ByteDance is reportedly in the process of developing an AI model which could rival Anthropic’s Mythos system as part of a bid to establish itself as a leader in China’s AI ecosystem.

Financial Times (FT) reported the TikTok owner is in the early stages of training a model with as many as 10 trillion parameters.

This would put ByteDance’s offering ahead of Anthropic’s cutting-edge Mythos and make it three-times larger than Moonshot Kimi K3, the biggest model to come out of China so far.

While Anthropic does not reveal the size of its models, FT reported the most advanced Mythos 5 system has 8 trillion parameters and its Fable 5 about 5 trillion.

ByteDance’s model is apparently being pre-trained, a stage in development which typically takes three-to-six-months. It would then be fine-tuned and released, provided there are no issues.

Chinese companies are showing their hand with regards to AI model development, with Moonshot and Alibaba highlighting releases in recent weeks.

FT added ByteDance invested more aggressively in AI than other Chinese companies, building data centres and attracting talent in the sector.

Founder Zhang Yiming reportedly told staff he wants the team to target “world-leading model capabilities” during an internal meeting last month.

The post ByteDance plots leading AI model to rival Mythos appeared first on Mobile World Live.

The European Commission (EC) and the European Space Agency (ESA) named Spain-based Hispasat as the lead contractor for its IRIS2 European satellite network, handing it a budget of more than €1.6 billion.

Owned by Spanish defence company Indra Group, Hispasat has been charged with providing antenna infrastructure, control systems and ground network connectivity for IRIS2.

It will also lead the development of the system’s very low Earth orbit layer, operating at altitudes below 750km. This layer is designed to provide direct connectivity services to mobile devices and for IoT, providing access for European startups and SMEs.

The agreement further guarantees Hispasat access to non-geostationary orbits in regions of high strategic interest, complementing its range of capabilities and geographic coverage.

To that end, Hispasat has committed an investment of up to €600 million in exchange for obtaining “exploitation rights” in regions of high strategic interest and commercial value for the business it already operates through its geostationary fleet.

The award comes following talks between the EC, ESA and SpaceRISE, a satellite consortium formed between Eutelsat, SES and Hispasat.

Luis Mayo, CEO of Hispasat, noted the agreement helps it strengthen its position among the leaders of the European space industry and in its main business areas.

The IRIS2 programme will deploy a network of 330 new LEO satellites and 12 medium earth orbit (MEO) birds, with initial services due to begin in 2030. In total, €15.6 billion will be invested in the project, with €11.6 billion coming from public funds and €4 billion from the SpaceRISE consortium.

The post Hispasat lands lead role in EC IRIS2 project appeared first on Mobile World Live.

China’s cyberspace agency launched an investigation into products sold in the country by Palo Alto Networks, as the government turns up the heat on US suppliers.

In a short statement, the Cyberspace Administration of China (CAC) explained it had launched a cybersecurity review into the company, as part of its bid to ensure safe and stable operation of critical information infrastructure, prevent risks and vulnerabilities, and safeguard national security.

It did not disclose potential penalties for violations or which products it is looking into.

Palo Alto states on its website it has offices in Shanghai, Beijing, Guangzhou, Shenzhen and Macau. It sells network, cloud and other cybersecurity products in the country, but does not break out individual sales figures for the market.

China’s review comes the same week its commerce ministry barred domestic companies from doing business with seven US entities, while beefing up export controls on the US drone market.

Reuters reported Beijing has said the moves are in response to recent US restrictions impacting the Chinese technology sector, including controls put on the nation’s telecoms operators.

The post China takes aim at Palo Alto Networks as US row deepens appeared first on Mobile World Live.

Deutsche Telekom CEO Tim Hoettges declared all areas of the business delivered impressive growth in the first half of 2026, highlighting continued momentum in the US and demand for its MagentaTV service in Germany fuelled by the football World Cup.

An earnings statement for Q2 2026 showed group revenue rose 4.4% year-on-year to €29.9 billion, while net profit was up 11.1% to €2.8 billion. The top line growth was driven by a 5.2% increase in the US to €19.6 billion and a 3.5% uplift in Germany to €6.5 billion. T-Mobile US earnings were reported on 23 July.

The rest of Europe rose 1.5% to €3.2 billion, while Systems Solutions was a sour point with revenue declining 13.4% to €998 million.

In Germany, the second quarter “was dominated by the FIFA World Cup”, with the company broadcasting all of the matches on MagentaTV. It said a million new customers signed up to the TV offering, with a large majority opting for its flex rate plans.

The company also pointed to an increase of 161,000 customers for its fibre offering and 218,000 new contract subscribers in mobile.

In Europe, it added 189,000 mobile contract customers, 56,000 broadband net additions and 16,000 across various TV offerings.

The decline in Systems Solutions was put down to low orders, however it still anticipates an overall rise over the full year.

For the full year, the company increased its guidance for free cash flow after leases from €19.8 billion to around €20 billion.

The post Deutsche Telekom nets in Q2 on US, World Cup growth appeared first on Mobile World Live.

Global revenue from private mobile networks is estimated to grow from $2.3 billion in 2025 to $22 billion in 2030, according to an STL Partners forecast, with the manufacturing industry vertical leading the way in terms of spending.

In a new report, the research company stated the five-year uplift represents a compound annual growth rate (CAGR) of 58%, as the number of private network sites increase from 4,000 by the end of 2025 to more than 40,000 by 2030.

Manufacturing is “by far the largest vertical” and will continue to be so, growing at 59% CAGR on the back of AI-driven automation and Industry 4.0 use cases. Logistics, meanwhile is emerging as a fast grower at a CAGR of 86%, driven by labour shortages and a push into automation.

It also cited growth in utilities and construction, which will move ahead of transport by the end of the decade.

Rosalind Craven, principal analyst at STL, said private networks remain a fast-growing market, but cautioned growth alone will not make success straightforward for vendors and telcos operating in it.

“Enterprises are not buying networks for their own sake; they are solving specific business problems and chasing measurable outcome,” she said. “Private networks can support many of those goals, but success hinges on communicating that value clearly and winning budget priority against competing investments.”

STL said it gathered data from across 15 vertical industry verticals, seven use case groups and nine value chain components across 33 countries for its report.

The post Private network revenue tipped to hit $22B by 2030 appeared first on Mobile World Live.

Samsung Electronics took the wraps off AI memory technology it stated can deliver eight-times the performance of its previous generation, as the vendor presented its vision on how to address growing demands for compatible infrastructure.

Unveiled at the Future of Memory and Storage 2026 event in Santa Clara, California, the company showcased the first concept models of zHBM and zNAND-O, which it stated represent the next-generation of 3D memory architectures.

It stated zHBM is designed for advanced AI computing, presenting a new architecture which vertically stacks high bandwidth memory directly above AI accelerators, “moving beyond conventional designs in which HBM is positioned alongside the processor”.

Samsung explained a next-generation interface system incorporating zHBM is expected to deliver an eightfold performance uplift over HBM5.

With next-generation wafer bonding technology, it can deliver more than 10-times the memory density of HBM5 while enhancing energy efficiency threefold and reducing thermal resistance by more than half.

It also apparently supports customer-specific designs, enabling customised IP to be integrated into “the interlayer between the memory and AI accelerator to expand memory capacity and enhance accelerator performance”.

The vendor positioned zNAND-O as a high-performance set-up combining high efficiency and low latency which is optimised for edge AI environments and capable of handling related real-time, data-intensive applications.

Finally, it showed off what it stated is the industry’s first V10 BV-NAND architecture featuring new wafer bonding with more than 400 layers capable of increasing storage density by 60% compared with previous generations.

$1 trillion market
Samsung is battling with the likes of SK Hynix and Micron to secure a leading position in the memory market and lucrative data centre contracts from major AI players.

In Q2, Samsung registered another all-time high in overall earnings, driven by memory sales.

Counterpoint Research estimates global semiconductor memory revenue will hit $1 trillion this year compared with $240 billion in 2025, due to demand from AI data centres and constraints on supply.

The post Samsung targets memory dominance with latest tech appeared first on Mobile World Live.