Samsung Electronics chairman Jay Y. Lee recently held high-level meetings with Meta, Qualcomm, and Amazon leaders to discuss AI, cloud services, and chip technology cooperation. Lee’s discussions with Meta’s Mark Zuckerberg, held at Zuckerberg’s home, included AI and virtual and augmented reality topics. In separate meetings, Lee explored semiconductor collaboration with Amazon CEO Andy Jassy and Qualcomm CEO Cristiano Amon, focusing on memory chips for Amazon’s data centres, cloud services, and chip manufacturing for Qualcomm’s mobile processors.
These discussions are part of Lee’s two-week visit to the United States and will inform Samsung’s upcoming company-wide strategic meeting at the end of June. The meetings aim to bolster Samsung’s position in the competitive semiconductor market, particularly in high-end memory for AI, where Samsung has faced challenges. The company’s newly appointed head of the semiconductor division has emphasised the need for a collective effort to navigate these challenges.
Samsung maintains strong relationships with these tech giants, supplying chips to Amazon for various applications and partnering on TV, mobile, and content projects. With Qualcomm, Samsung’s high-end smartphones feature Snapdragon mobile processors, and the partnership now extends to AI-enabled PCs. Meta, another key client, has collaborated with Samsung on projects like the Gear VR headset and foldable smartphones, showcasing the deep integration of their technologies.
Samsung Electronics is stepping up its game in the AI chip market by streamlining its contract manufacturing process to offer clients a comprehensive, one-stop shop. By integrating its top-tier memory chip, foundry, and chip packaging services, the tech giant aims to accelerate AI chip production by about 20%. The new approach allows clients to work with a single communication channel, significantly reducing the production time for AI chips. At a recent San Jose, California event, Siyoung Choi, President and General Manager of Samsung’s Foundry Business, highlighted the transformative impact of generative AI on the technology landscape.
Looking ahead, Samsung projects the global chip industry revenue to reach $778 billion by 2028, mainly driven by the demand for AI chips. Executive Vice President of Foundry Sales and Marketing, Marco Chisari, echoed this optimism, agreeing with OpenAI CEO Sam Altman’s predictions of a sharp increase in AI chip demand. Unlike other companies, Samsung’s unique combination of selling memory chips, offering foundry services, and designing chips under one roof has often been seen as a competitive disadvantage. However, the rising demand for highly integrated AI chips that process large datasets quickly and efficiently is turning this setup into a competitive advantage.
Samsung is also making strides in chip architecture with its cutting-edge gate-all-around (GAA) technology. The advanced transistor architecture improves chip performance while reducing power consumption, which is crucial for developing more powerful AI chips as they become smaller. While other competitors, including TSMC, are also advancing in GAA technology, Samsung has a head start and plans to mass-produce its second-generation 3-nanometre chips using GAA later this year. Furthermore, Samsung unveiled its 2-nanometre chipmaking process for high-performance computing chips, with mass production scheduled for 2027, positioning itself at the forefront of the AI chip revolution.
Two partnerships were unveiled at Apple’s yearly Worldwide Developer Conference on Monday; one on stage and one in the fine-print. A partnership with OpenAI to use GPT 4.0 within Siri’s Apple Intelligence was openly publicised, but use of Google chips to build the AI tools was not.
Initially, it would seem as though the two companies are at odds. Apple would be set to compete with Google’s Gemini over their AI systems, while the OpenAI partnership could potentially mean reduced access to customer data through Siri.
However, a technical document published by Apple after the event reveals that in order to build their AI models, Apple used its own framework software. This software depended on various pieces of Apple’s own hardware, but also tensor processing chips (TPCs). These are exclusively available for purchase through Google’s cloud. Google is among various companies competing with Nvidia’s AI-capable chips, which have been dominating the market recently.
Apple did not immediately reply after Reuters requested comment. It has not detailed how much it depends on third-party chips for the development of its new AI system.
MediaTek, a leading semiconductor company in Taiwan, is collaborating with Microsoft to design an ARM-based chip specifically for its Windows operating system AI-powered laptops. The strategic partnership marks a significant move in the tech industry. Last month, Microsoft unveiled a new generation of laptops featuring chips designed with ARM Holdings technology, providing the necessary power to run advanced AI applications. These applications are considered the future of consumer computing by Microsoft executives.
MediaTek’s new chip is set to be integral in this effort and is expected to bring substantial advancements in processing power and efficiency. That collaboration underscores the growing demand for high-performance, energy-efficient chips tailored for AI applications. Their ARM-based chip is designed to optimize AI tasks, leveraging ARM’s architecture known for its power efficiency and performance scalability. The development aligns with the industry trend of integrating specialized hardware to handle AI workloads more effectively, reducing the reliance on general-purpose CPUs.
For Microsoft, this partnership with MediaTek represents a strategic move to strengthen its position in the competitive AI hardware market. By incorporating MediaTek’s advanced chip technology, Microsoft aims to offer more capable AI laptops, appealing to both consumer and enterprise markets. The collaboration also takes direct aim at Apple, which has been using its own ARM-based chips for Mac computers for roughly four years. Also, Microsoft’s decision to optimise Windows for ARM could pose a significant challenge to Intel’s long standing dominance in the PC market. For decades, Windows machines have relied on chip architectures developed by Intel and AMD. MediaTek’s PC chip is expected for release late next year, coinciding with the end of Qualcomm’s exclusive deal to supply ARM-based chips for Windows laptops.
Why does it matter?
Enhanced AI capabilities: The integration of MediaTek’s specialised chip will enable Microsoft’s AI laptops to perform more complex AI tasks with greater efficiency and speed. That advancement is crucial as AI applications become more sophisticated and demand higher computational power.
Energy efficiency: ARM-based chips are known for their power efficiency, which means that AI laptops equipped with these chips will likely have longer battery life and reduced energy consumption. This is particularly important for users who require high performance without sacrificing mobility.
Market innovation: The partnership could set a new standard in the AI hardware market, encouraging other tech companies to develop and integrate specialised chips for AI applications. That could lead to a surge in innovation, driving the development of more advanced and capable AI devices.
Competitive edge: For both MediaTek and Microsoft, this collaboration provides a competitive edge. MediaTek can showcase its capacities in developing cutting-edge chips, while Microsoft can differentiate its AI laptops in a crowded market by offering superior performance and efficiency.
The US government is set to announce expanded sanctions on semiconductor chips and other goods sold to Russia, targeting third-party sellers in China. That move is part of a broader effort by the Biden administration to thwart Russia’s attempts to bypass Western sanctions and sustain its war efforts against Ukraine. The new measures will extend existing export controls to include US-branded goods, even those not made in the United States. They will identify specific Hong Kong entities involved in shipping goods to Moscow.
These upcoming sanctions come as President Joe Biden prepares to attend a summit with other Group of Seven (G7) leaders in southern Italy, where supporting Ukraine and weakening Russia’s military capabilities are top priorities. US officials have expressed increasing concern over China’s growing trade with Russia, which they believe is enabling Moscow to maintain its military supplies by providing essential manufacturing equipment. The broadened export controls aim to address this issue by encompassing a wider range of US goods.
Additionally, the US plans to impose significant new sanctions on financial institutions and non-banking entities involved in the ‘technology and goods channels’ that supply the Russian military. That decision comes amid efforts to ensure that Ukrainian President Volodymyr Zelenskiy can emphasise the critical situation facing Ukrainian forces in their ongoing struggle against Russia during his meetings with G7 leaders.
On Tuesday, the US Commerce Department announced plans to award Rocket Lab $23.9 million to significantly boost the production of compound semiconductors used in satellites and spacecraft. This funding aims to increase the production of space-grade solar cells by 50% over the next three years, bolstering the supply chain for critical space technologies. The funds come from the Biden administration’s $52.7 billion chip manufacturing and research subsidy program, benefiting major companies like Samsung, Intel, and TSMC.
In addition to federal support, the State of New Mexico has committed $25.5 million in financial assistance and incentives to help Rocket Lab expand its Albuquerque facility. Rocket Lab, founded in 2006 by New Zealander Peter Beck, specialises in producing highly efficient, radiation-resistant space-grade solar cells. These solar cells are essential for powering various US space programs, including missile awareness systems, the James Webb Space Telescope, and NASA’s Artemis lunar explorations.
Commerce Secretary Gina Raimondo emphasised the importance of solar cells in maintaining communication and space technology operations. The $23.9 million award will help Rocket Lab meet the growing demand for these components from the US military, NASA, and the commercial space industry. The expansion in Albuquerque is expected to create over 100 manufacturing jobs and enhance the company’s capacity to support US space missions.
Rocket Lab has secured numerous federal contracts worth hundreds of millions of dollars for satellite production and spacecraft launches. The proposed investment, which could be adjusted after further review by the Commerce Department, will also allow Rocket Lab to claim up to 25% in investment tax credits for qualified capital expenditures, part of the broader benefits provided by the Chips Act.
In a significant development for both the technology and geopolitical landscapes, Intel has reportedly halted its plans to expand a factory in Israel, a project initially valued at $25 billion. That decision, first reported by Israeli media, comes as a blow to the Israeli tech industry and raises questions about the future of major international investments in the region.
Intel, one of the world’s leading semiconductor manufacturers, has a longstanding presence in Israel, with multiple facilities that play a crucial role in the company’s global operations. The proposed factory at Intel’s Kiryat Gat site was intended to be a crucial component of Intel’s strategy to strengthen the global supply chain, complementing investments in Europe and the United States. Intel’s Kiryat Gat plant, known as Fab 28, currently produces 10-nanometer chips, labeled as Intel 7 technology. The new Fab 38 plant was slated to open in 2028 and operate through 2035. Intel employs nearly 12,000 people in Israel, and the suspension of this project raises concerns about future employment and economic stability.
While Intel has not officially confirmed the reasons behind the halt, several factors are believed to have influenced the decision, such as geopolitical tensions, economic considerations, and regulatory environment. Despite the halt, Intel emphasized its ongoing commitment to Israel, stating that ‘Israel continues to be one of our key global manufacturing and R&D sites and we remain fully committed to the region.’ The company acknowledged that managing large-scale projects in the semiconductor industry requires flexibility in timelines, driven by business conditions, market dynamics, and responsible capital management.
Why does it matter?
Israeli officials have yet to comment on the reports, but there is likely to be considerable concern over the potential economic and reputational impacts. Efforts to negotiate and address Intel’s concerns might be on the horizon to revive the project or mitigate the fallout. For Intel, this move might be part of a broader strategic realignment. The company has been undergoing significant changes under the leadership of CEO Pat Gelsinger, focusing on boosting its manufacturing capabilities in the US and other regions. The shift could reflect a more cautious approach to international investments, prioritising regions with more predictable political and economic climates.
This week, key players in the chip industry, including Nvidia, Intel, AMD, Qualcomm, and Arm, gathered in Taiwan for the annual Computex conference, announcing an ‘AI PC revolution.’ They showcased AI-enabled personal computers with specialised chips for running AI applications directly on the device, promising a significant leap in user interaction with PCs.
Intel CEO Pat Gelsinger called this the most exciting development in 25 years since the arrival of WiFi, while Qualcomm’s Cristiano Amon likened it to the industry being reborn. Microsoft has driven this push by introducing AI PCs equipped with its Copilot assistant and choosing Qualcomm as its initial AI chip supplier. Despite this, Intel and AMD are also gearing up to launch their AI processors soon.
Why does it matter?
The conference was strategically timed to precede Apple’s annual Worldwide Developers Conference, hinting at the competitive landscape in AI advancements. As the PC market shows signs of recovery, analysts predict a rise in AI PC adoption, potentially transforming how PCs are used. However, there needs to be more skepticism about whether consumer demand will justify the higher costs associated with these advanced devices, as the Financial Times reports.
In a recent interview, Intel CEO Pat Gelsinger presented how Intel intends to remain a relevant actor in the Chinese chip market while also scaling up production in the US. This was done at the occasion of the Computex tech conference in Taipei, where Intel released its new Xeon 6 processor, destined for data centres. Its release comes at a time when tech giants are challenging Nvidia’s chip dominance.
Gelsinger aims to build an Intel foundry in the US after the organisation was incentivised to increase facilities in the US with as much as $8.5 billion in grants and $11 billion in loans under the CHIPS and Science Act. In its release, the White House stated this is a step towards ‘protecting national security’ and increasing US share of global chip production to ‘20% […] by the end of the decade’.
“The capital is critical. We said that we have to have economic competitiveness if we build these factories in the US, and that’s what the CHIPS Act has done. It’s created a level playing field if I were building a factory in Asia versus US,” Gelsinger said.
Why does it matter?
At the same time, Gelsinger reiterated the importance of the Chinese market to his company. “China is a big market for Intel today, and one that we’re investing in to be a big market for Intel tomorrow as well,” he said. Intel has been competing to catch up its global market share since 2017, when South Korea’s Samsung overtook it as the largest chipmaker in terms of revenue. Since then, Taiwan Semiconductor Manufacturing Company reportedly overtook Samsung in 2023.
Chinese AI chip firms, including industry leaders such as MetaX and Enflame, are downgrading their chip designs in order to comply with Taiwan Semiconductor Manufacturing Company’s (TSMC) stringent supply chain security protocols and regulatory requirements. This strategic adjustment comes amidst heightened scrutiny and restrictions imposed by the US on semiconductor exports to Chinese companies, which includes limitations on accessing advanced manufacturing technologies critical for AI chip production.
The US has imposed strict export controls to obstruct China’s military advancements in AI and supercomputing. These controls include restrictions on sophisticated processors from companies like Nvidia, as well as on-chip manufacturing equipment crucial for advanced semiconductor production. That move has prevented TSMC and other overseas chip manufacturers using US tools from fulfilling orders for these restricted technologies.
In response to these restrictions, top Chinese AI chip firms MetaX and Enflame have reportedly submitted downgraded chip designs to TSMC in late 2023. MetaX, founded by former Advanced Micro Devices (AMD) executives and backed by state support, had to introduce the C280 chip after its more advanced C500 Graphic Processing Unit (GPU) ran out of stock in China earlier in the year. Enflame, also Shanghai-based and supported by Tencent, faces similar challenges.
Why does it matter?
The decision to downgrade chip designs to meet production demands reflects the delicate balance between technological advancement and supply chain resilience. While simplifying designs may expedite production and mitigate supply risks in the short term, it also raises questions about long-term innovation and competitiveness. The ability to innovate and deliver cutting-edge AI technologies hinges on access to advanced chip manufacturing processes, which are increasingly concentrated among a few global players.
