Thursday, 21 November 2024
26.9 C
Singapore

China’s AI market braces for consolidation amid high competition and US-China tech split

China’s crowded AI market faces consolidation as companies shift to revenue-driven models amid US tech restrictions, aiming to thrive locally.

According to industry analysts, China’s crowded market, with over 100 companies competing, is heading toward a significant shakeout. As the market grows more competitive and costly, companies are expected to shift their focus from basic research to practical, revenue-generating applications. Analysts suggest this pivot will help Chinese firms distinguish themselves from their American counterparts.

Pressure mounts for profitability over research

China’s generative AI industry, heavily centred on large language models (LLMs), has seen rapid expansion but now faces the challenge of saturation. Su Lian Jye, chief analyst at research firm Omdia, spoke at a recent conference about the increasing debates over whether the market can sustain so many players. “We’ve observed concerns this year about market saturation due to the presence of so many similar models,” Su explained.

Developing advanced AI models demands enormous resources, particularly in computing power. Building these models requires vast numbers of graphics processing units (GPUs), making AI development prohibitively expensive. This financial hurdle, Su noted, raises the barrier for companies trying to compete, inevitably pushing smaller players out of the market. “Only a limited number of companies will be able to sustain this costly pursuit,” he added, hinting at a future where only the most prominent players remain.

As a result, China’s AI start-ups are expected to shift their priorities. Instead of dedicating extensive resources to fundamental research, many companies will likely concentrate on creating products that can generate revenue more immediately. By focusing on commercial applications, these firms may carve out a more defined role in the global AI landscape, distinguishing themselves from their US counterparts, who often invest more in foundational research.

A divided ecosystem emerges amid tech restrictions

The ongoing US-China technology rivalry, especially in areas like semiconductor chips and AI, is driving a growing division in the global tech ecosystem. According to Wang Shen, principal analyst for data centres at Omdia, this split will result in two largely independent tech ecosystems: one primarily serving Chinese customers and the other catering to US and international markets. “The two countries will end up with separate ecosystems, each with its supply chains and industries,” Wang explained.

However, Wang sees this development as less of a setback and more as an opportunity for Chinese companies. With a digital market of 1.4 billion people, China has the consumer base to support its domestic AI industry. Wang pointed out that China’s vast, tech-savvy population could sustain numerous companies even as the market consolidates. This immense digital landscape provides a valuable competitive advantage that helps mitigate the effects of technology restrictions imposed by the US.

In response to these restrictions, more than a dozen Chinese AI and GPU companies have emerged to address the country’s surging demand for computing power. This demand has only grown stronger as restrictions from the US limit Chinese companies’ access to certain high-end GPUs, particularly from market leader . Despite these constraints, a robust domestic market allows Chinese firms to focus on local demand, further fuelling the sector’s resilience.

The domestic chip market also faces consolidation

Even with the emergence of local AI and chip start-ups, the market remains crowded and could soon see a similar wave of consolidation. According to Su, homogeneity in the AI chip industry could lead to an adjustment period where the market no longer needs such a wide array of vendors. While Nvidia remains the dominant force in the US and globally, China’s domestic chip start-ups face a challenging path forward, partly due to their products’ similarities.

The recent collapse of GPU start-up Xiangdixian Computing Technology highlights domestic firms’ difficulties in this competitive environment. The Chongqing-based company shut down in September, laying off all employees due to a lack of funding. Xiangdixian’s closure underscores the financial pressures confronting China’s tech start-ups, even as the country pushes for greater self-sufficiency in key technological areas.

Despite this, China’s extensive data centre infrastructure provides a foundation for growth in its AI industry. As local chip companies benefit from the country’s investment in data centres, analysts remain hopeful that the sector will mature and consolidate, leaving a smaller but potentially more resilient group of players.

Hot this week

Honda’s solid-state battery plans could revolutionise EVs

Honda's new facility in Japan accelerates solid-state battery production, promising affordable EVs with longer ranges by the late 2020s.

China’s chip production slows in October as US sanctions loom

China’s chip production slowed in October as potential US sanctions loomed. Legacy chips and EVs drove growth while advanced tech lagged.

Apple’s next AirTag to launch in 2025 with better privacy features

Apple plans to launch a second-gen AirTag in 2025, which will improve privacy with tamper-proof features and enhanced range for better tracking.

FPT and Sitecore launch ON.E, an AI-powered e-commerce accelerator

FPT and Sitecore launch ON.E, an AI-powered e-commerce accelerator aimed at transforming digital commerce for retailers globally.

Glancing becomes a digital trend in Indonesia’s smartphone market

Glance Smart Lock Screen is transforming how over 30 million Indonesians engage with their smartphones, boasting 117 billion glances and 2 billion taps in just nine months.

Elon Musk and Vivek Ramaswamy outline plans for government overhaul

Elon Musk and Vivek Ramaswamy outlined plans to cut federal costs and reduce regulations under Trump’s government efficiency task force.

Honda’s solid-state battery plans could revolutionise EVs

Honda's new facility in Japan accelerates solid-state battery production, promising affordable EVs with longer ranges by the late 2020s.

NVIDIA expands DLSS 3 support to over 600 games, including Stalker 2

NVIDIA expands DLSS 3 support to over 600 games, including Stalker 2 and Flight Simulator 2024, with improved visuals and performance.

ASUS-built supercomputer with NVIDIA HGX H100 ranked among the world’s top supercomputers

ASUS and Ubilink build a supercomputing facility ranked 31st on TOP500 and 44th on Green500, delivering 45.82 PFLOPS and unmatched efficiency.

Related Articles

Popular Categories