How Huawei went from telecommunications to the head of AI “Jack of All Trades” in China

Although he was beaten by years of American commercial restrictions, the Chinese telecommunications giant Huawei has quietly become one of the most ferocious competitors in the whole AI landscape.

Not only does the Shenzhen -based firm seem to represent Beijing’s response to the American darling of the Ai Nvidia chip, but it was also a first adopter of monetization of artificial intelligence models in industrial applications.

“Huawei has been forced to move and expand its commercial objective in the last decade … due to a variety of external pressures on the company,” said Paul Triolo, partner and vice-president of China to the advisory company DGA-ALBRIGHT Stonebridge Group.

This expansion has seen the company get involved in everything, from intelligent cars and operating systems to technologies necessary for AI boom, such as advanced semiconductors, data centers, fleas and models of large languages.

“No other technological enterprise could be competent in many different sectors with high levels of complexity and obstacles to entry,” said Triolo.

This year, the CEO of Nvidia, Jensen Huang, has become more and more expressed by calling Huawei “one of the most formidable technological companies in the world”. He also warned that Huawei would replace Nvidia in China if Washington continues to restrict exports from businesses to American fleas to the Asian country.

Nvidia exceeded 4 billions of dollars in market capitalization last week to become the most precious company in the world. Its cutting -edge processors and a related “CUDA” computer system remain the industry standard for the training of AI models and applications.

But this ditch can shrink, because Huawei proves that not only all of this, it does it well. While defying the American pillars of AI like Nvidia is a major challenge, the history of the company shows why it cannot be counted.

Huawei, who now employs more than 208,000 people on more than 170 markets, came from Humble Beginnings. Founded by the ambitious entrepreneur Ren Zhengfei in 1987 in an apartment in Shenzhen, the company began as a small distributor of telephone switches.

As he became a telecommunications player, he gained ground by targeting less developed markets such as Africa, the Middle East, Russia and South America, before finally expanding in places like Europe.

In 2019, Huawei would be well placed to capitalize on the global deployment of 5G, becoming a market leader. At that time, he also transformed into one of the largest smartphones manufacturers in the world and even conceived smartphones chips via his chip design subsidiary, Histilicon.

But Huawei’s success has also attracted a meticulous examination of governments outside of China, especially the United States, which has frequently accused Huawei technology of posing a national security threat. Chinese society has refuted these risks.

Huawei’s activities underwent a major setback in 2019 when it was placed on an American trade list, preventing US companies from doing business with it.

While the impact of sanctions has started, Huawei’s consumption activities – once the company’s most important company – reduced by about $ 34 billion in 2021 compared to the previous year.

The company has always succeeded in a breakthrough on the AI fleas, and was ahead despite additional American restrictions in 2020 which reduced the company of Chipmaker Taiwan Semiconductor Manufacturing Co. A year earlier, Huawei officially launched its Ascende 910 Processing Chip as part of a strategy to build a “AI compromise AI.

But Huawei’s American targeting also had the effect of transforming the company into a martyr type figure in China, based on the attention it received in 2018 when Meng Wanzhou, Huawei financial director and daughter of Ren, was arrested in Canada for alleged violations of Iranian sanctions.

While the American-Chinese technological war continued to develop and large advanced restrictions of fleas were imposed on China, Huawei was an obvious choice to become a national race champion, with more momentum and state support for its IA plans.

“Export controls ironically pushed Huawei into the arms of the Chinese government in a way that CEO Ren Zhengfei has always resisted,” said Triolo. In this way, the restrictions have also become “steroids” for the hardware and software of Huawei AI.

After another year of falling sales in the consumer segment, the unit began to turn around in 2023 with the release of a smartphone which, according to analysts, contained an advanced chip made in China.

The 5G chip was a shock for many in the United States, which did not expect Huawei to reach this level of progress so quickly without TSMC. Instead, Huawei would have worked with Chinese chip manufacturer Smic, a company that has also been put up black on the United States

While semiconductor analysts said that the scale that Huawei and Smic could produce these chips were seriously limited, Huawei had nevertheless proved that she was back in the advanced chip game.

It was also at this time that the reports began to surface on the new Huawei AI processor chip, Ascend 910b, with the company that seeks to grasp the gaps left by the export controls on the most advanced chips of Nvidia. The mass production of the new generation 910C would already be on the way.

To fill the vacuum left by Nvidia, Huawei “made great progress to reproduce the performance of high -end GPUs using lower flea combinations,” said Jeffrey Towson, director partner at Techmoat Consulting.

In April, Huawei unveiled its “AI Cloudmatrix 384”, a system that connects 384 Ascend 910C fleas in a cluster in data centers. Analysts said Cloudmatrix is able to outdo the NVIDIA system, the GB200 NVL72, on certain measures.

Huawei is not only catching up, “it is a question of redefining the functioning of the infrastructure of the AI,” said Forrester analysts in a report last month about Cloudmatrix.

Meanwhile, Huawei also developed its own “Cann” software system which acts as an alternative to the Cuda de Nvidia.

“Winning the AI race is not only faster tokens. It also includes the delivery of the tools that developers need to build and deploy large -scale models,” said the Forrester report, although the authors have noted that Huawei products are still not integrated with other tools commonly used so that developers quickly pass from Nvidia.

Although Huawei’s objective of overtaking Nvidia is considered a key development in China and the race in the United States for AI, it is important to note that the chips represent only one block of construction of the larger AA of Huawei.

Huawei now has his hands in the value chain of artificial intelligence, computer chips, AI models and AI applications. These different AI commercial routes also operate other areas of the vast technological empire of the company.

In fact, the company’s “ICT infrastructure” activity – which includes a 5.5 g cellular network deployment and industrial use AI systems – has become the largest revenue engine of the company at 362 billion yuan in 2023.

The company has deployed its Ai Ascenda and Ai Cloudmatrix 384 chips in its growing portfolio of AI data centers, which are used by its cloud computing unit, Huawei Cloud, created in 2017 to compete with tastes Amazon Web services and Oracle.

These data centers, in turn, provided the training capacities and the computing power used thereafter of Huawei AI models under its Pangu series.

Unlike other models for general use AI such as GPT-4 of OPENAI or Gemini Ultra 1.0 from Google, Huawei’s Pangu model is designed to take care of more specific apps for industry in the medical, financial, governmental, industrial and automotive sectors. Pangu has already been applied in more than 20 industries in the past year, the company announced last month.

The deployment of such AI applications often implies that Huawei Tech staff work for months on the project site, even if it is in a distant coal mine, Jack Chen, vice-president of the Marketing Department of the Commercial Petroleum, Gas and Mines of Huawei, which provides digital and intelligent solutions to transform these industries, said CNBC.

This research allowed the company in May to deploy more than 100 electric trucks which can independently transport dirt or coal using the 5G network of the telecommunications, AI and Cloud Computing company.

And that is not limited to China. Technology can “be reproduced on a large scale in Central Asia, Latin America, Africa and Asia-Pacific,” said Chen.

Huawei also opened the Pangu models, in a decision which, according to him, would help him develop abroad and continue his “strategy of the Ascend ecosystem”, which refers to its AI products built around its Ascend chips.

Addressing “Squawk Box Asia” from CNBC on Thursday, Patrick Moorhead of Moor Insights & Strategy said that he expected Huawei to pushed Ascend in Pays Part of the Chinese belt and road initiative – an investment and development project aimed at emerging markets.

Over a period of five to 10 years, the company could start to build serious market share in these countries, in the same way as it did once with its telecommunications business, he added.