TSMC leads the AI chip race, thriving on surging demand, while Samsung struggles with a 13% profit drop and ASML casts doubt on AI chip sustainability. Chinese tech giants adapt to U.S. trade limits with homegrown solutions, keeping the global competition fierce in the AI-driven market.
In this week's electrifying Tech Wars wrap-up, global tech titans are making bold moves that blend ambition with intrigue. The U.S. is setting ethical boundaries on AI in national security, aiming to outpace rivals like China without compromising democratic values—think AI with a conscience guarding the nuclear codes. Delta Electronics isn't waiting for the energy fairy; they're building their own renewable power plants in India and Thailand to fuel the skyrocketing demand for AI chips, proving that green is the new gold in tech innovation.
Meanwhile, the semiconductor arena heats up as China's arrest of a South Korean engineer on espionage charges underscores that microchips are the new pawns in geopolitical chess, with South Korea scrambling to protect its silicon supremacy. Amidst this high-stakes drama, TSMC surges ahead, riding the "insane" AI demand wave, leaving competitors like Samsung and ASML questioning if they've bet on the right horse.
Not to be outdone, Intel invests a cool $300 million to expand its Chengdu facility, deftly navigating the U.S.-China tech rivalry to keep its foothold in a market too big to ignore. It's a wild ride where sustainability meets strategy, and the race for AI dominance is reshaping the global tech landscape faster than you can say "semiconductor."
In a significant policy shift, President Joe Biden introduced new guidelines regulating artificial intelligence use within the US national security framework. Signed as a national security memorandum, the directive aims to set boundaries on AI applications within the Pentagon and intelligence agencies, ensuring they align with democratic values. The guidelines explicitly prohibit AI applications that could infringe upon rights, such as free speech, or bypass established nuclear control measures.
"Our memorandum directs the first-ever government-wide framework on our AI risk management commitments,"
remarked US national security adviser Jake Sullivan. He emphasized that the directive prioritises avoiding harmful bias, maximising accountability, and maintaining effective human oversight. Officials noted that the guidelines encourage responsible AI experimentation while discouraging practices misaligned with ethical standards.
The directive forms part of the Biden administration's broader efforts to responsibly foster AI use while remaining competitive with global powers like China. National security applications covered by the guidelines include cybersecurity, counter-intelligence, and military logistics.
Additionally, the memorandum assigns the AI Safety Institute in Washington the task of inspecting AI tools to mitigate risks before deployment. Notably, these guidelines are non-binding, meaning a potential future administration could modify or discard them. The administration hopes this step will inspire other nations to adopt a similarly cautious approach to AI, especially amid rising concerns over the technology’s dual-use potential in national security and civil liberties.
TAIPEI -- Delta Electronics, a key supplier for Nvidia's advanced AI chips, announced plans to build its own renewable energy plants in India and Thailand to meet the soaring energy demands of chip manufacturing. "We can't always rely on local governments to solve the issue," Jesse Chou, Delta's vice president and chief sustainability officer, told Nikkei Asia.
"We plan to build our own renewable energy plants to take a more active role to achieve our ultimate goal."
As the global race for artificial intelligence accelerates, the demand for GPUs and advanced chips has surged, placing immense pressure on energy resources. Delta's initiative aims to address the competitive environment between energy sustainability and increased capacity for GPU manufacturing. The company achieved 76% renewable energy usage globally last year but faced challenges in Southeast Asia, with only 41% in that region and 12% in India.
"Extreme weather doesn't have to be an irreversible and inevitable trend, and we, as part of the tech industry, have to do more to counter this,"
Chou emphasised. He added that the emergence of the AI boom is amplifying the challenge of energy supply, especially in Asia's tech economies that heavily rely on fossil fuels.
Delta is also exploring hydrogen fuel cell technology, aiming to start mass production by 2026.
"We have to spend extra efforts to achieve what we have promised. That's our priority,"
Chou stated. The company's proactive approach reflects a broader industry need to ensure sustainable energy solutions while scaling up production to meet AI-driven demand.
The move also has strategic implications amid the ongoing U.S.-China tech wars, as companies like Delta navigate geopolitical risks and supply chain diversification. By investing in renewable energy, Delta not only enhances its operational resilience but also contributes to the global efforts in combating climate change. As Chou noted,
"We have set goals to reach 100% renewable energy use by 2025 for all of our locations around the world, but we require further efforts in these two countries."
China's recent detention of a South Korean engineer on espionage charges underscores an intensifying tech rivalry that reaches beyond mere economic competition, exposing deeper geopolitical risks in the global semiconductor industry. The Financial Times confirmed that the engineer, currently held in China, is the first South Korean national detained under China’s revised anti-espionage laws—legislation that has already unsettled foreign businesses operating in the country.
This arrest not only spotlights China’s increasing focus on national security but also reveals the stakes for South Korea, which hosts the world’s two leading memory chip producers. South Korea now faces the urgent task of defending its technological edge in the highly competitive $158 billion semiconductor sector, where the gap with China is narrowing.
South Korea’s memory chip leaders, Samsung and SK Hynix, have maintained a technological lead, yet China’s aggressive investments and recent policy shifts indicate its determination to bridge that gap. Observers like Jaemin Lee, a trade expert at Seoul National University, predict that as both nations view semiconductors as vital to national security, cases of industrial espionage and security-based detentions may increase.
This climate of suspicion and strategic manoeuvring has prompted South Korea to tighten its own regulations against unauthorised technology transfers, aiming to safeguard its intellectual property amid fears of Chinese encroachment. Such actions reflect a broader trend: semiconductor technology is no longer merely an economic asset but a geopolitical one, as countries position it at the heart of national interests.
However, as the semiconductor industry becomes a battleground for political and security agendas, critical questions arise about the dangers of intertwining technology competition with national security policies. China’s actions, paired with South Korea’s defensive measures, raise concerns about the long-term impacts on innovation and international business cooperation.
When politics permeate tech competition to this extent, do national security justifications risk undermining the very foundations of global tech collaboration? As both nations and their tech industries navigate these uncharted waters, finding a sustainable balance between security concerns and open technological advancement will be essential for stability in the global semiconductor landscape.
The AI chip race is heating up, and the latest financial reports reveal a stark divide between the industry's frontrunners and those lagging behind. Taiwan Semiconductor Manufacturing Company (TSMC) is solidifying its position as a formidable leader, riding the wave of "insane" and enduring demand for AI chips. In contrast, heavyweights like Samsung and ASML are grappling with significant challenges that cast doubt on their ability to maintain a strong foothold in this rapidly evolving market.
Recent earnings paint a mixed picture of the AI boom's impact on chip manufacturers. Dutch semiconductor equipment maker ASML issued a pessimistic sales forecast, stirring concerns about the sustainability of AI-driven chip demand and sending ripples throughout the industry. Samsung, the world's largest memory-chip maker, reported a nearly 13% decline in third-quarter operating profit, missing analyst estimates and falling behind its South Korean rival, SK Hynix. Struggles with its latest AI memory chips have prevented Samsung from fully capitalising on the AI surge, highlighting a significant misstep in an otherwise booming sector.
In sharp contrast, TSMC confirmed the robust and ongoing demand for AI chips, positioning itself to continue reaping substantial benefits. SK Hynix also announced record-breaking quarterly profits, underscoring a widening gap in competitiveness within advanced memory chips. These developments highlight that the AI revolution is not a tide that lifts all boats; it favours those who can innovate and adapt swiftly. Industry leaders like Nvidia continue to dominate the market, while others scramble to keep pace.
Adding another layer to the competitive landscape, Chinese tech giants are creatively navigating U.S. trade restrictions by building a homegrown AI computing ecosystem. Chinese Manufacturers - SenseTime Group are utilising domestic chips from Huawei and Biren Technology—even though these firms are on U.S. trade blacklists. Chinese AI enterprises are also finding innovative ways to develop competitive models more cost-effectively, such as employing smaller datasets and engaging in price wars to reduce inference costs. This agility showcases how early challengers are carving their own paths and reshaping the market dynamics.
The high-stakes arena of AI chip development is becoming increasingly exclusive, favouring those with significant resources and long-term investment capacity. According to Noam Mizrahi, CTO of U.S. chipmaker Marvell, only a select few top developers can remain competitive in this capital-intensive field. While companies like Marvell benefit from cloud service providers investing in custom chips, the overall landscape favours industry leaders with the scale to dominate the market.
Samsung's recent performance exemplifies the challenges faced by established manufacturers. Despite its stature, Samsung has not managed to pass Nvidia's qualification tests for advanced high-bandwidth memory chips. Analysts remain skeptical about Samsung's prospects in the AI chip sector, anticipating major executive shake-ups as the company strives to regain lost ground.
Intel has announced a $300 million investment to expand its chip packaging and testing facility in Chengdu, China, reinforcing its commitment to its largest market despite escalating geopolitical tensions. The expansion will enhance the facility's capabilities and introduce a customer solutions centre to better support Chinese clients. This move comes as Intel faces increased scrutiny from Beijing, including calls from a government-backed cybersecurity group to review its products over security concerns.
The Chengdu plant, operational since 2003, handles over half of Intel's laptop processor packaging and testing. CEO Patrick Gelsinger has emphasised Chengdu's supportive business environment as key to the company's growth strategy in China, which accounted for 27% of Intel's revenue last year. However, Intel is grappling with competitive pressures in a market rapidly shifting toward artificial intelligence applications. The company reported a $1.6 billion loss in the second quarter, a significant downturn from a $1.5 billion profit a year earlier, citing
"weaker spending across consumer and enterprise markets, especially in China."
Intel's investment underscores the delicate balance between strategic ambitions and regulatory challenges amid the intensifying U.S.-China tech rivalry. Despite geopolitical headwinds, economic interdependence continues to drive strategic decisions in the tech industry.
OpenAI proposes bold U.S. alliances to outpace China in AI, advocating for advanced infrastructure and economic zones. Meanwhile, SMIC, China’s chip giant, faces U.S. restrictions but remains optimistic, leveraging AI-driven demand for legacy chips to sustain growth amid global challenges.
Amazon, Microsoft, and Google are turning to nuclear energy for AI data centers. Amazon invested in X-energy, Google partnered with Kairos Power, and Microsoft aims to revive the Three Mile Island plant, highlighting a shift toward nuclear power.
With OpenAI’s shift to a $157 billion for-profit model, CEO Sam Altman maintains its mission to "benefit humanity." However, as investors seek high returns and Altman stands to gain equity, doubts arise over who truly benefits from OpenAI’s growth—society or its shareholders?
Google is investing $1 billion in Thailand to expand AI and cloud infrastructure, while Meta is setting up manufacturing for its Quest 3S in Vietnam. Both moves position Southeast Asia as a key player in the global AI arms race, with tech giants racing to dominate the region’s digital economy.
