Norway’s $2.2tn sovereign wealth fund, the world’s largest, is using AI to screen companies for risks such as potential links to forced labour and corruption, and help avoid financial losses as a
result, it said on Thursday.

One of the world’s largest investors, the fund holds stakes in around 7,200 companies globally, owning about 1.5 per cent of all listed stocks. It has often set the pace on environmental,
social and governance issues.

The fund’s investments are measured against a benchmark index set by the finance ministry, with equities tracked against the FTSE Global All Cap index.

Norway SWF screens new companies on the index

Each time that index includes new companies, the fund’s operator, Norges Bank Investment Management (NBIM), must screen them before they enter the portfolio.

Since 2025, NBIM has used large language models to screen all companies on the day they enter the equity portfolio, rapidly scanning for public information that data vendors
typically do not provide.

“Within 24 hours of our investment, the AI tools flag new companies in the fund’s equity portfolio with potential links to, for example, forced labour, corruption or fraud,” NBIM said in its annual responsible investment report, published on Thursday.

“In multiple instances, we identified and sold these investments before the broader market reacted to the risks, avoiding potential losses.”

AI is especially useful for researching smaller companies in emerging markets, NBIM said, noting that data vendors often offer limited coverage and international media may not report on
them.

“News may be limited to small media outlets in local languages, and controversies suggesting systemic failures in risk management may go unreported in international media,” it said.

Saudi Arabia’s Real Estate Registry (RER) and Real Estate General Authority (REGA) sandbox platform has been described as a global first in sovereign-level tokenisation infrastructure. But beyond the headlines, the initiative signals something more structural: the digitisation of property markets at national scale.

For Adam Popat, CEO of SettleMint, the infrastructure provider behind the blockchain backbone, the significance lies in the scale of the problem being addressed.

“Saudi Arabia’s RER and REGA sandbox platform tackles three structural problems in real estate at scale: trust, liquidity, and access.”

Real estate markets have long struggled with opacity, fragmented data, and inconsistent due diligence. The Saudi model aims to reset that foundation.

“By issuing every resident a digitally verified blockchain credential and assigning every property a unique on-chain identity, the system creates a transparent, tamper-proof ownership and compliance layer,” Popat explains.

The result, he says, is a reduction in fraud and a standardisation of due diligence processes. Trust, often assumed but rarely systematised, becomes embedded into infrastructure.

“Powered by SettleMint DALP (Digital Asset Lifecycle Platform), the infrastructure runs on a secure, high-performance, scalable blockchain layer, integrating with government and banking systems while using smart contracts to automate settlement and compliance.”

At scale, Popat is clear that this is not experimentation. “At scale, this is not merely a sandbox, it is a production-grade national digital backbone for real estate, already powering live, regulated transactions in a controlled environment and redefining how property markets operate.”

In an ecosystem increasingly crowded with proptech platforms offering tokenised property investments, SettleMint positions itself differently.

“SettleMint’s role is purely that of a technology provider, not an operator,” Popat says. “We deliver the production-grade blockchain stack, smart contracts, digital asset lifecycle tooling, and secure integration layers that enable tokenisation at national scale.”

The distinction is deliberate. SettleMint does not issue assets, operate marketplaces, or intermediate transactions. Instead, it provides the underlying infrastructure on which regulated players can build.

Proptech firms, by contrast, sit at the distribution layer. They onboard investors, structure offerings, and operate marketplaces. “All of this innovation remains fully aligned with REGA guidelines and operates on RER’s sovereign blockchain infrastructure,” Popat adds, noting that deployment takes place within secure, Saudi Arabia-based data centres to ensure national control and regulatory integrity.

Liquidity as structural shift

Perhaps the most transformative development is the emergence of secondary markets. A regulated secondary trading platform is now in development, a move that could fundamentally change how property behaves as an asset class.

“Secondary liquidity fundamentally changes real estate from a static, long-hold asset class into a dynamic, price-discoverable market,” Popat says.

Historically, property valuations update only when entire assets change hands. With tokenised fractional units trading in smaller denominations, price signals can emerge continuously.

“When tokenised real estate units trade in smaller denominations, market demand and supply begin to shape transparent, real-time pricing signals. This improves valuation accuracy, reduces information asymmetry, and creates more efficient capital allocation.”

For investors, the implications are significant. Fractional ownership combined with secondary liquidity reduces lock-in periods and improves portfolio flexibility.

“Structurally, this moves real estate closer to capital market behavior, but within a regulated, compliant framework,” he says.

With Dubai Land Department launching its own secondary trading platform, questions of competition naturally arise. Popat frames the development differently.

“We see these initiatives as complementary and part of a broader regional evolution rather than direct competition.”

Tokenisation at sovereign scale is not a zero-sum game, he argues. Parallel initiatives across jurisdictions may, over time, support standardisation in digital identity, compliance frameworks, and interoperability.

“Rather than competition, this represents ecosystem building at a regional scale,” Popat says, adding that harmonisation strengthens cross-border capital flows and foreign direct investment within the GCC.

Institutional readiness

While tokenisation promises faster settlement and fractional access, institutional capital remains cautious. According to Popat, three issues must be resolved.

“Tokenisation promises fractional ownership and faster settlement, but institutional capital typically looks for three things before fully committing: legal enforceability, embedded compliance, and institutional-grade custody.”

Legal enforceability requires that tokens represent rights recognised under property law and anchored to official land registries. In Saudi Arabia, that linkage is addressed through REGA’s regulatory framework and RER’s sovereign infrastructure.

Compliance, meanwhile, must be embedded directly into code. “Powered by SettleMint DALP, the platform implements an enhanced ERC 3643 framework with more than 50 interconnected smart contracts governing identity verification, transfer restrictions, investor eligibility, and lifecycle controls directly on chain.”

Custody must also meet institutional standards, with RER serving as custodian of property records within a regulated sovereign environment.

Beyond regulation, national tokenisation requires resilient technical architecture.

“A national tokenisation backbone must be cloud native, fully containerised, and orchestrated through Kubernetes to allow automatic horizontal scaling as transaction volumes grow,” Popat explains.

Load balancers, distributed node clusters, stress testing, and embedded identity verification are not optional features but architectural requirements.

Importantly, the initiative builds on systems already operating at scale, including RER’s deed management system and national platforms such as Yakeen and SADAD, which serve millions securely.

Innovation under supervision

Saudi Arabia’s regulators have adopted what Popat describes as a structured sandbox model.

“Sovereign regulators in Saudi Arabia are balancing innovation with investor protection by designing the sandbox as a controlled pathway to full scale deployment, not an unrestricted experiment.”

Participation is permissioned, compliance obligations are defined upfront, and the phased rollout allows regulators to assess market behaviour before broader scaling.

The result is innovation under real market conditions, but within clear supervisory boundaries.

Popat does not see tokenised real estate as a niche experiment.

“I do not see tokenised real estate remaining a niche innovation. Over the next five years, it is far more likely to become embedded into mainstream property infrastructure across the GCC, with adoption progressing in a phased and regulator led manner.”

Tokenisation, he argues, will integrate into land registries, settlement systems, and regulated investment products. Fractional ownership and programmable compliance will increasingly become standard components of property markets.

As secondary markets mature and regulatory alignment strengthens, the GCC could become a global reference point for responsible, sovereign-scale digital asset infrastructure.

For SettleMint, the focus remains clear. “At SettleMint, we look forward to supporting governments worldwide in building digital asset platforms the right way, with security, compliance, and long term sustainability at the core.”

In an era where digital assets are often associated with volatility and hype, Saudi Arabia’s approach suggests a different trajectory — one where tokenisation is not speculative experimentation, but institutional infrastructure.

Samsung arrives at the Galaxy S26 launch with unusual momentum in this region. The company recently rolled out the Galaxy Z Tri-Fold to the UAE, one of just five markets globally selected for the debut of its most ambitious foldable yet. Fadi Abu Shamat, VP and head of the Mobile eXperience Division at Samsung Gulf Electronics, now discloses that three consecutive batches sold out within an hour each. For a device at that price point, in a category that remains genuinely niche, it was a signal about where the Gulf consumer sits on the early-adopter spectrum.

The S26 is a different kind of statement. Where the Tri-Fold was about form, the S26 is about behaviour, specifically, Samsung’s argument that the smartphone era is giving way to what it calls the agentic AI device, a phone that reads context and acts on it without being prompted. “Every year we start the year by raising the bar,” Shamat says, “and then the rest of the year literally everyone just tries to measure up.”

For starters, the S26 Ultra ships with the mobile industry’s first built-in Privacy Display, a redesigned camera system with wider apertures on the Ultra, and an AI framework running on the Snapdragon 8 Elite Gen 5.

The Privacy Display is the S26 Ultra’s most technically distinct feature and the most immediately testable. The technology is not a software overlay or an accessory; it is embedded in the display panel itself, controlling how pixels disperse light at the hardware level. When active, it preserves full brightness and colour accuracy for the primary user while blocking side-angle visibility.

That distinction from existing solutions is meaningful. Stick-on privacy films work by degrading the viewing cone for everyone, including the person holding the phone. Samsung’s implementation does not carry that tradeoff. The user experience is unchanged; only the side view is restricted.

Control is granular. Shamat describes the options as follows: you can run it as always-on across every application, or configure it to activate only in selected apps, so browsing stays open but WhatsApp triggers the privacy lock automatically the moment the icon is pressed. You can also restrict it to the notification bar only, so that when your phone is sitting on a table during a meeting, and a message arrives, only the top strip of the screen is masked. The feature ships on the Ultra exclusively. Scaling it to the broader S-series will depend on display production capacity, built-in privacy glass remains expensive to manufacture at volume.

Camera: Hardware inputs, AI processing

The camera upgrade on the S26 Ultra follows a logic Samsung has applied consistently across recent generations: improve what the sensor captures, then let AI operate on better source material. On the Ultra specifically, wider apertures across the camera system allow more light to reach the sensor. Shamat puts a figure on it: 47 per cent more light absorbed than the S25 generation.

The engineering principle holds regardless of the precise percentage. More light at capture means less computational noise reduction is required in processing, which matters most in low-light stills and video. As Shamat puts it, night shooting is fundamentally a light-deficit problem: less light means more grain, particularly in video, and the NPU working in real time to eliminate that noise is the S26’s answer — visible, he says, even in the darkest test conditions available at launch.

On the software side, AI ISP improvements reach the front camera for the first time, targeting skin tone accuracy in mixed lighting. The S26 Ultra also introduces APV codec support, a professional-grade standard designed to maintain visually lossless quality through repeated editing, aimed at creators working directly from the device.

The horizontal stabilisation lock in video, which adds a fixed horizontal axis option for smoother footage on uneven terrain, has broader appeal beyond that audience.

Agentic AI: The gap between demo and daily use

The S26’s most consequential and least provable claim is the agentic AI framework. Samsung introduced Galaxy AI with the S24, iterated on it with the S25, and is now positioning the S26 as the point at which AI becomes proactive rather than reactive. The feature doing the most work here is NowNudge, a contextual suggestion layer that surfaces relevant actions during natural use, without the user initiating a search or opening an application.

Shamat walks through three scenarios. The first: you are in conversation and mention meeting someone next Tuesday. The phone, understanding context, has already identified a 3pm conflict in your calendar and surfaces a NowNudge suggesting 4pm instead, tap once to inject the text. The second: you mention a recent family trip to Cairo. A friend asks to see photos. The phone surfaces a curated set of images geotagged in Egypt from the relevant dates, again via NowNudge, without you opening the gallery. The third: you have a 4pm meeting in downtown Dubai and you are running late. The phone identifies the situation from the calendar and location context, and offers to book a taxi through your preferred service — in the background, one confirmation tap required.

The appeal of the framework is clear. As Shamat describes it, the conventional interaction loop, home screen, search, locate the app, open it, complete the action, disappears entirely. “Agentic AI is carrying all that burden in the background on my behalf,” he says. “I don’t have to care about which app, I want the functionality, I want the service.” He goes further: the concept of app stores, in his view, is on its way out, replaced by a layer that routes requests to whatever service is most appropriate without the user needing to know which one.

That is a significant claim, and the honest caveat is that controlled launch demonstrations are the least reliable environment in which to assess it. Whether NowNudge performs consistently across ambient conversations, multiple languages, overlapping calendar entries and real-world context noise is a question that will only be answered over time. Samsung’s on-device processing model — AI runs locally by default, no cloud dependency unless the user opts in — addresses privacy but also places a ceiling on model capability relative to cloud-based alternatives. The integration of Gemini and Perplexity as optional agents alongside Bixby suggests Samsung is keeping its architecture open on which layer will ultimately carry the most weight.

Battery and thermal management

The S26 Ultra supports 60W wired charging, reaching 75 per cent from empty in 30 minutes. Shamat cites a 21 per cent improvement in heat dissipation over the previous generation; like the camera light figure, this comes from the interview rather than official specifications.

The thermal question has specific relevance in this market. A device running persistent AI workloads at 45°C ambient temperatures is under different stress than the same hardware in a temperate climate. Samsung’s redesigned Vapor Chamber positions thermal interface material along the sides of the processor, distributing heat across a larger surface area. The Snapdragon 8 Elite Gen 5 delivers a 39 per cent NPU improvement and up to 19 per cent CPU gain over the prior generation, meaning AI processes that previously required visible processing time or cloud round-trips can now run continuously in the background without measurable impact on responsiveness.

Ecosystem and what follows

The S26 launches alongside the Galaxy Buds 4 series.

The Buds 4 Pro introduce a pressure-sensitive stem that distinguishes between squeeze intensities to trigger different functions, a larger driver for improved bass, and head gesture controls for call management. Paired with the S26, AI features, including live translation, run on-device with no cloud processing by default.

For the remainder of 2026, Shamat signals new form factors in the second half without specifics.

The Ultra runs from 12GB/256GB to 16GB/1TB. Colours across the range: Cobalt Violet, White, Black, Sky Blue, with Pink Gold and Silver Shadow exclusive to Samsung.com.

Samsung Galaxy S26: Availability and offers

The Galaxy S26 Series is available for pre-order in the UAE until March 10, 2026, via Samsung.com/ae, Samsung stores, and select retail partners, and is available in four colours; Cobalt Violet, Sky Blue, Black, and White.

Pre-order offers on Galaxy S26 Series include:

• Memory upgrade: 512GB for the price of 256GB and 1TB for the price 512GB
• Samsung Members benefits: including a 1-year Samsung Entertainer membership, Amazon Prime (12 months), OSN+ (4 months), Anghami Plus (3 months), and Careem Plus (6 months)
• Trade-in: Up to Dhs2,599 saved with trade-in
• Samsung Care+: Optional coverage for added peace of mind, backed by genuine Samsung parts

The cost of hiring a very large crude carrier to ship 2 million barrels from the Middle East to China exceeded $200,000 a day on Thursday for the first time since 2020, LSEG data showed.

The benchmark freight rate, also known as TD3, rose to W218.52, or $206,141 per day, on the Worldscale industry measure used to calculate freight rates, the highest since April 2020.

It has nearly quadrupled from the start of the year, the data showed.

The surge in oil shipping costs follows increased crude exports from the Middle East as traders have accelerated charters ahead of possible tensions between the US and Iran. The rise is likely to reduce Asian refiners’ profits.

South Korean shipping group Sinokor’s VLCC buying spree has also supported freight rates, industry sources have said.

Adding to spot demand, Saudi Arabia’s biggest oil shipper Bahri has provisionally chartered three VLCCs – Nissos Anafi, DHT Jaguar and Maran Dione – to load crude on March 11-13 at W190-191 levels, said a shipbroker who declined to be named due to company policy.

Bahri did not immediately respond to a request for comment.