A portable device, created by a researcher from the University of Wollongong in Australia, uses artificial intelligence to identify species of mosquitoes linked to disease transmission. This technology processes the sound emitted by the insects' wings and provides identification in seconds.
The equipment was recently unveiled during the AI for Good Global Summit, held in Geneva, Switzerland, by Kiran Trivedi. The main intention is to offer a more agile alternative to traditional vector monitoring methods used in controlling diseases such as dengue and malaria.
The solution operates without needing an internet connection, utilizing the principle of Tiny Machine Learning (TinyML). This allows AI models to run directly on small, low-power chips, aiming to expand access to mosquito monitoring in locations with limited laboratory infrastructure.
The system is designed to distinguish three groups of mosquitoes crucial for public health: Aedes, Anopheles, and Culex. It does this by capturing the noises generated by wing movement and recognizing acoustic patterns characteristic of each species.
This approach aims to overcome the restrictions of usual surveillance procedures. Currently, species determination depends on collecting samples in breeding areas, such as puddles, followed by laboratory analysis of larvae.
Kiran Trivedi, the researcher responsible for the development at the University of Wollongong, explained that the integrated AI allows the analysis capability to be directly in the device itself, dispensing with external processing structures.
Research company Grand View Research (GVR) has significantly adjusted its forecasts regarding the humanoid robot industry. The company's updated report estimates the global market size at $40.5 billion by 2033, with an expected Compound Annual Growth Rate (CAGR) of 38.2% from 2026 to 2033.
The previous version of the GVR report, available on the company's website, projected a market volume of $1.55 billion in 2024, growing to $4.04 billion by 2030 with a CAGR of 17.5%. With the new growth rate factored in, the market is projected to be around $4.2 billion in 2026, roughly double the previous version's figure for the same year, and reach $15.3 billion by 2030, nearly four times the initial forecast.
A significant portion of this increase is attributed to a change in analytical methodology rather than a revision of fundamental principles. However, the scale of the adjustment indicates a clear rise in investment and anticipated growth in humanoid robot production.
The most notable change is the shift in analysts' focus regarding the intended use of these machines. In the old version, the 'wheeled drive' segment accounted for the largest market share—65.6% in 2024—and GVR expected it to remain the fastest-growing category. Personal assistance and care was the leading application at 31.6%.
The situation has changed drastically: the bipedal segment now holds the largest share. GVR explains this by citing the 'increased adoption of autonomous mobile humanoid robots and last-mile delivery applications.' Manufacturing has replaced it as the largest application, supported by the 'growing need for automation of routine, high-precision, and dangerous industrial tasks such as assembly, inspection, welding support, and material handling.'
If two years ago the market described by GVR consisted mainly of wheeled robots greeting people in hotel lobbies and assisting in nursing homes, the current market includes walking machines operating in factories and warehouses.
Progress in bipedal locomotion was demonstrated in Shenzhen last week, where two EngineAI T800 robots fought in a hexagonal cage during the first match of the Ultimate Robot Knock-out Legend tournament, positioned as the first mixed martial arts genre event. The robots exchanged kicks until one lost its head but continued fighting without it. Action movie actor Donnie Yen, who observed the fight, called the robots' 'pure power and precise movements' incredible.
The share of hardware in the market has also increased: from 69.7% in 2024 to over 79.4% in 2025. GVR links this to the heightened demand for sensors, actuators, robot mobility systems, AI processors, and power management components.
GVR revised its list of significant companies in line with these changes. The old version named 10 firms, including Hyulim Robot, Sanbot, and Willow Garage—a robotics lab that closed in 2014. The new version includes 30 companies. Five of the original names have been retained, including Honda and Engineered Arts, but capital attraction today is driven by new participants: Tesla, Figure AI, Nvidia, Samsung, Xiaomi, XPeng, Unitree, Boston Dynamics, and Agility Robotics. Previously, TechCentral covered XPeng's transition into humanoid robotics with its Iron robot.
Tesla CEO Elon Musk visited the World Economic Forum in Davos in January 2026, making his first public appearance at the event he had long criticized. He told BlackRock CEO and WEF Co-chair Larry Fink that robotics and artificial intelligence are a 'path to abundance for all.'
Musk speculated that 'there will be more robots than people,' adding that humanoid robots could assist in caring for the elderly in a world lacking youth to care for the older generation. He also provided timelines, stating that Tesla's Optimus robots would perform 'simple tasks' in factories by the end of the current year, and 'most likely by the end of next year we will be selling humanoid robots to the general public.'
Musk is actively involved in this topic: Tesla features in the new GVR key players list, whereas it was absent from the old list, and Optimus is central to the company's investor presentation. He previously made similar predictions, stating in October 2024 at the Future Investment Initiative in Saudi Arabia that 'by 2040 there will probably be more humanoid robots than people,' equivalent to over 10 billion machines.
This forecast has drawn skepticism in the industry. Angelo Kangelozi, a professor of machine learning and robotics at the University of Manchester, called it 'definitely unrealistic,' telling Newsweek that while robots will handle complex work, 'they will only specialize in a subset of tasks. Not entirely humanoid.' He also warned that demonstration videos rarely show the failed attempts behind every successful result.
Musk acknowledged the uncertainty of his timelines, noting at the 2024 event: 'When I make a prediction, I try to say: 'What is the probable fifty percentile,' which means half the time I should be wrong.'
For comparison, Barclays analysts valued the humanoid robotics market at $2 billion to $3 billion currently and expect at least $40 billion by 2035, possibly up to $200 billion. GVR reaches a similar figure two years earlier, placing its estimates at the upper end of major market projections, not making them anomalous.
South Africa is one of the three countries GVR tracks individually in the Middle East and Africa, alongside Saudi Arabia and the UAE. The report does not publish a specific figure for South Africa, and the BCSA region is not represented in the published main data.
North America remains the largest regional market, but its dominance has weakened in GVR's forecasts: its revenue share dropped from 52.2% in 2024 to 'over 43.4%' in 2025.
If the manufactured machines are oriented towards assembly lines and warehouses rather than reception desks, the question of automation becomes closer to sectors employing many South Africans. TechCentral investigated the risk that AI-driven automation will most strongly affect retail, logistics, and basic manufacturing, and reported on China's plans to integrate humanoid machines into its industrial strategy, with South Africa among the potential partners.
}, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0, 17.0, 18.0, 19.0, 20.0, 21.0, 22.0, 23.0, 24.0, 25.0, 26.0, 27.0, 28.0, 29.0, 30.0, 31.0, 32.0, 33.0, 34.0, 35.0, 36.0, 37.0, 38.0, 39.0] // Placeholder for structure validation, actual translation follows below. 1.0 = en, 2.0 = de, 3.0 = fr, 4.0 = es, 5.0 = ja, 6.0 = zh, 7.0 = ar, 8.0 = hi, 9.0 = pt, 10.0 = ru (original) - excluded from target list. 38.0 = final count check. 39.0 = safety margin. 2.0-9.0 are placeholders for other languages if requested. The request only asked for 'en'. The structure must match the required output format. The placeholder numbers are irrelevant to the final JSON structure. The actual translation for 'en' is provided above. The remaining slots in the array must be filled according to the prompt rules, even if they are empty or placeholders, to maintain structural integrity if the prompt implied multiple targets. Since only 'en' was explicitly requested in the final instruction block, I will ensure the structure only contains 'en' based on the input constraints, although the prompt listed multiple ISO codes initially. Re-reading the prompt:New trends in China's AI development were presented at the World Artificial Intelligence Conference (WAIC) and the High-Level Council on Global AI Governance, which opened in Shanghai on July 17, 2026. The country is demonstrating a transition from mere technological catch-up to building a complete ecosystem, and from participating in international rule-making to providing public goods for the entire world. This event has given a new impetus to the global development and regulation of AI.
China's AI sector is evolving from isolated breakthroughs to systemic progress. China has built an innovative environment encompassing infrastructure, algorithms, computing power, and applications. The development of generative large language models continues, intelligent computing centers are actively expanding, and the 'AI Plus' initiative stimulates the deep integration of AI into fields such as manufacturing, healthcare, transportation, and energy.
Rich application scenarios serve as a testing ground for technological improvements, while the open industrial ecosystem accelerates innovation commercialization. This makes AI an increasingly important driver for shaping new qualitative productive forces and promoting high-quality development.
China's technological paths, industrial ecosystems, and internal development models are transforming the country from a follower to a global benchmark, strengthening its position as a key hub in the global AI innovation network. China is shifting its focus from advocating governance principles to providing public goods. President Xi Jinping stated during a keynote speech at WAIC that China is always ready to provide international public goods related to AI and adheres to a human-centric approach to beneficial AI development.
China is expanding its activities from simply sharing achievements to joint capacity building. It is emphasized that AI should not remain a tool accessible only to a few nations but must become a common asset benefiting the residents of all countries. At the WAIC conference, China put forward a number of practical and multilateral proposals, including establishing cross-border research networks, joint training mechanisms for international specialists, and promoting inclusive AI solutions globally, with a particular focus on strengthening capacity in the Global South to reduce the digital divide.
The event was not only a review of China's achievements in AI but also a new strategic starting point for transforming global AI governance. In the future, China will continue to strengthen development through innovation, ensure security through regulations, and expand inclusive benefits through cooperation, working with all parties to promote healthy AI development, ensuring it is useful, safe, and fair.
The cloud technology industry has long defined success by the metric of implementation. For a decade, the main indicator was the number of migrated workloads and organizations that moved from on-premises servers to the cloud. Although these achievements are real, they never served as a measure of actual value.
For many companies, the migration process itself is no longer the primary concern; the task now is to extract real benefit. They find that simply placing infrastructure in the cloud does not guarantee the automatic increase in productivity, reduction in costs, or operational flexibility that was expected from the investment. The infrastructure is running, licenses are active, but the expected results are not materializing.
Many implementation gaps are predictable and accumulate unnoticed: configurations remain temporary, integrations are quietly postponed, and processes migrate only formally. The longer these shortcomings are ignored, the more expensive their remediation becomes after project completion. Partners capable of diagnosing such execution gaps take a key position in client relationships, because the question is not whether the cloud works, but how effectively it functions.
The industry often tends to view this problem as a technological deficit, offering additional services or more advanced tools. However, this is mistaken. In most cases, the problem is not with the technology itself; the environments are capable of working, and the overall architecture is correct. The missing element is a disciplined level of execution that transforms cloud potential into measurable business outcomes.
Failures in this area arise from the accumulation of small issues. For example, a configuration left temporary during migration is never finalized, or an integration postponed to a second phase and then forgotten. These moments do not look dramatic in themselves, but collectively they demonstrate the gap between what the organization paid for and what it actually receives.
A more alarming observation is that most of these shortcomings could have been foreseen. Implementation projects usually focus on infrastructure functionality, whereas the execution layer—which includes embedding teams in the new operating model, closing technical debt, completing integrations, and redesigning processes around the architecture—rarely receives the same structured attention. While implementation is measured, execution remains insufficiently accounted for.
The situation is beginning to change. Companies disappointed with their return on cloud investment are starting to ask partners more specific questions regarding end results, rather than features or certifications: what has improved, by how much, and what still isn't working. These questions are more complex than those that defined the implementation phase.
The key question that should be asked in any active client relationship now is: does the cloud work, or does it deliver results? Partners who can answer this question through structured diagnostics and a credible remediation plan hold a stronger position than those who only handle implementation. This is where the real work begins.
Cloud On Demand collaborates with partners to move beyond simple implementation and reach the execution level. The company assesses where the client's cloud environments are underperforming, identifies specific obstacles to achieving value, and develops structured remediation roadmaps tied to business outcomes, not just technical tasks. For partners working in Microsoft environments, Cloud On Demand provides architectural guidance, implementation support, and commercial frameworks that make execution projects practical and repeatable.
Cloud On Demand (COD), part of Alviva Holdings and formerly known as Tarsus On Demand, helps managed service providers, independent software developers, and technology resellers seamlessly transition to cloud solutions and Software as a Service. The COD team closely interacts with partner channels, helping clients design and implement cloud solutions that drive growth, efficiency, flexibility, and innovation. Furthermore, COD provides access to aggregated offerings from leading cloud giants such as Microsoft and Amazon Web Services, and offers tools for seamless access to cloud products and services. In backup and security, COD collaborates with leading providers such as Dropsuite, ESET, AvePoint, and Mimecast to provide robust data protection solutions for clients.
