Fractal selected as preferred services partner in Claude Partner Network

by Hansraj Agrawal | Jul 7, 2026 | Technology

AI and analytics solutions firm Fractal Analytics has been selected as a Preferred Services Partner in Anthropic’s Claude Partner Network Services Track.

As a Preferred Services Partner, Fractal can access Anthropic’s technical teams while helping clients solve business problems at scale using AI.

Through this partnership, Fractal will combine its deep industry and functional expertise with Cogentiq, its agentic AI platform, and Claude’s advanced reasoning capabilities to help organisations unlock enterprise value.

“Anthropic’s commitment to building capable, reliable and responsible AI aligns closely with how we help enterprises adopt AI at scale. Through the Claude Partner Network Services Track, Fractal will help clients identify, build and scale Claude-powered solutions that create measurable value across their businesses,” said Srikanth Velamakanni, Co-founder, Group Chief Executive and Vice-Chairman, Fractal.

Rich O’Connell, Head of Alliances at Anthropic, said that Fractal has built more than 300 Claude-certified practitioners already delivering production work, including a contract intelligence solution serving more than 350 legal and procurement users at 50 per cent higher productivity.

“By incorporating Claude into its delivery model, Fractal is expanding how its teams engage with clients and accelerate outcomes across industries. Fractal is bringing the most advanced and trusted enterprise AI to sectors including CPG, retail, financial services, healthcare and life sciences, and technology, media and telecom, combining a scaled Claude practice with deep customer delivery experience,” said O’Connell.

This partner status recognises Fractal’s expertise and track record in designing, deploying and scaling Claude-powered AI solutions across industries including CPG, retail, technology, media, telecom, healthcare and life sciences, financial services and insurance, and across all enterprise functions.

Source link

Beyond consoles: Xbox, Sony, and Nintendo chart gaming's next evolution

by Hansraj Agrawal | Jul 7, 2026 | Technology

Our business today is not healthy, said Xbox CEO Asha Sharma said in a public restructuring announcement on July 6, as Microsoft revealed sweeping changes across its gaming division.

The announcement included layoffs, management restructuring and changes to Xbox’s studio portfolio. But the most significant part of Sharma’s message was the acknowledgement that some of Xbox’s biggest strategic bets, including Game Pass, multiplatform publishing and years of content expansion, did not grow at the pace the company expected. The admission suggested that even after assembling one of gaming’s largest studio portfolios and investing billions into content, the economics of the business remain under pressure.

The timing is notable as, less than a week earlier, Sony announced that physical game disc production for all new PlayStation releases will end in January 2028, accelerating its transition towards a fully digital future.

Taken together, the two developments raise a bigger question than either announcement does on its own: If Xbox is rethinking the economics of its gaming business and Sony is preparing for a future without physical games, where exactly is the gaming industry headed?

Xbox’s unusual admission

For much of the last decade, Microsoft’s answer to Xbox’s competitive challenges was scale. Beginning in 2018, the company embarked on one of the largest acquisition drives in gaming history. It bought studios including Ninja Theory, Obsidian Entertainment, inXile Entertainment and Playground Games before moving on to far larger deals, including Bethesda parent ZeniMax Media and eventually Activision Blizzard in a transaction worth nearly $69 billion.

Microsoft’s strategy was straightforward: acquire more content, grow Game Pass to shift consumers from one-time purchases to recurring revenue, expand cloud gaming and embrace multiplatform publishing. Sharma’s statement suggests those bets failed to deliver the growth Xbox expected.

One of the most revealing passages in the announcement was Microsoft’s acknowledgement that it is now “neither possible nor desirable to own every great independent studio” — a remarkable statement from a company that spent years pursuing precisely that strategy. Notably, Sharma stressed that none of Xbox’s publicly announced first-party games are being cancelled because of this.

The broader message is that growth will not come simply from accumulating more studios and more content, and that points to a challenge facing the wider gaming industry.

The economics of gaming are becoming harder to ignore

Modern gaming has become a far more expensive business than it was a decade ago. AAA game development budgets have ballooned as studios chase increasingly ambitious projects. Development cycles that once lasted two or three years now regularly stretch beyond that. Teams often consist of hundreds of developers spread across multiple countries. Marketing budgets can potentially rival those of Hollywood productions.

At the same time, publishers have spent aggressively to secure future growth through acquisitions. The result is an industry where costs continue to rise even as hardware growth becomes harder to achieve.

Unlike previous generations, console makers are now competing not just against each other, but against mobile gaming, social media, streaming platforms and countless other forms of digital entertainment. That reality has forced gaming companies to ask a difficult question: If the traditional console model is becoming more expensive, what comes next?

Why owning more studios is no longer enough

For years, the gaming industry’s playbook revolved around exclusivity. Exclusive games sold consoles. Consoles created ecosystems. Ecosystems generated software sales. That formula helped define the rivalry between Xbox and PlayStation. Today, however, publishers increasingly care less about where people play and more about ensuring they play their games somewhere.

The economics are easy to understand. When development costs rise, limiting a game to a single platform can mean leaving millions of potential customers on the table. That is one reason why Xbox has increasingly published games beyond its own ecosystem. Sony, meanwhile, has steadily expanded its presence across platforms, bringing previously exclusive PlayStation titles to a broader audience.

Even Sharma’s restructuring announcement hints at this changing mindset. The Xbox chief noted that Mojang and King — the companies behind Minecraft and Candy Crush — will now report directly to her because they represent some of Xbox’s largest businesses by monthly active players. That is significant because neither franchise depends on Xbox hardware.

Minecraft thrives across consoles, PCs and mobile devices, while Candy Crush is primarily a mobile game. Their value comes from audience reach rather than platform exclusivity, a shift that also helps explain the growing importance of subscription services. Xbox Game Pass and PlayStation Plus are designed to create ongoing relationships with players and generate recurring revenue, but Xbox’s latest admission suggests subscriptions alone have not been enough to offset broader pressures within the business. Increasingly, gaming companies appear to be building ecosystems that combine hardware, software, services and digital marketplaces into a single experience. That is where Sony’s latest move becomes particularly interesting.

Sony is preparing for a world beyond physical games

In an announcement published on the official PlayStation Blog, Sony Interactive Entertainment confirmed that physical disc production for all new PlayStation releases will end in January 2028.

The decision reflects a broader shift that has been underway across entertainment for years. Physical media once sat at the heart of gaming’s business model, allowing players to build collections, resell games and maintain a sense of ownership while retailers played a central role in distribution.

Digital distribution eliminates manufacturing and logistics costs, allows publishers to sell directly to consumers and keeps players within an ecosystem built around updates, subscriptions and online services.

For Sony, the move is not simply about removing discs. It is about aligning PlayStation with a future where digital ecosystems generate a growing share of value. But that transition also raises an uncomfortable question for consumers.

If everything is digital, do you really own what you buy?

One of the biggest differences between physical and digital media is ownership. Unlike discs, which can be stored, lent or resold, digital purchases often provide access rights tied to a platform ecosystem rather than ownership of a tangible product.

According to Ars Technica, Sony informed PlayStation customers in the United Kingdom that hundreds of previously purchased StudioCanal movies and television shows could be removed from their libraries because of licensing changes. While the issue involved video content rather than games, it served as a reminder that digital purchases often depend on agreements between content owners and platform operators.

What happens when gaming becomes a service

If physical games are disappearing and subscriptions are becoming more important, the next logical question is whether hardware itself becomes less essential.

Cloud gaming is one potential answer. Services such as Xbox Cloud Gaming already allow players to stream games without relying entirely on local hardware. Sony has also invested in cloud-based gaming technologies through PlayStation. The technology still faces challenges around latency, infrastructure and internet quality.

Yet cloud gaming points towards a future where access matters more than ownership and where services matter more than devices. That does not mean consoles are disappearing any time soon, though. High-performance gaming hardware still offers advantages that cloud services cannot consistently replicate.

However, the direction of travel is becoming increasingly clear. Gaming companies want their content and services to reach users wherever they are, whether that is on a console, PC, smartphone or smart television. Interestingly, one company recognised that principle long before the current debate emerged.

Nintendo’s bet on characters over consoles

While Microsoft and Sony are rethinking distribution and business models, Nintendo has spent years expanding the reach of its intellectual property beyond dedicated gaming hardware.

According to comments from Nintendo’s recent investor Q&A reported by ScreenRant, Nintendo executive Shigeru Miyamoto said the company realised there was a limit to the number of people its consoles could reach, but not to the number of people its characters could reach through films, mobile devices and other media.

That insight helps explain Nintendo’s growing push into films, theme parks, mobile experiences and licensing partnerships. The success of The Super Mario Bros. Movie demonstrated that Nintendo’s biggest franchises can thrive well beyond traditional gaming platforms. A live-action Legend of Zelda film is also said to be in development.

While Microsoft’s answer to industry change is services and Sony’s answer is digital ecosystems, Nintendo’s answer is intellectual property. The objective, however, is remarkably similar: reach audiences wherever they are rather than relying exclusively on hardware sales.

The future may not belong to the box under your television

The biggest takeaway from Xbox’s restructuring and Sony’s digital push is not that consoles are disappearing. It is that consoles are no longer the centre of the business in the way they once were.

For decades, success in gaming was largely measured by how many boxes a company could place under consumers’ televisions. Today, the industry’s biggest companies increasingly measure success through audience reach, subscription revenue, ecosystem engagement and the strength of their intellectual property.

Xbox’s admission that Game Pass and content expansion did not grow as expected reflects the growing difficulty of relying on old assumptions. Sony’s decision to phase out physical game discs reflects confidence in a future built around digital distribution. Nintendo’s expansion beyond games shows how valuable gaming franchises can become when they evolve into broader entertainment brands.

Different strategies. Different companies. But all three point towards the same conclusion.

Gaming’s biggest companies are increasingly building businesses that extend beyond the console. The next era of the industry may still be played on a PlayStation, Xbox or Nintendo system, but it is increasingly being shaped by everything that exists beyond the hardware itself.

Source link

What Centre's proposed stake in Sarvam means for India's AI ambitions

by Hansraj Agrawal | Jul 7, 2026 | Technology

Last month, Sarvam AI became India’s newest artificial intelligence (AI) unicorn. Days later, The Economic Times reported that the Centre was considering taking a 1-2 per cent stake in the Bengaluru-based startup through compute support extended under the IndiaAI Mission.

As the race to build sovereign AI capabilities intensifies, India is seeking to move beyond exporting talent to building its own AI ecosystem. The IndiaAI Mission, which backs domestic compute infrastructure and indigenous foundation models, reflects that ambition.

But if the reported proposal takes shape, what it could mean for Sarvam’s competition, industry innovation and the future of India’s AI ecosystem?

Why Sarvam?

Sarvam is among a handful of companies selected by the Centre to develop indigenous, multilingual and domain-specific foundation models under the IndiaAI Mission, as part of its strategy to strengthen the country’s AI capabilities. Under the programme, the government is providing the selected companies access to graphics processing unit (GPU) compute at subsidised rates, covering up to 40 per cent of the cost.

Founded in 2023, Sarvam is building a full-stack AI business spanning foundation model development, inference infrastructure and enterprise applications, with a focus on Indian languages and domestic use cases. The company says its products are being deployed across sectors such as banking, insurance, government services, and defence.

Why could the government take equity?

What makes the proposal unusual is not the government’s support for Sarvam, but the form that support could take. Unlike conventional government support programmes, the proposed investment is not a direct cash infusion. The Centre’s stake would arise from compute infrastructure and other resources provided under the IndiaAI Mission, with the equity likely to be acquired through convertible instruments, The Economic Times had reported.

Experts say the move reflects a broader shift in how governments are beginning to view AI.

“An investment by the government into Sarvam AI could signify a paradigm shift in how India supports the creation and commercialisation of AI technologies,” said Kumar Rajagopalan, vice-president, strategic initiatives and India country head at Dexian. “It suggests the government wants to move from merely providing AI infrastructure to becoming a stakeholder in the country’s AI ecosystem.”

Rajagopalan said the proposal also reflects the government’s commitment to developing sovereign AI capabilities, which will support additional investments into cloud-based AI infrastructure.

What would this mean for India’s AI ecosystem?

Rajagopalan argued that equity investments should be reserved for projects with strategic national importance, while the government’s primary focus should remain on expanding shared AI infrastructure, such as compute capacity, high-quality datasets, cloud resources and research.

“Rationalising equity investments should only be given priority where there are strategic national implications and only if there are clear, fair and transparent policies governing their use,” he said.

Jaspreet Bindra, co-founder and chief executive of AI & Beyond, said government equity can be justified where scarce compute materially de-risks a frontier AI company, “but it should be the exception, not the model”.

India’s priority, he said, should remain building shared AI infrastructure rather than picking national champions.

Ritwik Batabyal, chief technology and innovation officer at Mastek, an IT services and digital engineering company, said, “AI development, especially at the foundation model level, requires significant investments in computing power, talent and research. If government support helps reduce these barriers, it could accelerate innovation.”

“At the same time, such initiatives should be backed by clear and transparent guidelines so that the wider startup community also has access to similar opportunities,” Batabyal said. “A healthy AI ecosystem is one where innovation is encouraged across the board, rather than concentrated in a few companies.”

Source link

Inside Claude: Anthropic finds AI uses a human-like reasoning workspace

by Hansraj Agrawal | Jul 7, 2026 | Technology

For years, the debate around artificial intelligence (AI) has centred on one question: Can machines ever think like humans? While AI systems can write essays, generate code and solve complex problems, they are still believed to lack qualities such as consciousness, self-awareness, emotions and genuine understanding. At the same time, rapid advances in AI have fuelled speculation about whether these systems could one day surpass human intelligence, making it increasingly important to understand what actually happens inside them before they generate a response.

While models such as Claude, ChatGPT and Gemini can perform increasingly complex tasks, the reasoning behind their responses remains largely hidden. As a result, researchers often describe them as “black boxes” because their outputs are visible, but their internal decision-making is difficult to explain, even for the companies that built them.

US-based AI entity Anthropic, in its research, tried to understand what happens inside its large language model (LLM) Claude before it generates a response. Instead of focusing on whether the model gives the right answer, the researchers studied how it solves problems internally. They found evidence of what they call a “global workspace”, a temporary internal space where important information is gathered and processed before Claude produces its final response.

Anthropic makes it clear that the findings do not mean Claude is conscious or has human-like awareness. Instead, the research suggests that advanced AI models may have developed internal reasoning mechanisms similar to those described in theories of human cognition. The company believes this could help researchers better understand how AI works and improve its transparency and safety.

From neuroscience to artificial intelligence

Anthropic says the idea behind the research comes from neuroscience, particularly how scientists study conscious and unconscious processing in humans. Much of the brain’s activity happens automatically, without people being aware of it. Only a small amount of information enters conscious awareness, where it can be used for reasoning, planning and problem-solving. The researchers wanted to examine whether a similar distinction exists inside Claude.

To explore this, Anthropic turned to “Global Workspace Theory”, a framework that explains how information becomes consciously accessible. According to the theory, the brain selects a small amount of important information and makes it available to different parts of the brain for reasoning and problem-solving. The researchers investigated whether Claude has a comparable internal workspace that supports these functions.

Anthropic wanted to investigate whether something comparable emerges inside a LLM. Importantly, the researchers were not trying to prove that AI has consciousness. Instead, they wanted to determine whether AI develops organisational structures that resemble some of the computational mechanisms described in human cognition.

Their experiments identified what Anthropic calls the ‘J-space’, an internal reasoning space where Claude appears to gather and process information before generating a response.

What is the ‘J-space’

To explore Claude’s internal activity, Anthropic developed techniques to monitor patterns inside the neural network while the model worked through different tasks. Rather than identifying individual thoughts, the researchers found recurring patterns that appeared to represent concepts the model was actively using. They collectively referred to this internal representation as J-space. Anthropic says the J-space was not programmed into Claude but emerged naturally during the model’s training.

J-space can be thought of as an internal working area where Claude temporarily stores ideas while solving a problem. These ideas are not always visible in the model’s final response. Instead, they exist behind the scenes and influence how Claude arrives at its answer.

Unlike a conversation with a user, where only the finished response appears on screen, J-space captures information that remains hidden during normal interactions. According to Anthropic, this makes it possible to observe parts of the model’s reasoning process that would otherwise remain inaccessible.

The researchers argue that this workspace behaves less like a collection of stored facts and more like an active reasoning environment where information is combined, updated and manipulated before the final response is generated.

What the experiments revealed

To understand whether J-space played a meaningful role, Anthropic conducted a series of experiments. One involved mathematical reasoning. Claude was asked to solve arithmetic problems while providing only the final answer. Although the visible response contained no intermediate calculations, the researchers observed changing internal representations that corresponded to successive steps in the calculation.

This suggests that the model was not jumping directly to the final answer. Instead, it appeared to work through intermediate reasoning internally before generating its response.

Another experiment examined whether Claude could deliberately maintain an internal idea while performing a different task. Researchers instructed the model to think about a specific object while simultaneously copying unrelated text. Although the visible output focused entirely on the copying task, the internal workspace continued to represent concepts associated with the requested object.

This finding suggests that Claude can maintain internal representations independently of what it is writing externally. In other words, the model appears capable of separating its internal reasoning from its outward response, at least in certain situations.

Anthropic also tested whether this internal workspace could be suppressed. Even when instructed not to think about a particular concept, traces of that concept continued appearing in the internal representations. This resembles the difficulty humans often experience when trying not to think about something.

Other key findings

J-space evolves with training: Anthropic found that J-space exists before post-training but develops a distinct “Claude” perspective after the model is trained to act as an AI assistant.

It detects risks internally: In one test, words like “warning” and “dangerous” appeared in J-space before Claude responded to a potentially harmful user query.

It supports self-monitoring: During roleplay, J-space activated words such as “fictional” and “disclaimer”, suggesting Claude internally tracks when it is acting as a character.

It helps generate natural responses: Disabling J-space made Claude’s responses more mechanical, indicating the workspace supports richer, experience-based language.

Why switching off the workspace mattered

One of the most significant experiments involved reducing access to the internal workspace while leaving the rest of Claude’s neural network intact. The results revealed an important distinction between language generation and reasoning.

Without this workspace, Claude could still perform relatively straightforward tasks. It remained capable of writing fluent text, answering simple factual questions and producing grammatically correct responses.

However, when the tasks required multiple reasoning steps or the integration of different pieces of information, performance declined noticeably. According to Anthropic, this suggests the internal workspace is particularly important for higher-order reasoning rather than basic language generation. It appears to help the model organise information before producing more complex responses.

Although the experiments do not explain every aspect of Claude’s behaviour, they indicate that advanced reasoning may rely on specialised internal mechanisms rather than emerging solely from next-word prediction.

Does this mean Claude is conscious

The company says its research does not show that Claude has feelings, subjective experiences or human-like consciousness. Instead, the study focuses on what researchers call “access consciousness,” a concept from neuroscience that refers to information a system can access, reason with and use to make decisions.

Anthropic argues that the J-space appears to perform many of these functions. It stores information that Claude can deliberately use while solving problems, while much of the model’s other processing continues automatically in the background. However, the researchers emphasise that this should not be confused with phenomenal consciousness, which refers to the ability to have subjective experiences or emotions.

The company says there is currently no scientific method to determine whether an AI system possesses that kind of awareness. One of the study’s key findings is that J-space was not intentionally designed by Anthropic. Instead, it emerged naturally during Claude’s training, suggesting that advanced AI models can develop internal reasoning mechanisms on their own.

At the same time, Anthropic notes that Claude’s internal workspace is fundamentally different from the human brain. While human conscious thought combines images, sounds, memories and actions, Claude’s workspace is almost entirely built around language because words are the model’s primary way of interacting with the world. The company also points out that human working memory fades over time, whereas Claude can retrieve information from earlier parts of a conversation using its attention mechanism.

Rather than claiming AI is conscious, Anthropic says the research offers a new way to understand how language models reason internally. The company believes this could improve AI transparency and safety, while adding that questions around machine consciousness remain unresolved and require broader scientific and ethical debate.

What does this mean for the future of AI

Anthropic believes interpretability research will become increasingly important as AI models grow more capable. Current frontier models already influence software development, customer service, scientific research and enterprise decision-making. Yet their internal processes remain difficult to explain.

Improving transparency could help developers identify hidden biases, reduce hallucinations, strengthen safety testing and increase confidence in AI-assisted decisions. For regulators, such research may also support future governance efforts by providing more reliable methods for auditing advanced AI systems. The study reflects a broader shift within AI research, from measuring what models can do to understanding how they do it.

Source link

India's 5G performs well on speed, but falls short where AI needs it most

by Hansraj Agrawal | Jul 7, 2026 | Technology

Ask an artificial intelligence (AI) assistant on your phone connected to mobile internet to read through a long PDF and write you a summary. Watch how long it takes for that document to leave your phone before any answer comes back. That upload gap, more than the download speed operators love to advertise, is where India’s 5G network is struggling most, according to a new report from connectivity intelligence provider Ookla.

In its report, titled “Beyond Download Speed: Benchmarking 5G Mobile Networks Against AI Workloads,” Ookla tested 86 operators across 22 markets against the network conditions that AI applications actually need to work well. India ranks ninth globally on headline 5G download speeds. On the two things that matter most for AI workloads, upload capacity and latency, it lands in the bottom tier.

Why upload matters

Every mobile network splits its capacity between two directions: downlink, which is data flowing to your phone, and uplink, which is data flowing from your phone back to the network. For years, networks were built around a roughly 90:10 split, 90 percent of capacity reserved for downloads and only 10 percent for uploads, because people mostly consumed content such as videos and web pages rather than sending large amounts of data out, according to the Ericsson Mobility Report of June 2026.

AI upends that math. When someone sends a prompt to a chatbot, especially one attached to a document, a photo or a long conversation, all of that content has to travel upstream before any answer comes back. Ericsson’s data shows text-based AI chat already runs closer to a 29:71 uplink-to-downlink split, meaning uploads now take up nearly three times the share they used to. Voice AI and AI agents push that further, closer to an even 50:50 split between what goes out and what comes in, and AI glasses or camera-based AI, which continuously stream what they see and hear back to the cloud, push uplink demand higher still.

India networks allocates 7.53 percent of its 5G throughput to upload, delivering a median upload speed of 15.75 Mbps, Ookla found. That falls short of the 20 Mbps target the report sets for AI modalities like augmented reality and multimodal vision, a bar only ten of the 22 markets studied managed to clear. There is a bright spot here. India’s upload share actually grew by 1.53 percentage points between 2023 and 2025, at a time when 12 of the 22 markets studied saw their upload share shrink or stay flat, Ookla noted. The improvement is real, it is just starting from a low base.

Part of the reason is technical. Much of India’s mid-band 5G spectrum uses Time Division Duplex (TDD), where uplink and downlink share the same frequency band and take turns using it. Giving uploads more room directly eats into download capacity, and every operator using the same band in a market has to coordinate its timing to avoid interference, so no single telco can simply decide to fix this alone.

Where the delays creep in

The other constraint is latency, or how quickly data makes a round trip. Ookla measured what it calls multi-server latency, the baseline responsiveness a network offers under normal conditions, and found India’s figure stands at 51.6 ms. That puts India in a group of just four markets, along with South Korea (53 ms), the US(50.5 ms) and Spain (50.2 ms), that miss the sub-50 ms mark Ookla sets as the target for text-based AI chat and AI agents. Eighteen of the 22 markets studied cleared that threshold. Ookla chose this particular metric to benchmark against AI requirements because, in its own words, “those thresholds describe what AI applications require under typical operating conditions, and multi-server latency is measured under those same conditions.”

For voice AI, which needs latency under 40 ms to sound natural rather than stilted, only 13 of the 22 markets qualify. For AR and multimodal vision, which needs latency under 10 ms, not a single market in the study clears the bar.

There is better news buried in the same report. Ookla also measured how much networks slow down under heavy load, a figure it calls the degradation ratio. India’s ratio comes in at 4.0x, among the best in the study and far better than markets like Thailand (11.4x) or Singapore (9.2x). So while India’s baseline latency is weak, its network does not fall apart badly once traffic gets heavy, which is not nothing.

The bottleneck beyond the network

There is a second delay most people never think about, the one that happens after data leaves the mobile network altogether and heads to the cloud servers where AI models actually run. This is not something telecom operators fully control, since it depends on where a cloud company’s data centres sit and how well an operator’s network connects to them, but it shapes how fast an AI reply comes back just as much as the mobile network does.

Ookla found India’s median latency to reach these cloud servers runs to 114 ms for Amazon Web Services, 109 ms for Microsoft Azure, 121 ms for Google Cloud and 158 ms for Oracle Cloud Infrastructure. Set against the rest of the world, that is a weak showing.

Regions like Europe and East Asia lead here, with South Korea reaching AWS in just 40 ms, Germany in 42 ms and the UK in 44 ms. This means users in these areas get more than double India’s headroom before an AI response even starts forming. But India also trails markets much closer to home. Singapore reaches AWS in 74 ms and Indonesia in 63 ms, both comfortably ahead of India despite facing similar distances to major cloud regions.

The gap points to something specific: India’s problem is not primarily geography, since Southeast Asian markets sitting at similar distances from cloud infrastructure do better, but the quality of interconnection between Indian networks and the hyperscalers’ edge locations.

Are network operators ready for the future?

For today’s most common AI use case, text-based chat, the answer is yes for now. Every one of the 22 markets studied, India included, meets the minimum bar for text LLMs, AI-generated video and AI agents at the median, Ookla found.

The concern is what happens next. Ericsson’s own scenario modelling projects that under a medium AI adoption scenario, additional AI traffic alone could make uplink demand three times higher by 2031 compared with 2025. Under a high-adoption scenario, that multiple rises to five times. Ookla’s report is even starker on the more demanding modalities ahead: not one of the 86 operators it studied anywhere in the world currently meets the target for multimodal AI.

Do operators need to redesign their networks

The industry itself seems to think so. At the Mobile AI Industry Summit held during MWC Shanghai in June this year, telecom executives reached what Tech Wire Asia described as an explicit consensus that improving uplink is now the single most urgent priority for mobile networks. Huawei used the event to unveil a solution it calls GigaUplink, built specifically to address this gap. In a separate keynote at the same event, reported by South Korean outlet The Elec, Huawei’s vice chairman David Wang listed “establishing sustainable and future-oriented spectrum planning and allocation” and “defining standards for AI-native core networks” among six priorities he expects to shape the industry over the next decade.

Ericsson made a similar case in a separate blog post this April, titled “Four moves for operators in the AI-native era.” It said AI traffic is “more uplink-heavy, more latency-sensitive, and more session-rich” than anything mobile networks were designed to carry, and that “best-effort” connectivity, the standard approach until now, is turning into a bottleneck on both user experience and operator revenue. Ookla’s own recommendation for operators is more concrete: deploy 5G Standalone architecture, turn on uplink carrier aggregation, and strike direct peering deals with cloud providers to cut down the latency gap.

Will India’s spectrum planning start favouring upload

There are early signs of this. TRAI released its recommendations for the country’s next spectrum auction on February 24, 2026, covering nine bands from 600 MHz to 26 GHz.

One recommendation speaks directly to the upload problem. TRAI has asked the Department of Telecommunications to explore using the 1427-1518 MHz band purely for what is called Supplementary Uplink, spectrum meant only to add upload capacity without cutting into download speeds elsewhere. It has also proposed setting aside some TDD spectrum for private and machine-to-machine networks, the kind of connections industrial AI systems depend on.

This falls short of the uplink-first approach Ookla’s report calls for. But treating upload as spectrum in its own right, rather than leftover capacity, is new. Whether it survives into the final auction terms remains to be seen.

Source link

What is SALT, India's first mobile 'liquid tree' to fight pollution?

by Hansraj Agrawal | Jul 7, 2026 | Technology

Researchers at the Council of Scientific and Industrial Research-Central Institute of Mining and Fuel Research (CSIR-CIMFR) said they have developed a mobile “liquid tree”, an air-purification system that offers a solution for polluted urban areas where planting conventional trees is difficult.

Called the Smart Algal Liquid Tree, or SALT, the compact system has already been installed at the CSIR-CIMFR campus in Dhanbad and at Northern Coalfields Limited in Singrauli, Madhya Pradesh.

But what exactly is a liquid tree, and how can algae help clean polluted urban air?

What is a liquid tree and how is SALT different?

A liquid tree is essentially a container filled with water and microalgae. Microalgae are microscopic, single-celled organisms that typically live in waterbodies. They use sunlight to convert carbon dioxide into oxygen and are responsible for generating about half of the global oxygen, according to ScienceDirect.

Like plants and trees, microalgae use light and carbon dioxide to grow. During this process, they absorb carbon dioxide from the surrounding environment and release oxygen.

A liquid tree puts this biological process inside an enclosed, compact unit that can be installed where there is little room for conventional trees.

Unlike earlier fixed liquid-tree installations, SALT is designed to be mobile, allowing deployment in different urban and industrial locations.

How does a liquid tree reduce air pollution?

The process is similar to photosynthesis:

Air containing carbon dioxide comes into contact with the algae-based system

The microalgae use light to absorb carbon dioxide during photosynthesis

Oxygen is released as a by-product of the process

In many designs, polluted urban air is drawn or bubbled through the tank. As the air interacts with the liquid and algae, some pollutants can be trapped, absorbed or biologically processed. This helps reduce pollution in the atmosphere.

“The primary purpose of this innovation is to combat poor air quality in densely populated and space-constrained urban areas where there is little or no room to plant large trees,” Vetrivel Anguselvi, senior principal scientist at CIMFR who led the project, told news agency PTI.

SALT can also operate with artificial light, allowing the biological process to continue even when sunlight is unavailable. The unit can run on solar power as well as electricity.

Can liquid trees replace real trees?

No. Liquid trees are mostly designed for places where conventional plantation is difficult because of a shortage of space.

Real trees don’t just produce oxygen. They support biodiversity, provide shade, cool urban areas, absorb rainwater and contribute to ecosystems.

SALT instead seeks to use the carbon-absorbing ability of microalgae in a smaller physical space. Its developers say the enclosed system does not require soil, needs little maintenance and is less vulnerable than conventional trees to pests and harsh urban conditions.

Where could SALT be used?

SALT is designed for use in densely populated and space-constrained areas where poor air quality coincides with limited room for large-scale tree planting.

CSIR-CIMFR sees potential for the system at transport hubs, industrial sites, educational institutions, shopping centres, parks and other crowded public spaces.

The institute is also exploring commercial production of the device. Researchers are working to make it affordable enough for possible use in residential areas and localities facing severe air pollution.

Source link

« Older Entries

Next Entries »

Times Top News

Times Top News