackwell AI chips at TSMC’s new plant in Arizona, according to insider sources. The Blackwell chips, unveiled in March, are designed for generative AI and accelerated computing, boasting speeds up to 30 times faster for tasks like chatbot responses. Production at the Arizona facility is expected to commence next year.
While TSMC will handle the front-end chip production in Arizona, the chips will require shipment back to Taiwan for essential packaging. Arizona’s plant lacks chip-on-wafer-on-substrate (CoWoS) capacity, which is crucial for Blackwell chips. TSMC currently reserves all CoWoS operations for its Taiwan facilities.
The agreement, if finalised, would expand the client base of TSMC’s Arizona site, which already counts Apple and Advanced Micro Devices among its customers. Neither TSMC nor Nvidia commented on the ongoing discussions, and Apple and AMD have yet to respond to inquiries.
TSMC’s investment in Arizona includes three advanced facilities, supported by US government subsidies aimed at revitalising domestic semiconductor production. The collaboration with Nvidia highlights the increasing global demand for AI chip technology.
