Samsung stumbles; Huawei rejects plagiarism claims

The Friday File: Mobile World Live brings you our top three picks of the week as US chip controls dragged Samsung’s financial outlook, Huawei rebuffed claims of AI plagiarism and Capgemini ramped up its agentic AI ambitions with a $3.3 billion acquisition.

US chip controls take toll on Samsung

What happened: In the same week it made a big foldable device splash, Samsung Electronics flagged a sharp downturn in its second quarter operating profit, attributing the slump in part to US export controls on the supply of advanced AI chips to China.

Why it matters: The South Korean vendor projected a 56 per cent year-on-year drop in operating profit and flat revenue, attributed mainly to its Device Solutions unit, which includes Samsung’s memory, system chips and foundry operations, It blamed “inventory value adjustments and the impact of US restrictions on advanced AI chips for China”. However, analysts argue a key factor to the decline is Samsung’s delays in supplying its latest HBM3E chips to major US customer Nvidia. According to Reuters, a few months ago Samsung forecasted that progress on its 12-layer HBM3E chip could come by June, but ultimately offered no timeline for Nvidia supply in its Q2 update. The silence is telling, particularly as rivals SK Hynix and Micron continue gaining momentum in AI-driven memory demand. “The key issue remains regaining competitiveness… Everything ultimately comes back to HBM,” stated Ryu Young-ho, senior analyst at NH Investment and Securities, in a statement to Reuters. Moreover, the potential for new US tariffs could further squeeze Samsung’s margins, particularly across its chip and mobile lines. “It will also be difficult to raise prices immediately due to competition,” Ryu added, underscoring the pressure Samsung now faces.

Huawei denies copying rival AI models

What happened: Huawei’s AI research division reportedly denied accusations that it lifted code from competitors to develop its large language model (LLM), the Pangu Pro Mixture of Experts (MoE) platform.

Why it matters: The controversy kicked off after Huawei published the source code for Pangu Pro MoE on GitHub last week. Later, an online group called “HonestAGI” contained allegations the model included uncredited components from Alibaba’s Qwen 2.5 14B model, Bloomberg reported. Responding in a post on WeChat, Huawei stated the Noah’s Ark Lab, the team behind Pangu Pro, “strictly adheres to the requirement of open source licenses and clearly mark copyright statements in the relevant source files”. While the tech giant didn’t directly name Alibaba in its response, it firmly pushed back against any implications of foul play as they strike at Huawei’s credibility at a pivotal moment in China’s AI journey. Indeed, according to Matthias Bastian, co-founder of publication The Decoder, “the accusations come at a sensitive time for Huawei, which is working to show that its Pangu models – trained on Ascend chips – can compete with global AI leaders”. This is of growing significance “as China seeks alternatives to Nvidia amid increasing US export restrictions”, Bastian explained.

Capgemini ups agentic AI play with $3.3B WNS buy

What happened: Capgemini dialled up its agentic AI push with a $3.3 billion deal to acquire technology outsourcing company WNS in a bid to leverage growing enterprise interest in AI.

Why it matters: The French consultancy stated the move would enable it to deliver more intelligent operations, citing agentic AI’s role in reshaping enterprise operation. Capgemini described WNS as a “springboard” for this evolution due to its expertise in digital business process services (BPS). The company stated the acquisition will serve as a means to better serve enterprise customers looking to overhaul traditional processes with more intelligent AI-powered systems. Indeed, Capgemini CEO Aiman Ezzat argued that BPS is now a “showcase” for agentic AI innovation and said the deal would accelerate its ability to lead the transition. “Capgemini’s acquisition of WNS will provide the group with the scale and vertical sector expertise to capture that rapidly emerging strategic opportunity,” he explained. WNS chief Keshav Murugesh echoed the sentiment, pointing to demand from clients who are “already digitised” and now aim to embed AI deeper into their operations. The deal also deepens Capgemini’s reach in the US market, currently WNS’s largest base, while unlocking cross-sell potential between both companies’ client bases.