Digital business models Archives | Page 30 of 540

Rising DRAM prices push memory to the centre of AI strategy

The cost of running AI systems is shifting towards memory rather than compute, as the price of DRAM has risen sharply over the past year. Efficient memory orchestration is now becoming a critical factor in keeping inference costs under control, particularly for large-scale deployments.

Analysts such as Doug O’Laughlin and Val Bercovici of Weka note that prompt caching is turning into a complex field.

Anthropic has expanded its caching guidance for Claude, with detailed tiers that determine how long data remains hot and how much can be saved through careful planning. The structure enables significant efficiency gains, though each additional token can displace previously cached content.

The growing complexity reflects a broader shift in AI architecture. Memory is being treated as a valuable and scarce resource, with optimisation required at multiple layers of the stack.

Startups such as Tensormesh are already working on cache optimisation tools, while hyperscalers are examining how best to balance DRAM and high-bandwidth memory across their data centres.

Better orchestration should reduce the number of tokens required for queries, and models are becoming more efficient at processing those tokens. As costs fall, applications that are currently uneconomical may become commercially viable.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

China boosts AI leadership with major model launches ahead of Lunar New Year

Leading Chinese AI developers have unveiled a series of advanced models ahead of the Lunar New Year, strengthening the country’s position in the global AI sector.

Major firms such as Alibaba, ByteDance, and Zhipu AI introduced new systems designed to support more sophisticated agents, faster workflows and broader multimedia understanding.

Industry observers also expect an imminent release from DeepSeek, whose previous model disrupted global markets last year.

Alibaba’s Qwen 3.5 model provides improved multilingual support across text, images and video while enabling rapid AI agent deployment instead of slower generation pipelines.

ByteDance followed up with updates to its Doubao chatbot and the second version of its image-to-video tool, SeeDance, which has drawn copyright concerns from the Motion Picture Association due to the ease with which users can recreate protected material.

Zhipu AI expanded the landscape further with GLM-5, an open-source model built for long-context reasoning, coding tasks, and multi-step planning. The company highlighted the model’s reliance on Huawei hardware as part of China’s efforts to strengthen domestic semiconductor resilience.

Meanwhile, excitement continues to build for DeepSeek’s fourth-generation system, expected to follow the widespread adoption and market turbulence associated with its V3 model.

Authorities across parts of Europe have restricted the use of DeepSeek models in public institutions because of data security and cybersecurity concerns.

Even so, the rapid pace of development in China suggests intensifying competition in the design of agent-focused systems capable of managing complex digital tasks without constant human oversight.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

From Milan-Cortina to factory floors, AI powers Zhejiang manufacturing

As Chinese skater Sun Long stood on the Milan-Cortina Winter Olympics podium, the vivid red of his uniform reflected more than national pride. It also highlighted AI’s expanding role in China’s textile manufacturing.

In Shaoxing, AI-powered image systems calibrate fabric colours in real time. Factory managers say digital printing has lifted pass rates from about 50% to above 90%, easing longstanding production bottlenecks.

Tyre manufacturing firm Zhongce Rubber Group uses AI to generate multiple 3D designs in minutes. Engineers report shorter development cycles and reduced manual input across research and testing.

Electric vehicle maker Zeekr uses AI visual inspection in its 5G-enabled factory. Officials say tyre verification now takes seconds, helping eliminate assembly errors.

Provincial authorities in China report that large industrial firms are fully digitalized. Zhejiang plans to further integrate AI by 2027, expanding smart factories and industrial intelligence.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Study says China AI governance not purely state-driven

New research challenges the view that China’s AI controls are solely the product of authoritarian rule, arguing instead that governance emerges from interaction between the state, private sector and society.

A study by Xuechen Chen of Northeastern University London and Lu Xu of Lancaster University argues that China’s AI governance is not purely top-down. Published in the Computer Law & Security Review, it says safeguards are shaped by regulators, companies and social actors, not only the central government.

Chen calls claims that Beijing’s AI oversight is entirely state-driven a ‘stereotypical narrative’. Although the Cyberspace Administration of China leads regulation, firms such as ByteDance and DeepSeek help shape guardrails through self-regulation and commercial strategy.

China was the first country to introduce rules specific to generative AI. Systems must avoid unlawful or vulgar content, and updated legislation strengthens minor protection, limiting children’s online activity and requiring child-friendly device modes.

Market incentives also reinforce compliance. As Chinese AI firms expand globally, consumer expectations and cultural norms encourage content moderation. The study concludes that governance reflects interaction between state authority, market forces and society.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Hollywood groups challenge ByteDance over Seedance 2.0 copyright concerns

ByteDance is facing scrutiny from Hollywood organisations over its AI video generator Seedance 2.0. Industry groups allege the system uses actors’ likenesses and copyrighted material without permission.

The Motion Picture Association said the tool reflects large-scale unauthorised use of protected works. Chairman Charles Rivkin called on ByteDance to halt what he described as infringing activities that undermine creators’ rights and jobs.

SAG-AFTRA also criticised the platform, citing concerns over the use of members’ voices and images. Screenwriter Rhett Reese warned that rapid AI development could reshape opportunities for creative professionals.

ByteDance acknowledged the concerns and said it would strengthen safeguards to prevent misuse of intellectual property. The company reiterated its commitment to respecting copyright while addressing complaints.

The dispute underscores wider tensions between technological innovation and rights protection as generative AI tools expand. Legal experts say the outcome could influence how AI video systems operate within existing copyright frameworks.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Qwen3.5 debuts with hybrid architecture and expanded multimodal capabilities

Alibaba has released Qwen3.5-397B-A17B, the first open-weight model in its Qwen3.5 series. Designed as a native vision-language system, it contains 397 billion parameters, though only 17 billion are activated per forward pass to improve efficiency.

The model uses a hybrid architecture that combines sparse mixture-of-experts with linear attention via Gated Delta Networks. According to the company, this design improves inference speed while maintaining strong results across reasoning, coding, and agent benchmarks.

Multilingual coverage expands from 119 to 201 languages and dialects, supported by a 250k vocabulary and larger visual-text pretraining datasets. Alibaba says the model achieves performance comparable to significantly larger predecessors.

A hosted version, Qwen3.5-Plus, is available through Alibaba Cloud Model Studio, with a 1-million-token context window and built-in adaptive tool use. Reinforcement learning environments were scaled to prioritise generalisation across tasks rather than narrow optimisation.

Infrastructure upgrades include an FP8 training pipeline and an asynchronous reinforcement learning framework to improve efficiency and stability. Alibaba positions Qwen3.5 as a base for multimodal agents that support reasoning, search, and coding.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

UAE launches first AI clinical platform

A Pakistani American surgeon has launched what is described as the UAE’s first AI clinical intelligence platform across the country’s public healthcare system. The rollout was announced in Dubai in partnership with Emirates Health Services.

Boston Health AI, founded by Dr Adil Haider, introduced the platform known as Amal at a major health expo in Dubai. The system conducts structured medical interviews in Arabic, English and Urdu before consultations, generating summaries for physicians.

The company said the technology aims to reduce documentation burdens and cognitive load on clinicians in the UAE. By organising patient histories and symptoms in advance, Amal is designed to support clinical decision making and improve workflow efficiency in Dubai and other emirates.

Before entering the UAE market, Boston Health AI deployed its platform in Pakistan across more than 50 healthcare facilities. The firm states that over 30,000 patient interactions were recorded in Pakistan, where a local team continues to develop and refine the AI system.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Quebec examines AI debt collection practices

Quebec’s financial regulator has opened a review into how AI tools are being used to collect consumer debt across the province. The Autorité des marchés financiers is examining whether automated systems comply with governance, privacy and fairness standards in Quebec.

Draft guidelines released in 2025 require institutions in Quebec to maintain registries of AI systems, conduct bias testing and ensure human oversight. Public consultations closed in November, with regulators stressing that automation must remain explainable and accountable.

Many debt collection platforms now rely on predictive analytics to tailor the timing, tone and frequency of messages sent to borrowers in Quebec. Regulators are assessing whether such personalisation risks undue pressure or opaque decision making.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Security flaws expose ‘vibe-coding’ AI platform Orchids to easy hacking

BBC technology reporting reveals that Orchids, a popular ‘vibe-coding’ platform designed to let users build applications through simple text prompts and AI-assisted generation, contains serious, unresolved security weaknesses that could let a malicious actor breach accounts and tamper with code or data.

A cybersecurity researcher demonstrated that the platform’s authentication and input handling mechanisms can be exploited, allowing unauthorised access to projects and potentially enabling attackers to insert malicious code or exfiltrate sensitive information.

Because Orchids abstracts conventional coding into natural-language prompts and shared project spaces, the risk surface for such vulnerabilities is larger than in traditional development environments.

The report underscores broader concerns in the AI developer ecosystem: as AI-driven tools lower technical barriers, they also bring new security challenges when platforms rush to innovate without fully addressing fundamental safeguards such as secure authentication, input validation and permission controls.

Experts cited in the article urge industry and regulators to prioritise robust security testing and clear accountability when deploying AI-assisted coding systems.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Five lesser-known SPACs tapping AI, quantum and digital asset innovation

In a recent episode of Ticker Take, financial analysts spotlight five SPACs that fly under the radar but are linked with next-generation tech sectors such as quantum computing, artificial intelligence infrastructure, tokenised assets and genomics/health tech.

The list reflects renewed investor interest in SPACs as an alternative route to public markets for early-stage innovators outside mainstream IPO pipelines.

Crane Harbor Acquisition Corp (CHAC) is targeting Xanadu Quantum Technologies, a Canadian quantum computing company planning to go public via SPAC, aiming to accelerate quantum hardware development.

Churchill Capital Corp X (CCCX) is set to merge with Infleqtion, a firm building quantum computers and precision sensing systems, in an ~$1.8 billion deal.

Cantor Equity Partners II (CEPT) is associated with Securitize, a digital securities platform enabling regulated tokenisation of real-world assets (including potentially AI/tech-linked assets).

Willow Lane Acquisition (WLAC) is linked to Boost Run, an AI-enabled delivery-optimization platform, offering exposure to logistics tech with generative and predictive capabilities.

Perceptive Capital Solutions Corp (PCSC) is connected to Freenome, a company focused on AI-driven early cancer detection and genomics, blending AI with life-science innovation.

Together, these SPAC deals illustrate how blank-check vehicles are resurfacing in markets for AI, quantum and digital transformation, offering investors early access to companies that might otherwise take longer to reach public markets.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!