BREAKING LIVE All breaking →
News

Baidu: the Tech giant ready to challenge Huawei at home and Nvidia abroad, but first it must be face reality

by UCapital MediaUCapital News newsroom2 min read
Baidu: the Tech giant ready to challenge Huawei at home and Nvidia abroad, but first it must be face reality

Amid chip shortages and new restrictions, Baidu is striving to become China’s AI powerhouse with an ambitious roadmap and massive investments. A challenge embedded in the broader technological rivalry between China and the United States, though the obstacles ahead are substantial…


China is accelerating its race for AI chips, and Baidu is emerging as one of its most dynamic players. The tech giant, historically known for its search engine, is transforming itself into a pillar of AI hardware thanks to its subsidiary Kunlunxin, which is developing semiconductors capable of competing with Huawei’s offerings and filling the gap left by Nvidia, now hampered by U.S. export restrictions.


Baidu’s new five-year roadmap - featuring the launch of Kunlun M100 chips in 2026 and M300 chips in 2027 - aims to strengthen an increasingly autonomous computing ecosystem. The company already uses its proprietary chips to power its ERNIE models and provides computing power through its cloud platform, positioning itself as a “full-stack” AI supplier: from chips to servers, data centers, and applications.


Orders from giants like China Mobile and the optimistic forecasts of Deutsche Bank and JPMorgan confirm the trend: domestic demand for AI capacity in China is explosive, and local tech champions are rapidly reducing reliance on U.S. products. According to projections, Baidu’s chip sales could multiply by 2026, while Kunlunxin alone could reach a valuation of 28 billion dollars.


AI chip shortages hit China

Baidu’s push into chips comes at a critical moment for China’s tech sector, which is grappling with an intensifying shortage of semiconductors. Major players - from Alibaba to Tencent - have raised alarms: supply simply cannot keep pace with the soaring demand for AI infrastructure. Alibaba’s CEO, Eddie Wu, has described a “significant obstacle” in the coming years, while Tencent has begun scaling back investments.


Amid trade restrictions and complex geopolitical tensions, companies are attempting to adapt by optimizing their models and drawing down accumulated inventories, while national champion SMIC struggles to keep up with rivals like TSMC. In this context, Baidu sees an opening: with its Kunlun chips, it aims to fill the void left by U.S. GPUs and become a key supplier for a multibillion-dollar domestic market.


As the economic cold war between China and the United States continues, leadership in technology and artificial intelligence hangs in the balance. While China has made major advances in 5G - with a modest footprint in Vietnam - its ambitions in AI currently appear less promising. Baidu, long envisioned as China’s answer to OpenAI, must confront job cuts and financial instability.


Donald Trump’s economic hostility toward the Asian superpower may have softened slightly in recent months, but the fierce battle for technological supremacy is far from over. OpenAI is expected to continue growing over the next five years; the company has stated that it aims to reach 220 million paying ChatGPT users by 2030, securing for the United States a significant advantage in the sector.


Andrea Pelucchi e Klevis Gjoka

BaiduNvidiaHuaweiArtificialIntelligenceMicrochipTech
DISCLOSURE

UCapital Asset Management LLP or group companies may have commercial relationships with the companies mentioned. This content is not financial advice.

✦ Ask the AI about this article

Sign up for free to ask the AI about this article.

The macro picture, read by the AI agents

Growth, inflation and rates across the major economies, with the impact on your business.

Start the macro analysis →

AI use policy. UCapital AI supports the newsroom with news analysis, summaries, market reactions, translations and the generation of cards and charts. AI-generated or AI-assisted content is labelled in the byline (AI Desk, a named AI profile marked «AI», or joint byline) and is reviewed by editors before publication. News received via API feeds from third-party providers is attributed to the original source. AI can make mistakes: report any inaccuracies to the newsroom. Users’ personal data is not used to train the models.