AI chip startup Positron has raised $23.5 million in a bid to compete with industry leader Nvidia. The Reno-based company, which manufactures its chips in Arizona, claims its processors consume less than a third of the power of Nvidia’s high-performance H100 chips while maintaining similar capabilities. Investors in the funding round included Valor Equity Partners, Atreides Management, and Flume Ventures.
Positron’s chips are designed for AI inference, the stage where trained AI models are used rather than developed. While demand is currently higher for training chips, analysts predict that inference chips could soon become the more sought-after option as AI applications expand. This shift has led major players such as OpenAI, Google, and Meta to invest heavily in AI infrastructure, with spending expected to reach tens of billions of dollars this year.
Although Nvidia dominates roughly 80% of the AI chip market, rising costs and concerns over reliance on a single supplier have pushed major tech firms to seek alternatives. With its latest funding, Positron positions itself as a strong contender in the growing US and global AI chip industry, offering a more energy-efficient option for future AI applications.
For more information on these topics, visit diplomacy.edu
Groq, a US semiconductor startup, has secured a $1.5 billion commitment from Saudi Arabia to expand its AI chip delivery. The Silicon Valley firm, founded by a former Alphabet engineer, specialises in AI inference chips designed for speed and efficiency in pre-trained model execution. A partnership with Aramco Digital has already established an AI hub in the region.
Funding will be received throughout the year to support the expansion of Groq’s data centre in Dammam. The company’s chips, used for chatbots and large language models, comply with US export regulations, with necessary licences already secured for shipments to Saudi Arabia.
The investment was announced at Saudi Arabia’s LEAP 2025 technology event, where the country confirmed $14.9 billion in new AI investments. Groq’s chips will help power the Dammam data centre, which will support Allam, a Saudi-developed AI language model operating in Arabic and English.
Groq reached a $2.8 billion valuation last August after raising $640 million in funding from Cisco Investments, Samsung Catalyst Fund, and BlackRock Private Equity Partners.
For more information on these topics, visit diplomacy.edu.
Powerchip Technology, a Taiwanese IT company, is facing intense competition from Chinese foundries like Nexchip, which has rapidly gained market share in the legacy chip sector. This shift, driven by steep discounts and aggressive capacity expansion, has been accelerated by China’s localisation efforts, forcing Powerchip to retreat from the once-profitable business of making integrated circuits for Chinese flat panels. The increasing dominance of Chinese companies in the $56.3 billion legacy chip market is causing concern in Taiwan, with companies like Powerchip and UMC now focusing on more advanced technologies to stay competitive.
Chinese foundries, supported by strong government funding and low margins, have significantly increased their production capacity, undercutting Taiwanese rivals on price. By 2027, China is projected to surpass Taiwan’s global mature node manufacturing capacity. Taiwanese executives are exploring specialisation and diversification, shifting focus from legacy chips to more advanced processes like 3D stacking, which integrates logic and DRAM memory to improve performance.
The rising competition from China is compounded by geopolitical tensions, as some customers are now requesting chips made outside China. This shift is partly due to the US’s trade policies and worsening relations between Beijing and other nations. Taiwanese companies are seeing more orders directed to their local fabs, with some customers explicitly avoiding ‘Made in China’ products.
While Taiwanese companies still have an edge in terms of process stability and production yields, the pressure from Chinese competitors is forcing them to rethink their strategies and adapt to the changing landscape. The future of the industry may depend on how Taiwan navigates both the rising Chinese competition and the geopolitical challenges shaping global supply chains.
OpenAI announced on Thursday that it is evaluating US states as potential locations for data centres supporting its ambitious Stargate project, which aims to secure the US’s lead in the global AI race. The project is seen as crucial for ensuring that AI development remains democratic and open, rather than falling under authoritarian control, according to Chris Lehane, OpenAI’s chief global affairs officer.
Stargate, a venture backed by SoftBank, OpenAI, Oracle, and other investors, is set to receive up to $500 billion for AI infrastructure. A significant portion of this investment, $100 billion, will be deployed immediately, with the rest scheduled over the next few years. Texas has been designated as the flagship location for Stargate’s data centres. An initial site under construction in Abilene is expected to begin operations later this year.
The announcement follows the rise of DeepSeek, a Chinese AI model that challenges the traditional view that AI development requires large, specialised data centres. DeepSeek’s use of cheaper chips has raised concerns among investors, leading to a significant drop in tech stock values, including a record $593 billion loss for Nvidia, the leading AI chipmaker.
OpenAI is considering data centre locations in approximately 16 states, with plans to expand the Stargate network to five to ten campuses in the coming months.
Thailand plans to draft a strategic plan for its semiconductor sector within 90 days, aiming to attract new investments amid the growing trade tensions between the US and China. The country’s national semiconductor board will engage a consultancy to create an industry roadmap, with Narit Therdsteerasukdi, secretary-general of the Thailand Board of Investment (BOI), leading efforts to promote the sector. As part of these efforts, Narit is also organising roadshows in the US and Japan to draw in semiconductor investments.
The semiconductor industry has faced significant disruption due to the US-China trade war, and further instability is expected as US President Donald Trump’s renewed tariffs on Chinese imports continue. Despite this, Thailand’s semiconductor sector has seen growth, with inbound investment applications reaching a decade-high of 1.14 trillion baht ($33.5 billion) in 2023. The country aims for 500 billion baht in new investments by 2029, focusing on power electronics, including semiconductors for electric vehicles, data centres, and energy storage systems.
Thailand is positioned as a key player in the global semiconductor market, ranking second among emerging economies for semiconductor manufacturing. Companies like Analog Devices, Sony, Toshiba, and Infineon have facilities in Thailand, and investment in printed circuit boards, essential for electronic devices, has also surged. Thailand’s neutral position in the ongoing trade conflict makes it an attractive destination for investors seeking stability.
However, Thailand faces stiff competition from other Southeast Asian countries, particularly Malaysia, which is aiming for over $100 billion in semiconductor investments. Despite this, Thailand’s growth potential remains strong, driven by its growing reputation as a manufacturing hub for electronics.
Chinese companies are increasingly backing DeepSeek‘s AI, marking a pivotal moment for the industry. Firms like Moore Threads and Hygon Information Technology are enabling their computing clusters to support DeepSeek’s R1 and V3 models, which use domestically produced graphic processing units (GPUs). Analysts have hailed this as a ‘watershed moment,’ particularly as these models rival those run on global high-end chips.
Huawei has also joined the trend, integrating DeepSeek’s models with its Ascend cloud service and partnering with AI infrastructure start-up SiliconFlow. This integration showcases the growing potential of Chinese-made chips to support competitive large language models, reducing reliance on US hardware. Additionally, major Chinese tech companies such as Alibaba, Baidu, and Tencent have made DeepSeek’s models available through their cloud services.
DeepSeek’s rise has captured significant attention, especially after the launch of its free AI assistant, which surpassed ChatGPT in app downloads within days. The company’s approach, requiring far less computing power than its US counterparts, has further fueled its success. While DeepSeek is gaining traction globally, some countries, including Italy and the Netherlands, have raised privacy concerns, leading to investigations and blocks on its app.
AMD is set to report strong fourth-quarter results, with revenue expected to rise over 22% to $7.53 billion. However, competition in the AI chip market is intensifying, as Nvidia maintains its dominance and major tech firms such as Microsoft, Amazon, and Meta develop custom silicon. Analysts warn that the shift towards in-house AI processors could limit AMD’s growth in the long term.
The rise of Chinese AI startup DeepSeek has also raised concerns, as its models reportedly rival Western alternatives at a lower cost. Demand for AI chips remains high, but supply constraints continue to pose challenges. While contract manufacturer TSMC is working to expand capacity, Nvidia’s ramp-up of its latest Blackwell AI chips may restrict AMD’s access to manufacturing resources.
Despite these pressures, AMD’s AI chip sales could reach $10 billion in 2024, double the company’s initial forecast. The data centre chip segment is expected to see significant growth, contributing over half of total revenue. Meanwhile, AMD’s personal computer unit is expanding, with sales set to rise nearly 33% as the company gains market share from Intel.
German semiconductor materials supplier Siltronic has forecast stagnant sales for 2025 after a 7% decline last year, citing high inventory levels and weak demand in key sectors. The company’s shares fell more than 15% following the announcement, continuing a downward trend from 2024. While AI is driving some growth, it has not yet offset lower demand for automotive, PC, and memory chips.
CEO Michael Heckmeier stated that wafer demand is unlikely to recover soon, with the first half of 2025 expected to be weaker than late 2024. Siltronic plans to halt production of smaller silicon wafers at its Burghausen facility by the end of July, which will slightly impact sales but have little effect on earnings. The company also acknowledged that its mid-term targets set for 2028 will not be met, though it did not provide a new timeline.
Analysts reacted negatively to the outlook, with Stifel’s Juergen Wagner warning of possible downward revisions to 2025 earnings forecasts. Last week, STMicroelectronics also reported continued weakness in the automotive and industrial chip markets. Siltronic’s preliminary 2024 revenue stood at €1.41 billion, slightly above expectations, but the company has cut its dividend. A more detailed financial outlook will be presented in its annual report on 6 March.
The US administration under President Donald Trump is weighing stricter controls on Nvidia’s H20 chips, which were specifically designed for the Chinese market. Discussions, still in the early stages, build on previous export restrictions established by former President Joe Biden to limit the shipment of advanced AI chips to China. Nvidia’s H20 chips comply with existing US regulations but could face additional curbs due to concerns about China’s rapid progress in AI technology.
The potential restrictions come amid mounting unease about China‘s competitive edge in AI. Recent developments, such as the launch of DeepSeek‘s cost-efficient AI assistant, have intensified worries that China may be narrowing the gap with the US in AI development. RAND researcher Lennart Heim revealed that the possibility of further controls on AI chip exports had been discussed for over six months, originating during Biden’s tenure.
Nvidia has expressed its willingness to collaborate with the US government as the administration formulates its policy on AI exports. The company’s stock, already trading lower, saw additional losses following the report. Despite previous restrictions, Nvidia’s H20 chips have so far remained compliant with US regulations and available for shipment to China.
Intel has received $2.2 billion in federal grants as part of the US CHIPS and Science Act, supporting its efforts to boost domestic semiconductor production. The funding, awarded by the Department of Commerce, is part of a total $7.86 billion grant announced last November, aimed at expanding Intel’s manufacturing and advanced packaging operations across several states.
While an additional $5.66 billion is still to be disbursed, concerns have emerged over the future of the US CHIPS Act under the Trump administration. A proposed federal funding freeze, currently blocked by a judge, could impact the Commerce Department’s work on semiconductor subsidies. Despite this uncertainty, Intel’s leadership remains optimistic, citing ongoing discussions with the new administration.
Intel executives have expressed confidence in continued government support, highlighting shared goals of strengthening US semiconductor production. The company plans to use the funds to enhance its facilities in Arizona, New Mexico, Ohio, and Oregon, reinforcing America’s position in the global chip industry.
