Category: 3. Business

‘Chaotic’ Liverpool care home’s residents found ‘wet through’
David HumphreysLocal Democracy Reporting Service

Google

A spokesperson for Rowan Garth Care Home said it had appointed a “turnaround manager to lead improvements”

Incontinent residents were “wet through” and in need of a change of clothes at a “chaotic” care home where checks for a deadly disease were also not always completed, inspectors have found.

The Care Quality Commission (CQC) has put Rowan Garth Care Home in the Anfield area of Liverpool back into special measures following an inspection in June.

The care regulator, which previously put the home under special measures in November 2022, found “serious failings” this summer, despite a plan having been made last year to improve conditions. The CQC said “no effective action had been taken”.

Wellington Healthcare Ltd, which runs the site, said it had taken “immediate action”.

The care home on Lower Breck Road provides accommodation for older people requiring nursing or personal care.

At the time of inspection, only three of its five units were in operation.

There were 82 people living there.

The CQC downgraded the home’s overall rating from “requires improvement” to “inadequate”.

Assessments for being “safe, effective and well-led” were rated as “inadequate” while the “responsive and caring” assessment criteria were rated as “requires improvement”.

PA Media

The home was described as “very chaotic” by staff, the CQC said
The findings in the CQC report include:

the management of medicines was unsafe, with residents not receiving them at the right time, exposing people to “often painful or uncomfortable symptoms”

staff did not have sufficient clinical guidance on people’s clinical risks and medicines for complex health conditions such as diabetes and epilepsy

Medicines were not always stored at the right temperatures, increasing the risk of them being ineffective

Some people’s continence records showed they were wet through and in need of a change of clothes on multiple occasions, indicating people’s continence care was insufficient

Some residents did not have access to a working bath or an accessible call bell to ring for staff support when needed

Infection control standards were “poor”

Furnishing, fittings and equipment was not always clean or in a good state or repair

Checks to monitor for the risk of Legionella bacteria in the home’s water system were not always completed.

An agency nurse told inspectors how, from their point of view, there were not enough staff and it was “too much”.

The home was described as “very chaotic” with staff described as “knackered”.
‘Expect rapid improvements’

Andrew Peck, of the CQC, said inspectors found “serious failings in leadership that placed people at unnecessary risk of harm”.

He said some residents received time-critical medications hours late which was “especially serious for people with conditions like Parkinson’s disease, where timing is vital”.

He added: “Leaders didn’t ensure the environment was safe and we saw broken equipment and inadequate facilities.

“The call bell system wasn’t fit for purpose and although the provider had been aware of this for over six months, no effective action had been taken to ensure people were able to call for staff help when needed.”

He said: “While we found staff were kind and caring, they weren’t supported by leaders to deliver safe care.

“Leaders also didn’t ensure staffing levels were sufficient, meaning people often experienced delays in receiving support.”

Mr Peck said the regulator expected to see “rapid and continued improvements” and would continue to monitor the home closely to keep people safe.

“We have begun the process of taking regulatory action in order to protect people further.”

Residents’ safety ‘paramount’

A spokesperson for the care home said it was “disappointed” with the “inadequate” CQC rating and said “our priority is to learn from this and take immediate corrective action”.

They said it had “implemented a comprehensive improvement plan to address all concerns raised”.

“We acknowledge there were areas where we did not meet the high standards our residents and their families rightfully expect and deserve.

“The safety and wellbeing of our residents is paramount. We have appointed a highly experienced turnaround manager to lead the improvements at Rowan Garth and ensure sustainable change.

“We remain committed to delivering the quality of care our residents deserve and look forward to demonstrating significant progress at the CQC’s next inspection.”
Continue Reading
December 11, 2025
Slower buses causing passenger number fall, London assembly told

Kumail JafferLocal Democracy Reporting Service

Getty Images

Westminster and Camden have the capital’s slowest bus speeds

Bus speeds in London have slowed to their lowest level in years, causing a fall in passenger numbers, the London Assembly has heard.

Average speeds on the capital’s bus network fell to 9.17mph in 2024–25, down from 10.27mph four years earlier, according to City Hall data. In August, the latest month available, buses were travelling at 9.06mph on average.

Passenger numbers also fell last year for the first time since the pandemic, dropping from 1.869bn journeys to 1.842bn.

Transport for London (TfL) said its Bus Action Plan would speed up travel, with 15.5 miles (25km) of new bus lanes, 1,900 signals prioritising buses and 52.8 miles (85km) of existing lanes operating 24 hours a day.

The assembly’s transport committee was told this week slower services and “endless traffic” were making buses less attractive.

Paul Lynch, managing director of Stagecoach London, said conditions had “worsened over the last few years to a point where somebody who works for me… and has been around for 40 years operating buses in London says it’s the worst he has ever seen”.

He added: “It’s making them less attractive and less reliable… It’s got to be one of the reasons why bus passenger numbers are declining at the same time that bus speeds are.”

TfL’s latest Travel in London report recorded a 1.5% fall in bus journeys compared with last year, alongside rises in passenger numbers on the Underground and Elizabeth line.

‘Bad for London’

Michael Roberts, chief executive of London TravelWatch, told members that slower journey times “mean reduced patronage, which in turn means reduced income to TfL”.

He said slower speeds also increased operating costs because “you need more buses to run a given level of service”, adding that buses are “an effective use of road space” and declining use was “bad for London”.

“For every 10% reduction in journey speeds, there’s a 6% reduction in demand,” he said.

London TravelWatch estimates that meeting the mayor’s aim for 80% of trips to be made by walking, cycling or public transport by 2041 would require bus journeys to rise by 40%

TfL analysis suggests daily trips must grow from 5.1m to 9m.

Some boroughs experience far slower services than others, with average speeds under 7mph in the City of London, Camden and Westminster.

Bexley, Hillingdon and Havering recorded average speeds above 11mph.

Continue Reading

December 11, 2025
Exclusive: China's ZTE may pay more than $1 billion to the US over foreign bribery allegations, sources say – Reuters
1. Exclusive: China’s ZTE may pay more than $1 billion to the US over foreign bribery allegations, sources say Reuters
2. ZTE May Pay US Govt USD1B+ to Settle Overseas Bribery Allegations: Report AASTOCKS.com
3. Justice Department has moved ahead with a U.S. investigation into ZTE for allegedly violating Foreign Corrupt Practices Act in South America and other regions -sources marketscreener.com
4. ZTE Corp shares slide on report of over $1 bln fine to US govt Investing.com
5. ZTE Communicating with US Department of Justice Over Compliance Probe Related to Foreign Corrupt Practices Act, Will Protect Rights Through Legal Means AASTOCKS.com
Continue Reading
December 11, 2025
Graph-augmented transformer ensemble framework for robust and scalable fake news detection in social media ecosystems
Almandouh, M., Alrahmawy, M. F., Eisa, M., Elhoseny, M. & Tolba, A. S. Ensemble based high performance deep learning models for fake news detection. Sci. Rep. 14, 26591. https://doi.org/10.1038/s41598-024-26591-5 (2024).

Google Scholar
Praseed, A., Rodrigues, J. & Santhi, T. P. Disinformation detection using graph neural networks: A survey. Artif. Intell. Rev. 57 (2024).
Liu, J., Wu, F., Jin, H., Zhu, X. & Jing, X. Y. Inter-modal fusion network with graph structure preserving for fake news detection. In: Neural Information Processing. Springer, 1–9 https://doi.org/10.1007/978-981-99-8076-5_20 (2024).
Jiang, Y. & Liu, X. Deep learning for fake news detection: A survey. Artif. Intell. 303, 101235. https://doi.org/10.1016/j.artint.2022.101235 (2022).

Google Scholar
Harris, S., Hadi, H. J., Ahmad, N. & Alshara, M. A. Fake news detection revisited: an extensive review of theoretical frameworks, dataset assessments, model constraints, and forward-looking research agendas. Technologies 12, 222. https://doi.org/10.3390/technologies12110222 (2024).

Google Scholar
Wang, B., Feng, Y., Xiong, X. C., Wang, Y. H. & Qiang, B. H. Multi-modal transformer using two-level visual features for fake news detection. Appl. Intell. 53, 10429–10443. https://doi.org/10.1007/s10489-022-04055-5 (2022).

Google Scholar
Lee, D. & Kim, S. Detecting fake news using social media and graph neural networks. J. Comput. Social Sci. 5, 125–137. https://doi.org/10.1007/s42001-021-00110-3 (2022).

Google Scholar
Song, C., Teng, Y., Zhu, Y., Wei, S. & Bin, B. Dynamic graph neural network for fake news detection. Neurocomputing 505, 362–374. https://doi.org/10.1016/j.neucom.2022.07.013 (2022).

Google Scholar
Xu, W., Wang, X. & Chen, Q. Leveraging attention-based networks for fake news detection in social media. IEEE Access. 10, 47115–47127. https://doi.org/10.1109/ACCESS.2022.3181273 (2022).

Google Scholar
Roumeliotis, K. I., Tselikas, N. D. & Nasiopoulos, D. K. Fake news detection and classification: A comparative study of convolutional neural networks, large Language models, and natural Language processing models. Future Internet. 17, 28. https://doi.org/10.3390/fi17010028 (2025).

Google Scholar
Folino, F., Folino, G., Guarascio, M. & Tagarelli, A. Towards data- and compute-efficient fake news detection: an approach combining active learning and pre-trained Language models. SN Comput. Sci. 5, 470. https://doi.org/10.1007/s42979-024-02809-1 (2024).

Google Scholar
Sudhakar, M. & Kaliyamurthie, K. P. Efficient prediction of fake news using novel ensemble technique based on machine learning algorithm. In: Information and Communication Technology for Competitive Strategies. Springer, 1–10 https://doi.org/10.1007/978-981-19-0098-3_1 (2023).
Luo, P. & Xie, Y. Fake news detection via multi-task learning on graph networks. Inf. Sci. 613, 533–550. https://doi.org/10.1016/j.ins.2022.09.040 (2023).

Google Scholar
Xu, X., Sun, C. & Wang, S. Fake news detection via multi-view graph convolutional network. ACM Trans. Inform. Syst. 41, 47. https://doi.org/10.1145/3458986.3458991 (2023).

Google Scholar
Zhang, X. & Zhao, Q. Fake news detection using graph neural networks: A comprehensive survey. Neural Comput. Appl. https://doi.org/10.1007/s00542-023-08257-w (2023).

Google Scholar
Rani, S. & Kumar, M. Multi-modal topic modeling from social media data using deep transfer learning. Appl. Soft Comput. 160, 111706. https://doi.org/10.1016/j.asoc.2024.111706 (2024).

Google Scholar
Kumar, Y., Bhardwaj, P., Shrivastav, S. & Mehta, K. Predicting paediatric brain disorders from MRI images using advanced deep learning techniques. Neuroinformatics 23 (2), 9. https://doi.org/10.1007/s12021-024-09707-0 (2025).

Google Scholar
Modi, N. et al. Physiological signal-based mental stress detection using hybrid deep learning models. Discover Artif. Intell. 5, 166. https://doi.org/10.1007/s44163-025-00412-8 (2025).

Google Scholar
Alzahrani, M. A. & Aljuhani, M. A. Enhancing fake news detection with word embedding: A machine learning and deep learning approach. Computers 13, 239. https://doi.org/10.3390/computers13090239 (2024).

Google Scholar
Zamani, A. S., Hashim, A. H. A., Mohamed, S. S. I. & Alam, M. N. Optimized deep learning techniques to identify rumors and fake news in online social networks. J. Comput. Cogn. Eng. 2, 1–12. https://doi.org/10.47852/bonviewJCCE52023348 (2023).

Google Scholar
Wei, L. & Zhang, T. Fake news detection using deep learning-based fusion of graph convolutional networks and transformer models. Inform. Fusion. 81, 146–159. https://doi.org/10.1016/j.inffus.2022.11.004 (2023).

Google Scholar
Abduljaleel, I. Q. & Ali, I. H. Deep learning and fusion mechanism-based multimodal fake news detection methodologies: A review. Eng. Technol. Appl. Sci. Res. 14, 15665–15675. https://doi.org/10.48084/etasr.7907 (2024).

Google Scholar
Li, H., Liu, L. & Wang, Y. Fake news detection using a transformer-based framework with attention mechanisms. Expert Syst. Appl. 182, 115129. https://doi.org/10.1016/j.eswa.2021.115129 (2022).

Google Scholar
Jing, J., Wu, H., Sun, J., Fang, X. & Zhang, H. Multimodal fake news detection via progressive fusion networks. Inf. Process. Manag. 60, 103120. https://doi.org/10.1016/j.ipm.2022.103120 (2023).

Google Scholar
Dixit, D. K., Bhagat, A. & Dangi, D. An accurate fake news detection approach based on a levy flight honey Badger optimized convolutional neural network model. Concurrency Computation: Pract. Experience. 35, e7382. https://doi.org/10.1002/cpe.7382 (2023).

Google Scholar
Yang, L. & Lee, K. Fake news detection in social media using a hybrid model of deep neural networks. Neural Netw. 145, 202–213. https://doi.org/10.1016/j.neunet.2021.09.003 (2022).

Google Scholar
Singhania, S., Fernandez, N. & Rao, S. 3HAN: A deep neural network for fake news detection. ArXiv Preprint (2023). ArXiv:2306.12014
Kikon, J. M. & Bania, R. K. Towards development of machine learning models for fake news detection and sentiment analysis. In: Proceedings of the NIELIT’s International Conference on Communication, Electronics and Digital Technology (NICEDT 2024). pp. 99–107 https://doi.org/10.1007/978-981-97-3601-0_8 (Springer, Singapore, 2024).
Fu, X., Guo, C. & Yuan, Z. A survey of fake news detection in social media: Methods, trends, and challenges. Computers 12, 80. https://doi.org/10.3390/computers12050080 (2023).

Google Scholar
Patel, R. & Gupta, P. Fake news detection via text and graph-based fusion learning. Comput. Intell. Neurosci. https://doi.org/10.1155/2022/6574754 (2022).

Google Scholar
Truică, C. O., Apostol, E. S., Marogel, M., Paschke, A. & GETAE Graph information enhanced deep neural network ensemble architecture for fake news detection. Expert Syst. Appl. 275, 126984. https://doi.org/10.1016/j.eswa.2025.126984 (2025).

Google Scholar
Zhang, L. & Chen, X. An improved method for fake news detection using attention-based neural networks. Appl. Soft Comput. 115, 108273. https://doi.org/10.1016/j.asoc.2021.108273 (2022).

Google Scholar
Papageorgiou, E., Varlamis, I. & Chronis, C. Harnessing large Language models and deep neural networks for fake news detection. Information 16, 297. https://doi.org/10.3390/info16040297 (2025).

Google Scholar
Jin, W. et al. Veracity-oriented context-aware large Language models–based prompting optimization for fake news detection. Int. J. Intell. Syst. 40, 5920142. https://doi.org/10.1002/int.5920142 (2025).

Google Scholar
Jin, W. et al. -FND: A multi-role fake news detection method based on argument switching-based courtroom debate. J. King Saud Univ. – Comput. Inform. Sci. 37, 33. https://doi.org/10.1016/j.jksuci.2024.101033 (2025).

Google Scholar
Khattar, D., Goud, J. S., Gupta, M. & Mvae, V. V. Multimodal variational autoencoder for fake news detection. In: The World Wide Web Conference, pp. 2915–2921 (2019).
Jin, W. et al. A veracity dissemination consistency-based few-shot fake news detection framework by synergizing adversarial and contrastive self-supervised learning. Sci. Rep. 14, 19470. https://doi.org/10.1038/s41598-024-19470-0 (2024).

Google Scholar
Jin, W. et al. A prompting multi-task learning-based veracity dissemination consistency reasoning augmentation for few-shot fake news detection. Eng. Appl. Artif. Intell. 144, 110122. https://doi.org/10.1016/j.engappai.2025.110122 (2025).

Google Scholar
Abulaish, M., Kamal, A. & Zaki, M. J. A survey of figurative Language and its computational detection in online social networks. ACM Trans. Web. 14, 1–52. https://doi.org/10.1145/3383212 (2020).

Google Scholar
Kamal, A. & Abulaish, M. Contextualized satire detection in short texts using deep learning techniques. J. Web Eng. 23, 27–52 (2024).

Google Scholar
Abulaish, M. & Kamal, A. Self-deprecating sarcasm detection: An amalgamation of rule-based and machine learning approach. In: 2018 IEEE/WIC/ACM International Conference on Web Intelligence (WI). pp. 415–422 https://doi.org/10.1109/WI.2018.00-25 (IEEE, 2018).
Kamal, A., Anwar, T., Sejwal, V. K., Fazil, M. & BiCapsHate Attention to the linguistic context of hate via bidirectional capsules and hatebase. IEEE Trans. Comput. Social Syst. 11, 1781–1792. https://doi.org/10.1109/TCSS.2022.3228775 (2023).

Google Scholar
Mohankumar, P., Kamal, A., Singh, V. K. & Satish, A. Financial fake news detection via context-aware embedding and sequential representation using cross-joint networks. In: 2023 15th International Conference on Communication Systems & Networks (COMSNETS), pp. 780–784 https://doi.org/10.1109/COMSNETS56262.2023.10041125 (IEEE, 2023).
Kamal, A., Mohankumar, P. & Singh, V. K. Financial misinformation detection via RoBERTa and multi-channel networks. In: International Conference on Pattern Recognition and Machine Intelligence. pp. 646–653 https://doi.org/10.1007/978-3-031-40375-8_53 (Springer, Cham, 2023).
Ilie, V. I., Truică, C. O., Apostol, E. S. & Paschke, A. Context-aware misinformation detection: A benchmark of deep learning architectures using word embeddings. IEEE Access. 9, 162122–162146. https://doi.org/10.1109/ACCESS.2021.3132502 (2021).

Google Scholar
Truică, C. O., Apostol, E. S. & MisRoBÆRTa Transformers versus misinformation. Mathematics 10, 569. https://doi.org/10.3390/math10040569 (2022).

Google Scholar
Truică, C. O. & Apostol, E. S. Fake news detection using sentence Transformers. CEUR Workshop Proc. 3180, paper–61 (2022). https://ceur-ws.org/Vol-3180/paper-61.pdf

Google Scholar
Truică, C. O. & Apostol, E. S. It’s all in the embedding! Fake news detection using document embeddings. Mathematics 11, 508. https://doi.org/10.3390/math11030508 (2023).

Google Scholar
Petrescu, A., Truică, C. O. & Apostol, E. S. Language-based mixture of Transformers for EXIST2024. CEUR Workshop Proc. 3740, paper–108 (2024). https://ceur-ws.org/Vol-3740/paper-108.pdf

Google Scholar
Truică, C. O., Apostol, E. S. & Karras, P. Deep neural network ensemble architecture for social and textual context-aware fake news detection. Knowl. Based Syst. 294, 111715. https://doi.org/10.1016/j.knosys.2024.111715 (2024).

Google Scholar
E Almandouh M, Alrahmawy MF, Eisa M, Elhoseny M, Tolba AS. Ensemble based highperformance deep learning models for fake news detection. Scientific Reports 14 (1), 26591 (2024).
Ghosh, A. et al. Proactive network immunization for misinformation control. In: Proceedings of the ACM Conference on Computer and Communications Security (CCS), pp. xxx–xxx https://doi.org/10.1145/3459637.3482481 (2021).
Truică, C. O., Apostol, E. S., Nicolescu, R. C. & Karras, P. M. C. W. D. S. T. A minimum-cost weighted directed spanning tree algorithm for real-time fake news mitigation in social media. IEEE Access. 11, 125861–125873. https://doi.org/10.1109/ACCESS.2023.3331220 (2023).

Google Scholar
Apostol, E. S., Coban, Ö. & Truică, C. O. CONTAIN: A community-based algorithm for network immunization. Eng. Sci. Technol. Int. J. 55, 101728. https://doi.org/10.1016/j.jestch.2024.101728 (2024).

Google Scholar
Apostol, E. S., Truică, C. O., Paschke, A. & ContCommRTD A distributed content-based misinformation-aware community detection system for real-time disaster reporting. IEEE Trans. Knowl. Data Eng. 36, 5811–5822. https://doi.org/10.1109/TKDE.2024.3417232 (2024).

Google Scholar
Truică, C. O., Constantinescu, A. T. & Apostol, E. S. StopHC: A harmful content detection and mitigation architecture for social media platforms. In: Proceedings of the IEEE International Conference on Intelligent Computer Communication and Processing (ICCP), pp. xxx–xxx https://doi.org/10.1109/ICCP63557.2024.10793051 (2024).
Almeida, F. et al. Virality detection in social media. In: Proceedings of the International Conference on Extending Database Technology (EDBT), pp. xxx–xxx https://doi.org/10.5441/002/edbt.2021.69 (2021).

Continue Reading

Why Big Tech is doubling down on investing in India

A slogan related to Artificial Intelligence (AI) is displayed on a screen in Intel pavilion, during the 54th annual meeting of the World Economic Forum in Davos, Switzerland, January 16, 2024.

Denis Balibouse | Reuters

Big Tech is doubling down on investing billions in India, drawn by its abundance of resources for building data centers, a large talent and digital user pool, and market opportunity.

In under 24 hours, Microsoft and Amazon pledged more than $50 billion toward India’s cloud and AI infrastructure, while Intel on Monday announced plans to make chips in the country to capitalize on its growing PC demand and speedy AI adoption.

While India trails the U.S. and China in the race to develop a native AI foundational model, and lacks a large domestic AI infrastructure company, it wants to leverage its expertise in the information technology sector to create and deploy AI applications at enterprise level, also offering Big Tech companies a huge opportunity.

Having a model or computing is not enough for any enterprise to use AI effectively, and it requires companies making application layer and a large talent pool to deploy them, S. Krishnan, secretary at India’s Ministry of Electronics and Information Technology, told CNBC.

Stanford University ranks India among the top four countries along with the U.S., China and the UK in the global and national AI vibrancy ranking. GitHub, a community of developers, has ranked India at the top with the global share of 24% of all projects.

India’s opportunity lies more in “developing applications” which will be used to drive revenues for AI companies, Krishnan said.

On Tuesday, Microsoft announced $17.5 billion in investment in the country, spread over 4 years, aimed at expanding hyperscale infrastructure, embedding AI into national platforms, and advancing workforce readiness.

“This scale of capex gives Microsoft first‑mover advantage in GPU‑rich data centers while making Azure the preferred platform for India’s AI workloads, as well as deepening alignment with the government’s AI public infrastructure push,” said Tarun Pathak, research Director at Counterpoint Research.

Amazon on Wednesday announced plans to invest over $35 billion, on top of the $40 billion it has already invested in the country.

Over the past few months, AI and tech majors such as OpenAI, Google, and Perplexity have offered their tools for free to millions in India, with Google also firming up its plans to invest $15 billion toward building data center capacity for a new AI hub in southern India.

“India combines a huge digital user base, rapidly growing cloud and AI demand, and a high-talent IT ecosystem that can build and consume AI at scale, making it more than just a market for users and instead a core engineering and deployment hub,” Pathak said.

Data center opportunity

India has several advantages when it comes to building data centers. Markets such as Japan, Australia, China and Singapore in the Asia Pacific region have matured. Singapore, one of the oldest data center hubs in the region, has limited room to deploy large-scale data centers due to land availability issues.

India has abundant space for large-scale data center developments. When compared with data center hubs in Europe, power costs in India are relatively low. Coupled with India’s growing renewable energy capacity — critical for power-hungry data centers — and the economics begin to look compelling.

Local demand, fueled by the rise of e-commerce — a major driver of data center growth in recent years — and potential new rules for storing social media data, strengthens the case.

Put simply: India is entering a sweet spot where global cloud providers, AI players, and domestic digitalization all converge to create one of the world’s hottest data center markets.

“India is a pivotal market and one of the fastest‑growing regions for AI spending in Asia Pacific,” said Deepika Giri, associate vice president and head of research, big data & AI, at International Data Corporation.

“A major gap, and therefore a significant opportunity, lies in the shortage of suitable compute infrastructure for running AI models,” she added. Big Tech is looking to capitalize on the infrastructure opportunity in India by investing heavily in the cloud and data center space.

Global companies are expanding capacities closer to service bases in IT cities such as Bangalore, Hyderabad and Pune from traditional centers like Mumbai and Chennai which are closer to landing cables, as they build data centers in India for the world, Krishnan said.

— CNBC’s Dylan Butts, Amitoj Singh contributed to this report.

Continue Reading

Stocks Slide as Oracle Spoils Mood After Fed Cut: Markets Wrap

This content was published on

December 11, 2025 – 06:17

(Bloomberg) — A global equities rally spurred by the Federal Reserve’s interest-rate cut evaporated as disappointing results from Oracle Corp. weighed on tech shares and focus turned to the US central bank’s outlook for further easing next year.

Futures on Nasdaq 100 slumped more than 1.5%, while a selloff in technology stocks in Asia caused the regional equity gauge to reverse earlier gains. S&P 500 futures were down 0.8%. Shares of Oracle, whose fate is deeply tied to the artificial intelligence boom, plunged more than 10% in extended US trading after second-quarter cloud sales fell just short of analysts’ estimates. In a sign of waning risk appetite, Bitcoin lost more than 2%.

The moves came after the S&P 500 rose 0.7% on Wednesday, ending just short of all-time highs, as the Fed cut rates for a third consecutive time and Chair Jerome Powell voiced optimism that the economy will strengthen as the inflationary impact from tariffs fades away. The result marked the first time since 2019 that three officials voted against a policy decision, with dissents on both ends of the policy spectrum. Officials maintained their outlook for a single cut in 2026.

“While most of the focus was on the FOMC, a key risk for markets overnight was Oracle,” said Billy Leung, investment strategist at Global X Management. Its result was a key test for the AI infrastructure trade given Oracle’s role as a bellwether for hyperscale data center spending and has added to broader tech pressure, he said.

Oracle reported a jump in spending on AI data centers and other equipment, rising outlays that are taking longer to translate into cloud revenue than investors wanted. Powered by an AI-driven rally, MSCI Inc.’s index of global stocks has surged about 20% this year and is on course for its best annual advance since 2019.

A gauge of Asian technology stocks was down more than 1%, versus a 0.5% decline in the broader regional benchmark. Shares of SoftBank Group lost more than 8% in Tokyo.

Traders in Asia were also assessing the impact of Mexican lawmakers’ final approval for new tariffs on the region’s imports. Also in focus was Thursday’s interest-rate decision in the Philippines, where the central bank is predicted to cut its key interest rate for a fifth straight meeting.

In Japan, bonds gained after an auction of 20-year government debt drew its strongest demand since 2020. Yields across the curve have recently climbed to multi-year highs on renewed fiscal concerns as well as rising expectations for a Bank of Japan rate hike at its meeting next week.

Elsewhere in markets, a gauge of the dollar edged higher after falling 0.4% on Wednesday. In commodities, oil prices were in focus after the US seized a sanctioned tanker off Venezuela, deterring more shipments from the South American producer and raising the risk of a conflict.

‘Fairly Choppy’

US Treasuries rallied on Wednesday, with the policy-sensitive two-year yield sliding eight basis points, as the Fed’s quarter-point rate reduction was accompanied by the authorization of fresh Treasury bill purchases to rebuild bank reserves. The yield fell one more basis point on Thursday.

The 10-year yield, which dropped around four basis points in the previous session, was two basis points lower on Thursday. The declines have stalled a prior run up in yields that had driven one global gauge to its highest since 2009.

Powell pushed through the quarter percentage point cut not only over the objection of a few voters. A much larger group of regional Fed bank presidents who participated in the debate but weren’t among this year’s voting roster also signaled they opposed the cut.

The fractures could foreshadow what’s to come in 2026, when a new chair may struggle even more than Powell to marshal consensus at the Fed.

The Fed chair suggested the central bank had now acted sufficiently to help stabilize the labor market while leaving rates high enough to continue weighing on price pressures. He also underscored the importance of upcoming economic reports while advising caution on assessing household jobs readouts, given technical distortions after a government shutdown caused a data blackout.

Nick Twidale, chief analyst at AT Global Markets in Sydney, said he is “hesitant” on how much momentum the Fed’s cut will bring to global markets.

“The forward guidance was probably less dovish than most investors were hoping for,” he said. “We may see some fairly choppy markets in the sessions ahead as the market digests what Jerome Powell had to say.”

Corporate News

Private equity firm TPG Inc. is considering options for APM Monaco, including a possible stake sale or an initial public offering of the jeweler, according to people familiar with the matter. SK Hynix Inc. fell after South Korea’s main bourse issued a higher-level warning on investing in the stock following strong gains sparked by expectations of a listing in New York. Investment banks are on track to take home their smallest slice of underwriting fees from Hong Kong listings in years, even as share sales in the city have staged a blistering rebound. President Donald Trump signaled he’ll oppose a Warner Bros. Discovery Inc. deal that doesn’t include new ownership of CNN, a potential wrinkle for the bid from Netflix Inc. Shares of Jingdong Industrials Inc., the supply-chain unit of Chinese e-commerce giant JD.com Inc., fell in their Hong Kong trading debut after a HK$2.98 billion ($383 million) initial public offering. Japan’s stock market is witnessing a record wave of large private transactions known as block trades, stemming from companies reducing cross-shareholdings to improve corporate governance. Chinese artificial intelligence startup DeepSeek has relied on Nvidia Corp. chips that are banned in the country to develop an upcoming AI model, according to a new report in The Information. Coca-Cola Co. said Chief Executive Officer James Quincey is stepping down and will be replaced at the end of March by Henrique Braun, the company’s chief operating officer. Some of the main moves in markets:

Stocks

S&P 500 futures fell 0.8% as of 2:16 p.m. Tokyo time Japan’s Topix fell 0.7% Australia’s S&P/ASX 200 rose 0.1% Hong Kong’s Hang Seng was little changed The Shanghai Composite fell 0.6% Euro Stoxx 50 futures fell 0.1% Currencies

The Bloomberg Dollar Spot Index rose 0.1% The euro was little changed at $1.1690 The Japanese yen was little changed at 155.91 per dollar The offshore yuan was little changed at 7.0611 per dollar The Australian dollar fell 0.6% to $0.6637 Cryptocurrencies

Bitcoin fell 2.4% to $90,140.14 Ether fell 4.3% to $3,197.85 Bonds

The yield on 10-year Treasuries declined two basis points to 4.12% Japan’s 10-year yield declined three basis points to 1.925% Australia’s 10-year yield declined nine basis points to 4.72% Commodities

West Texas Intermediate crude was little changed Spot gold fell 0.4% to $4,212.65 an ounce This story was produced with the assistance of Bloomberg Automation.

–With assistance from Richard Henderson.

Continue Reading

Starling eyes UK acquisition to boost corporate lending

Stay informed with free updates

Starling Bank is exploring plans to buy another UK lender, an overhaul to its business model that would allow the digital bank to put its £12bn worth of customer deposits to more profitable use.

The fintech is actively looking at several acquisition options that would allow it to expand its lending capacity and deploy deposits into areas that could generate higher returns, such as corporate lending, according to two people familiar with the discussions. One said the acquisition could be “substantial” in scale.

The acquisition is part of an effort by the fintech to diversify its income. Currently, Starling customer deposits are either held at the Bank of England collecting interest, invested in securities such as bonds, or channelled into mortgage lending through Fleet, a small lender that Starling bought in 2021.

John Cronin, an independent bank analyst, said Starling “has struggled to grow its loan book commensurate with the size of its deposit book”. He said Starling’s latest set of accounts showed that £4.7bn worth of its £12.1bn deposits were net loans, which was lower than peers.

Starling has thought about growing its capacity organically but has decided that an acquisition would be faster, according to the people familiar with the situation.

Starling declined to comment.

The neobank, which was founded in 2014 by former Allied Irish Banks executive Anne Boden, has been tied to an acquisition before. In April, Sky News reported that Shawbrook, the SME lender, had approached Starling about a potential combination, but no deal emerged.

One person familiar with the business said Starling was not eyeing up mergers but acquisitions of a mixture of cash or cash and shares.

Starling was previously criticised by the government’s former anti-fraud minister, Lord Theodore Agnew, for expanding its lending during Covid-19. The neobank built out a large portfolio of loans from the government-backed “bounce back loans”, which were issued by banks to support struggling businesses quickly during the pandemic.

This grew to become a significant chunk of Starling’s lending book, but it announced in May that it was setting aside £28mn to cover loans that may not have complied with the lending scheme’s requirements. The UK financial regulator fined Starling in 2024 for “shockingly lax controls” against financial crime during its expansion from 2017 to 2023. The bank said it had since invested heavily in strengthening its controls.

Starling has also built software called Engine, which allows lenders to design and build their own digital capabilities. The software, which Starling believes will be key to growth, is already being used by a bank in Australia and another in Romania. Last month, Starling announced that it had signed Canada’s Tangerine Bank as its latest Engine customer.

Starling is also weighing up buying a bank in the US to accelerate its global expansion, the Financial Times has previously reported. The plans could also include a US initial public offering. Part of the ambition is to plug Engine into an American bank and attract more local customers.

Raman Bhatia, the group’s chief executive officer, told the FT’s Global Banking Summit last week that Starling was also open to buying a European business in the next three to five years to further boost expansion.

Starling acquired in August the accounting software start-up Ember, which it hopes to provide to its small-business customers. “We have focused on high-quality acquisitions but now we want to turbocharge that engine in the UK so there is so much to do here in terms of market share and growth expansion,” said Bhatia at the FT’s banking summit.

Continue Reading

Policy levers to support viability and increase sustainable finance – United Nations Environment – Finance Initiative

In this new policy brief, UNEP FI and the European Banking Federation (EBF) share seven potential policy levers to help strengthen the investment case for the EU chemical sector and help it attract finance for its sustainable transition.

The brief discusses the sector’s market and decarbonization challenges and how new policy packages anchored in the European Green Deal—including the 2025 EU Chemicals Industry Package and the 2025 Clean Industrial Deal (CID), supported by the Industrial Decarbonization Bank—present opportunities by aiming to create the policy certainty and financing tools needed for large-scale sectoral transformation.

This is the first in a new EU sectoral policy brief series developed by UNEP FI and EBF, intended to strengthen the connection between financial institutions, policymakers, and the real economy.

Continue Reading

New materials could boost the energy efficiency of microelectronics | MIT News

MIT researchers have developed a new fabrication method that could enable the production of more energy efficient electronics by stacking multiple functional components on top of one existing circuit.

In traditional circuits, logic devices that perform computation, like transistors, and memory devices that store data are built as separate components, forcing data to travel back and forth between them, which wastes energy.

This new electronics integration platform allows scientists to fabricate transistors and memory devices in one compact stack on a semiconductor chip. This eliminates much of that wasted energy while boosting the speed of computation.

Key to this advance is a newly developed material with unique properties and a more precise fabrication approach that reduces the number of defects in the material. This allows the researchers to make extremely tiny transistors with built-in memory that can perform faster than state-of-the-art devices while consuming less electricity than similar transistors.

By improving the energy efficiency of electronic devices, this new approach could help reduce the burgeoning electricity consumption of computation, especially for demanding applications like generative AI, deep learning, and computer vision tasks.

“We have to minimize the amount of energy we use for AI and other data-centric computation in the future because it is simply not sustainable. We will need new technology like this integration platform to continue that progress,” says Yanjie Shao, an MIT postdoc and lead author of two papers on these new transistors.

The new technique is described in two papers (one invited) that were presented at the IEEE International Electron Devices Meeting. Shao is joined on the papers by senior authors Jesús del Alamo, the Donner Professor of Engineering in the MIT Department of Electrical Engineering and Computer Science (EECS); Dimitri Antoniadis, the Ray and Maria Stata Professor of Electrical Engineering and Computer Science at MIT; as well as others at MIT, the University of Waterloo, and Samsung Electronics.

Flipping the problem

Standard CMOS (complementary metal-oxide semiconductor) chips traditionally have a front end, where the active components like transistors and capacitors are fabricated, and a back end that includes wires called interconnects and other metal bonds that connect components of the chip.

But some energy is lost when data travel between these bonds, and slight misalignments can hamper performance. Stacking active components would reduce the distance data must travel and improve a chip’s energy efficiency.

Typically, it is difficult to stack silicon transistors on a CMOS chip because the high temperature required to fabricate additional devices on the front end would destroy the existing transistors underneath.

The MIT researchers turned this problem on its head, developing an integration technique to stack active components on the back end of the chip instead.

“If we can use this back-end platform to put in additional active layers of transistors, not just interconnects, that would make the integration density of the chip much higher and improve its energy efficiency,” Shao explains.

The researchers accomplished this using a new material, amorphous indium oxide, as the active channel layer of their back-end transistor. The active channel layer is where the transistor’s essential functions take place.

Due to the unique properties of indium oxide, they can “grow” an extremely thin layer of this material at a temperature of only about 150 degrees Celsius on the back end of an existing circuit without damaging the device on the front end.

Perfecting the process

They carefully optimized the fabrication process, which minimizes the number of defects in a layer of indium oxide material that is only about 2 nanometers thick.

A few defects, known as oxygen vacancies, are necessary for the transistor to switch on, but with too many defects it won’t work properly. This optimized fabrication process allows the researchers to produce an extremely tiny transistor that operates rapidly and cleanly, eliminating much of the additional energy required to switch a transistor between off and on.

Building on this approach, they also fabricated back-end transistors with integrated memory that are only about 20 nanometers in size. To do this, they added a layer of material called ferroelectric hafnium-zirconium-oxide as the memory component.

These compact memory transistors demonstrated switching speeds of only 10 nanoseconds, hitting the limit of the team’s measurement instruments. This switching also requires much lower voltage than similar devices, reducing electricity consumption.

And because the memory transistors are so tiny, the researchers can use them as a platform to study the fundamental physics of individual units of ferroelectric hafnium-zirconium-oxide.

“If we can better understand the physics, we can use this material for many new applications. The energy it uses is very minimal, and it gives us a lot of flexibility in how we can design devices. It really could open up many new avenues for the future,” Shao says.

The researchers also worked with a team at the University of Waterloo to develop a model of the performance of the back-end transistors, which is an important step before the devices can be integrated into larger circuits and electronic systems.

In the future, they want to build upon these demonstrations by integrating back-end memory transistors onto a single circuit. They also want to enhance the performance of the transistors and study how to more finely control the properties of ferroelectric hafnium-zirconium-oxide.

“Now, we can build a platform of versatile electronics on the back end of a chip that enable us to achieve high energy efficiency and many different functionalities in very small devices. We have a good device architecture and material to work with, but we need to keep innovating to uncover the ultimate performance limits,” Shao says.

This work is supported, in part, by Semiconductor Research Corporation (SRC) and Intel. Fabrication was carried out at the MIT Microsystems Technology Laboratories and MIT.nano facilities.

Continue Reading

Taiwan probes leaks of vital chip technology

Taiwan has begun trade secrets investigations in its critical chipmaking sector under newly broadened national security laws, but the probes have raised eyebrows for who they are targeting: not companies from China but from the island nation’s closest allies.

Last week, prosecutors charged the local subsidiary of Japanese chip equipment maker Tokyo Electron with failing to prevent alleged theft of trade secrets from Taiwan Semiconductor Manufacturing Company.

A week earlier, prosecutors raided two homes of former TSMC executive Lo Wei-jen, who joined Intel after he left the Taiwanese company in July, as part of a probe into whether the 75-year-old was sharing “national core critical technology” with his new employer.

The probe came after TSMC sued Lo for violating his non-compete agreement, saying it was highly probable he “uses, leaks, discloses, delivers or transfers TSMC’s trade secrets and confidential information to Intel”.

Legal and industry experts in Taiwan said they were glad to finally see investigators getting serious about protecting a technology that had made Taiwan indispensable to the global economy. TSMC is the world’s largest chipmaker and dominates the market for cutting-edge semiconductors.

But, unexpectedly, the first trade secrets cases under national security laws are not implicating companies from China — long seen as the main culprit in technology theft and the biggest threat to Taiwan’s security — but Tokyo Electron, a supplier, and Intel, a customer and rival. Both companies are from countries viewed as Taiwan’s closest partners.

The cases have emerged amid concern in Taipei over the reliability of its main security backer given US President Donald Trump’s desire to make a “deal” with China, as well as his remarks about Taiwan “stealing” America’s chip business and allegedly freeriding on defence support.

An executive at a Taiwanese chip company in the US likened the investigations to a “man bites dog” scenario, saying the probes went against both the narrative that Beijing was poaching Taiwanese talent and Trump’s position that Taiwan had stolen the US’s technological leadership.

A US executive at a fund invested in semiconductor companies warned that Taipei’s more aggressive protection of its economic security could create risks for its geopolitical security by offending the US.

“This is not a good look for Taiwan right now,” said the executive. “Do they really think they can afford to go after US efforts to revive its chip manufacturing industry?”

Under pressure from the Trump administration, TSMC raised its US investment commitment by $100bn to $165bn in March. But Washington has made clear this is not enough. Trump administration officials have said they want 50 per cent of chip manufacturing to happen onshore, far more than TSMC’s expanded capacity can deliver.

In August, Washington agreed to take a 10 per cent stake in Intel as it aimed to resurrect the struggling company as a national champion of semiconductor manufacturing.

Neither TSMC nor the prosecution has targeted Intel directly or suggested its involvement in alleged technology theft. Prosecutors are only investigating Lo and have not brought charges against him. But observers suggested Washington could apply political leverage on any Taiwanese legal case.

“Taiwan has very limited options to refuse US requests and pressure” because it was pursuing a trade deal to lower Washington’s 20 per cent tariff on Taiwanese exports, said James Chen, a professor at Tamkang University in Taipei. It is also seeking US support for President Lai Ching-te’s tough approach towards China.

Provisions introduced in 2022 made the unauthorised transfer of “national core critical technology” to a foreign entity a national security offence for the first time, with a clear focus on China, to which Taiwan has been losing chip engineers for years.

In one of the most prominent controversies, Liang Mong-song, a former TSMC research and development executive, joined Semiconductor Manufacturing International Company, China’s largest chipmaker, in 2017 and is now its co-chief executive. He and the many TSMC engineers who followed him are credited with helping SMIC narrow its technology gap with the Taiwanese chipmaker.

The national security law amendments have raised the risk of such moves and stipulate much higher fines for leaking trade secrets to China than to allies such as Japan and the US.

But experts said the law still fell short. Tsai Ing-wen, Lai’s predecessor, initially aimed for a government role in initiating trade secrets cases through broad national security powers. The law adopted by parliament only allows prosecutors to move when a Taiwanese company makes a complaint, mirroring the US Economic Espionage Act and similar laws in Japan.

Investigators are now under pressure to build solid cases. In the instance of Tokyo Electron’s subsidiary, prosecutors have charged former TSMC staff with technology theft, but the indictment of the company only lists a failure to prevent such behaviour, not an accusation of theft itself.

“They have established the precedent that companies are responsible for building strong internal compliance mechanisms to protect against trade secrets theft,” said Jeremy Chang, chief executive of the Research Institute for Democracy, Society and Emerging Technology under Taiwan’s technology ministry.

“That could become a key task for everyone in the semiconductor supply chain,” he said, especially as more countries try to onshore chip manufacturing.

The former TSMC staff who have been charged declined to comment.

Tokyo Electron said the indictment of its subsidiary did not allege it had directed or encouraged its employee to improperly obtain TSMC technology. The company added that it had measures in place to prevent such behaviour and would strengthen its compliance systems.

Lo declined to comment. Intel stressed its commitment to internal controls that prohibit the use of third-party technology and said it had no reason to believe there was merit to the allegations involving Lo.

Observers cautioned that as Taipei worries about securing continued support from Washington, politics might play into prosecutors’ decisions.

“The government might have some thoughts of intervening or using leverage, but they cannot directly intervene in the judicial system,” said Chen. “This is a very politicised and sensitive issue.”

Category: 3. Business

‘Expect rapid improvements’

Residents’ safety ‘paramount’

‘Bad for London’

Data center opportunity