News related to
demand

BlackRock’s iShares Bitcoin Trust has overtaken all other funds in its suite to become its most profitable exchange-traded product, generating approximately $244 million in fee revenue over the past year, analysts estimate. Its assets under management are approaching the $100 billion mark, driven by massive capital inflows into the spot Bitcoin ETF space.

IBIT has pulled in surging investor interest, with around $970 million flowing into it in a single trading session — the largest share of the $1.2 billion that U. S. spot Bitcoin ETFs collectively attracted that day. The fund now sits about $2.2 billion short of the $100 billion milestone, a level it is on track to cross more rapidly than any ETF in history.

That speed is notable: IBIT has amassed nearly $98–$100 billion in under 435 days, compared to Vanguard’s flagship S&P 500 ETF, which required over 2,000 days to reach the same threshold. This acceleration reflects investor appetite for regulated, exchange-traded access to digital assets.

BlackRock’s fee structure — currently 0.25% of AUM on IBIT — means that each incremental inflow directly boosts revenue. Analysts Eric Balchunas and James Seyffart have identified IBIT as now bringing in more income than long-established heavyweights in BlackRock’s portfolio, such as the iShares Russell 1000 Growth ETF and iShares MSCI EAFE ETF.

The inflow pattern underscores a broader shift. Since the U. S. Securities and Exchange Commission authorised spot Bitcoin ETFs in January 2024, the category has seen multiple single-day inflows exceeding $1 billion. In several past instances, such inflow spikes have coincided with short-term peaks in Bitcoin’s price, raising questions among market watchers whether these are signs of a turning point.

Bitcoin itself has responded: it pushed past $125,000 this month, buoyed by demand for regulated crypto exposure and safe-haven demand amid macroeconomic pressures. This run-up in price further amplifies IBIT’s appeal, because asset growth happens both through net inflows and valuation gains on its holdings.

BlackRock is also deepening its Bitcoin product strategy. It has filed to create a Bitcoin Premium Income ETF, which would generate yield by selling covered call options on Bitcoin futures — a move aimed at appealing to yield-focused investors seeking regulated crypto exposure. BlackRock has avoided entering the altcoin ETF arena for now, focusing its efforts on Bitcoin-linked products where regulatory clarity is more established.

Still, the meteoric rise of IBIT carries risks. Fee compression remains a structural challenge across the ETF industry, and competitors are poised to launch rival crypto products. The volatility inherent in Bitcoin could reverse gains sharply. Moreover, regulatory shifts — whether through tax treatment, disclosure mandates, or limits on crypto exposure — could reshuffle the competitive landscape.

Edgecore Networks will reveal its latest open infrastructure solutions at the OCP Global Summit in San Jose and GITEX Global in Dubai, with demonstrations of next-generation AI, networking, and compute platforms designed to ease deployment and boost performance.

The Taiwanese firm plans four flagship showcases: new Tomahawk 6–based data centre switches, a proof-of-concept for co-packaged optics, its Nexvec™ turnkey AI solution, and an upgraded Nous infrastructure controller offering composable compute management. The exhibitions will take place during the week of 13 October: at the OCP Summit from 13–16 October and at GITEX from 13–17 October.

Edgecore emphasises that the Tomahawk 6 switches can deliver more than 100 Tbps of bandwidth, supporting AI cluster demands with low latency and high density. The company adds that its CPO concept demonstrates integration of optical elements directly into switch silicon, reducing power and footprint. Its Nexvec™ offering is a pre-validated stack combining open networking with GPU compute, while Nous enables dynamic allocation of GPU, memory and compute resources.

Mingshou Liu, President of Edgecore, is quoted as saying that as AI workloads accelerate, simplifying deployment is crucial, and that the firm seeks to lead in the open infrastructure community with turnkey and scalable offerings.

Edgecore’s push comes at a time when hyperscale data centre operators are under pressure to manage burgeoning AI compute demands without ballooning costs or power draws. Industry trends point to co-packaged optics, disaggregated architectures, and composable resource frameworks as priorities for the next wave of infrastructure evolution. At the OCP Summit, the AI HW/SW co-design and open systems tracks will explore modular compute, memory architectures and photonic interconnects for AI clusters.

In parallel, GITEX Global has become a high-stakes stage for showcasing AI infrastructure, especially in the UAE’s drive to scale AI capabilities and attract global tech investment. With over 45 editions, the event draws global vendors and governments presenting AI, semiconductors, networking, and startup innovation. Edgecore’s presence signals its intention to engage key markets in the Middle East, where data centre growth, AI deployment, and regional cloud expansion are strong levers.

Edgecore is not alone in spotlighting AI infrastructure at these events. Major networking suppliers, chip vendors, and cloud players are expected to present competing solutions, particularly around optical integration, disaggregation, and hardware-software co-design. Observers will closely track how well Edgecore’s offerings stack up in power efficiency, interoperability, and ease of deployment.

Within its broader strategic frame, Edgecore continues to emphasise open, standards-based infrastructure. Its public messaging highlights disaggregated networking, composable compute, and software-defined manageability. The timing of product demonstrations at both OCP and GITEX gives the company a dual audience of deeply technical integrators and regional enterprise buyers.

At the OCP Summit, Edgecore’s participation also coincides with evolving community priorities—AI, sustainability, modular compute, and photonics interconnect are central tracks. During the event, Edgecore’s representative Tim Zhou will present on simplifying large-scale “agentic AI” deployments in a session scheduled for 16 October.][5])

Omnix International will unveil its HOT Systems suite—built in collaboration with PNY—at GITEX Global Dubai from 13 to 17 October, spotlighting AI workstations, liquid-cooled laptops and a new system-monitoring tool dubbed HOT Guard.

Visitors to the Hall 5, Booth 30 installation will see across-industry demonstrations—architecture, engineering, media, research—illustrating how tailored hardware accelerates digital transformation, reduces thermal throttling, and boosts reliability under sustained loads. HOT Guard introduces a continuous monitoring layer that oversees thermal behaviour, utilisation and system health, integrated with Omnix’s Hardware Optimization Technology stack.

The partnership with PNY ensures integration of high-performance memory and NVIDIA RTX graphics. PNY’s regional manager for MEA, Talus A Arukalil, emphasised that this collaboration positions the company to deliver “technology that doesn’t just keep pace with the future but defines it in the region.” On the HOT Systems side, Joseph John, Regional Sales Manager, asserted that the hardware is engineered “for today’s demands and tomorrow’s innovation,” targeting sectors where sustained compute power and system stability matter most.

According to Omnix’s own announcements, the new laptop models under the HOT Systems brand will be equipped with cutting-edge silicon—Intel Core Ultra 9 200HX or AMD Ryzen 9000 HX—matched with NVIDIA RTX 5080/5090 GPUs, DDR5 memory up to 128 GB, PCIe Gen 5 SSDs, Wi-Fi 7 and optional liquid cooling. These designs aim to balance high thermal headroom with energy efficiency and sustained performance even under intense workloads such as CAD, simulation and AI/ML inference.

Liquid cooling in a laptop format is still relatively rare. During previews, the company has framed this design choice as essential for holding consistent clock speeds over long renders or training sessions, reducing throttling under heat stress. As TBreak put it, the “pitch is stable clocks over flash specs,” targeting firms that can’t afford dips in performance mid-job.

With the industry steadily shifting from experimental AI use cases to full deployment, enterprises are increasingly looking for turnkey, dependable solutions—not just proof-of-concepts. In this environment, Omnix seeks to differentiate itself by offering not just hardware, but optimised stacks, monitoring, service support and aligned component integration.

GITEX Global 2025 is expected to host over 6,800 exhibitors from around 180 countries, showing advances across semiconductors, data centres, quantum computing, biotech and AI infrastructure. With global AI infrastructure investments accelerating sharply, regional companies and integrators are keen to showcase solutions tailored to local conditions—particularly in arid climates and metropolises with high operating temperatures.

Cisco will unveil its latest innovations at GITEX GLOBAL 2025 in Dubai, aiming to equip enterprises across the Middle East and broader region with infrastructure built specifically for the AI era. The showcase focuses on AI-ready data centres, resilient network architectures, and strengthened security frameworks designed to support advanced workloads in data-intensive environments.

By joining the AI Infrastructure Partnership alongside BlackRock, Microsoft, NVIDIA, xAI and others, Cisco has committed to accelerating capital flows into datacentre and infrastructure projects. The partnership expects to unlock $30 billion in initial funding, scaling up to $100 billion including debt financing. Cisco brings its network, security and systems expertise to support the platform’s mission of secure, at-scale AI infrastructure.

At the heart of Cisco’s push is its new Cisco Data Fabric, launched in integration with Splunk, which provides federated data access across systems such as Amazon S3, Iceberg, Delta Lake and Snowflake, routing workloads intelligently to the best analytics or storage engine. This architecture is intended to reduce data movement, shorten analysis latency, and enable enterprises to transform machine data into AI-ready intelligence.

To support retrieval-augmented generation and agentic AI deployments, Cisco introduced the Secure AI Factory solution with NVIDIA and VAST Data. This validated architecture accelerates data extraction and retrieval at scale while enforcing security guardrails for every token processed. AI PODs built under this architecture offer pre-validated, high-performance building blocks for enterprise AI.

Networking and security are being redefined for the AI workload environment. At Cisco Live 2025, the company launched a next-generation network architecture for campus, branch and industrial environments, delivering unified management and devices purpose-built to support AI traffic patterns. It also introduced Hybrid Mesh Firewall and Universal Zero Trust Network Access to embed security into network fabric across hybrid and distributed environments. These features aim to simplify policy enforcement, improve visibility, and reduce complexity in securing both human users and AI agents.

Cisco CEO Chuck Robbins disclosed that fiscal year 2025 saw record AI infrastructure orders. In its final quarter, the company secured over $800 million from webscale clients, bringing total AI-focused orders to more than $2 billion—more than double its initial target. He described the pipeline as steadily growing into “hundreds of millions” as enterprises demand scalable, secure solutions.

In the Middle East, Cisco plans to demonstrate the regional relevance of its technologies under the theme “Make AI Work for You.” According to Cisco’s new AI Readiness Index, a vast majority of organisations in the UAE expect to expand their data centre capacity for AI within five years. The company will also present its work with Splunk on security and observability platforms at the event.

Cisco’s collaboration with Saudi Arabia’s HUMAIN is intended to build out an open, scalable AI infrastructure ecosystem aligned with Vision 2030. The multi-year programme includes deployment of cloud-based AI infrastructure, joint research initiatives, and efforts to grow local talent and innovation. The partnership underscores Saudi ambitions to position itself as a global AI hub.

Cisco sees accelerating demand for network modernisation. Its internal research reports that 97 percent of IT leaders view modernised networks as critical to deploying AI, cloud and IoT workloads, and 98 percent rate secure networking as essential for growth. However, the same study notes that only 40 percent of organisations have begun deploying advanced segmentation or intelligent control, illustrating an execution gap.

To strengthen its AI strategy, Cisco also expanded its presence in Indonesia by signing a Letter of Intent with BRIN to develop AI and cybersecurity capabilities via its Country Digital Acceleration programme. The plan includes building an AI Centre of Excellence and delivering training in AI, security and networking to 500,000 Indonesians by 2030.

Splunk’s observability portfolio itself is being reworked to embrace agentic AI, deploying AI agents to automate tasks such as root cause detection, alert correlation and issue resolution. This is intended to lighten the operational burden on IT and engineering teams while driving real-time insight.

Telecom operator du and technology vendor Nokia have completed a trial of an artificial intelligence–driven automation system aimed at simplifying and expediting the expansion and management of optical networks in the UAE.

The deployment tested Nokia’s WaveSuite AI, which combines traditional AI methods with generative models to assist du engineers with tasks such as planning, troubleshooting and documentation retrieval. The trial reportedly halved the time needed for optical network planning and produced designs 30 percent more efficient, while reducing errors during deployment.

During the trial, du’s engineering team used a single natural language interface to query live network status, access accurate documentation instantly, and simulate potential network evolutions. The system flagged possible inconsistencies or conflicting configurations early, allowing corrective adjustment before full-scale rollout. According to du, the result was faster troubleshooting, fewer manual errors and improved resource utilisation.

Saleem AlBlooshi, chief technology officer at du, said the trial “shows how innovation can transform network operations to face challenges brought on by increasingly sophisticated networks and ever-higher traffic volumes.” He emphasised that the automation of routine tasks and provision of intelligent tools would lead to more reliable, SLA-backed connectivity for customers.

From Nokia’s standpoint, Ron Johnson, senior vice president and general manager of Optical Networks, called WaveSuite AI “a demonstration of the real value of automation solutions with both classical and generative AI for optical networking.” He noted that the system reduces friction in planning, documentation search, and operational processes, and helps service providers accelerate provisioning of higher-speed, more reliable services.

The trial builds on du’s broader effort to embed AI and automation more deeply across its systems. Earlier this year, du partnered with Microsoft, Nokia, Khalifa University’s 6G Research Center and the ITU to launch an “Arabic Telecom LLM,” a large language model tailored for internal telecom operations. That model is designed to handle internal workflows, resolve device issues, process complaints and provide operational insights in Arabic and English. The initiative is part of du’s strategy to blend regional research leadership and global AI tools while retaining language and cultural fidelity in its operations.

This AI-driven automation trial comes amid growing pressures on telecom operators worldwide to scale quickly to meet surging demand for bandwidth, driven by AI workloads, data centers and new real-time applications. As networks grow in complexity, manual processes no longer scale efficiently, pushing operators into an era of closed-loop or autonomous operations. Nokia, in its public literature, frames network automation as essential to reducing manual intervention, improving performance and enabling faster service delivery across domains such as core, mobile, IP and optical networks.

In the UAE context, this trial signals du’s ambition to advance its optical infrastructure ahead of demand peaks, while benchmarking its operations for future 6G readiness. The success of automation in planning and deployment could reduce operational costs, increase agility, and support more complex services such as network slicing, ultra-low latency applications and differentiated service tiers.

Google has chosen Saudi startup Lahint for inclusion in its Growth Academy: AI for GovTech 2025 cohort, positioning the company among 24 global government-technology innovators supported by Google. The move aims to strengthen Lahint’s capacity to deliver AI-enabled public services and further Saudi Arabia’s push for digital transformation under Vision 2030.

Lahint, established in 2023 by Ahmed Saber and Mohamed Ibrahim, secured a $1 million pre-seed investment earlier this year to scale its AI and automation offerings in government services. The platform currently manages more than 60 digital government services and serves over 200 businesses, integrating robotic process automation and AI to streamline licensing, permits and administrative workflows.

By integrating Lahint into its accelerator, Google pledges technical support, mentorship and access to cloud infrastructure—tools that could accelerate Lahint’s roadmap to full automation of priority public services. Lahint has already earned multiple ISO certifications and is accredited by the Saudi Data & AI Authority with a Waee badge for AI ethics compliance.

Lahint has advanced its internal timeline for key automation milestones. The company is now targeting completion of its GovTech automation program by early 2026, a shift from an earlier 2027 target. This acceleration comes alongside newly signed partnerships with government entities to deploy real-time status reporting and predictive compliance tools covering domain areas such as licensing, taxation and corporate registration.

Saudi Arabia’s broader AI landscape is scaling rapidly. In 2025, the kingdom announced over $14.9 billion in investments at LEAP and pursued MoUs under SDAIA to reinforce its data and AI infrastructure. The establishment of HUMAIN, a state-backed AI company under the Public Investment Fund, is central to that strategy. HUMAIN, launched in May 2025, has already inked collaboration agreements with Nvidia, AMD and Qualcomm to develop AI factories and Arabic large language models.

Google’s AI programme in Saudi Arabia is part of a wider arrangement. A $10 billion joint investment by Google Cloud and the Public Investment Fund backs a national AI hub to reinforce the local ecosystem and serve regional demand.

Teams at GITEX Global 2025 in Dubai will put liquid-cooled AI laptops under extreme thermal stress, testing whether advanced cooling systems can uphold sustained performance in one of the world’s harshest climates. Omnix International’s HOT Systems division, in partnership with PNY, will unveil a lineup of AI workstations, liquid-cooled laptops and the new HOT Guard monitoring suite, aiming to prove that these machines can maintain stability in high ambient temperatures.

At the heart of the exhibit is the challenge posed by regional heat: daytime temperatures in Dubai frequently exceed 40 °C, while indoor exhibition halls still pose cooling constraints. For high-performance AI workloads—particularly those involving CAD, BIM and machine learning—thermal throttling can cripple throughput. HOT Systems claims its integrated cooling and hardware optimisation can circumvent that performance drop. The accompanying HOT Guard software will monitor temperatures, fan and pump behaviour and security in real time on site.

Liquid cooling is no novelty in high-density server enclosures, but applying it in portable form factors presents a tougher engineering problem. In data centres, the shift from air to liquid systems has accelerated because conventional cooling struggles to keep up with power densities above 10–15 kW per rack. Liquid systems offer reduced power use, higher thermal headroom and often reclaimed waste heat reuse. But translating that to mobile form demands compact and robust fluid paths, efficient pumps, and risk control against leakage.

Advances in chip cooling are reinforcing that liquid systems represent more than a marginal improvement. Microsoft engineers have developed a microfluidics cooling technique that embeds coolant channels within the silicon substrate itself, enhancing heat removal efficiency at the chip level. That suggests future AI hardware may increasingly rely on liquid solutions at a micro scale.

Regional data centre operators are already embracing the shift: Khazna Data Centres, active across the UAE, has rolled out the country’s first liquid-cooled AI data centre and is developing more AI-optimised campuses. The move reflects recognition that standard cooling infrastructure cannot scale profitably with AI workloads.

Technical design choices behind liquid cooling include direct-to-chip systems, which place cold plates atop CPUs or GPUs, and immersion cooling, where liquid envelops hardware. The former is more suited for modular, retrofit deployments; the latter can offer full heat removal but introduces maintenance and material challenges, especially for localised devices. A recent review in AI systems architecture emphasises single-phase direct cooling as a flexible compromise for power levels below extreme thresholds, though it demands high reliability in fluid delivery and sealing.

Within the exhibition context, the HOT Systems demo will likely emphasise sustained throughput rather than peak benchmarks. The company plans real-world workflows—such as rendering, simulation and AI model inference—to demonstrate how the cooling system maintains clock speeds under load. PNY’s involvement underlines the reliance on qualified memory, GPUs and driver support, ensuring that thermal gains translate into usable computing stability.

End-users in architecture, media, engineering and AI fields are closely watching such developments. If heat throttling can be mitigated in this climate, portable AI rigs could expand in markets previously constrained to air-conditioned labs. But adoption depends on how well the solutions resist failure under harsh dust, vibration and heat cycles inherent to Gulf environments.