As per the start-up, XCENA’s technology, built on the open Compute Express Link (CXL) standard, expands a computer’s memory beyond traditional CPU limits, apart from executing computation where data resides, reducing latency, energy use and total cost of ownership.

Founded by semiconductor veterans from Samsung and SK Hynix (South Korean semiconductor major), XCENA combines deep hardware expertise with a full-stack software SDK to enable rapid deployment and seamless workload acceleration for hyperscalers, telcos, and research institutions.

In today’s episode of the “Start-up of the Week”, International Finance will talk in detail about XCENA, which, by bridging the gap between compute and memory, is powering a new class of intelligent, efficient, and scalable AI infrastructure.

Addressing AI’s data relay bottleneck
Every prompt, every question asked to ChatGPT, starts a data relay race where information leaves memory, passes through a CPU for preprocessing, travels to a GPU (Graphics Processing Unit) for heavy computation, and then makes its way back to the user in the form of “AI-generated replies”.

XCENA sees this as a structural bottleneck, which results in data routing through some of the most expensive and power-intensive chips in the industry on every single user prompt. To solve this, the four-year-old company, with a secondary presence in the United States, has come up with a chip that places compute capabilities much closer to DRAM (Dynamic Random-Access Memory). DRAM is known as the type of volatile computer memory used as the main memory (RAM) in PCs, laptops, and smartphones that temporarily stores data that users’ CPUs need to actively run programs and apps.

XCENA is using DRAM, the fast, short-term memory chips, as the main mechanism enabling routine data operations to be handled near memory without the costly round data trips between CPUs, GPUs, and memory. XCENA was in the news in the first week of June 2026 by raising USD 135 million in a Series B at a valuation of USD 570 million, which brought its total capital raised to USD 185 million.

XCENA CEO Jin Kim co-founded the start-up in 2022 alongside CTO Dohun Kim and CPO Harry Juhyun Kim, veterans of Samsung and SK Hynix, the memory giants also known for supplying chips powering Nvidia’s GPUs.

CPUs and GPUs have both gotten smarter over the decades. Memory never did. XCENA wants to change that. The recent rise in memory prices and related stocks points to a broader shift in AI infrastructure toward memory-centric architectures,” Jin Kim told TechCrunch.

While the chipmaking trio of Samsung, SK Hynix, and Micron crossed a trillion-dollar valuation for the first time, further consolidating their hold on the global semiconductor ecosystem, XCENA is taking a different path by putting all of its money on inference, which, as per Kim, “is increasingly becoming a memory scaling problem.”

The World’s First CXL 3.2 Computational Memory
XCENA’s chip, the MX1, connects to the CPU through CXL (Compute Express Link), becoming a dedicated express lane in the process between the processor and memory by processing data before it ever needs to leave the memory module. The breakthrough technology brings compute to the data, not the other way around. As per the start-up, what used to require 10 servers could potentially run on just one.

“While GPUs excel at matrix multiplication — the heavy math behind AI model training — much of the surrounding data orchestration, including preprocessing, KV cache management (the system that stores prior conversation context so a model doesn’t have to reprocess it), and data caching, still runs on CPUs. Our chip handles those tasks directly within the memory module itself,” Kim said.

MX1, the world’s first computational memory controller supporting CXL 3.2, is built to break the memory wall in AI data centres. It integrates up to 2 TB of DDR5 memory with thousands of RISC-V cores for near-data processing, reducing data movement and computation.

workload. MX1 also supports PCIe 6.0 SSD-backed expansion, enabling scalable and cost-efficient memory for large-scale AI workloads. “The company explained its breakthrough product through these following words.

Among the core features of the MX1, we have CXL memory expansion, with the chip enabling memory expansion up to 2 TB via CXL. The functions achieve up to 128 GB/s bandwidth over PCIe 6.0. The chip also maximises system efficiency by reducing unnecessary data movement in AI and HPC (High-Performance Computing) applications.

MX1 also brings “near-data processing” capability, as thousands of RISC-V cores process data near memory, reducing CPU load and latency. The chip has been optimised for data-intensive workloads such as AI inference, RAG, vector databases, and KV cache.

The chip also possesses enhanced RAS (Reliability, Availability, and Serviceability) features like Chipkill, SSD RAID, and ECC for fault recovery and data integrity. These capabilities ensure high availability and serviceability for mission-critical environments like data centres.

MX1’s “memory compression” supports software-based compression and hardware-based decompression by reducing overall cost by allowing more data to be stored and processed in the same physical footprint.

MX1’s scalable architecture, combining DRAM and PCIe 6.0 SSDs, enables petabyte-scale memory by minimising latency while maximising capacity through DRAM caching.

And last but not least, MX1 also provides developer-friendly tools across all layers, including runtime. It offers an integrated software environment optimised for various workloads, including SQL (Structured Query Language), vector processing, and graph analytics.

All set for the mass production stage
As the AI wave becomes the new normal in the tech industry, the demand for memory solutions has surged since the second half of 2025. XCENA is reportedly in touch with several global memory vendors. The company is eyeing the hyperscalers as its ideal customers. While these hyperscalers are spending tens of billions a year on AI infrastructure, XCENA believes, through MX1, even a small gain in memory efficiency will bring hundreds of millions in savings for these companies.

While the flagship product itself has been in the prototype stage, mass production is scheduled to roll off Samsung’s foundry lines by the 2026-end, with the company eyeing revenue flow from 2027 onwards. While neural processing unit (NPU) makers are looking to become the competitors for chipmaking giant Nvidia when it comes to training workloads, XCENA is taking a different but a crucial path altogether: targeting the memory-intensive layer that sits underneath all of the AI chip ecosystem.

It’s not like XCENA doesn’t have rivals. It needs to compete with Nasdaq-listed Astera Labs and Marvell, both working on next-generation memory connectivity. While Marvell is a large, established player, the start-up has thousands of cores. Based on publicly available specifications, Marvell’s approach relies on a handful of general-purpose cores by comparison. So XCENA is looking to win the game here through sheer numbers.

XCENA’s cores are built on RISC-V, an open-source chip design blueprint optimised specifically for data processing. Beyond the cores (tailored as small and efficient beings), XCENA designs its own internal memory hierarchy, interconnect bus, and DRAM controller and research and development activities, which its rivals mostly outsource.

The post Start-up of the Week: XCENA eyes solving AI’s “memory bottleneck” appeared first on International Finance.

While the Mark Zuckerberg-led Meta has been selling AI-enabled Ray-Ban glasses since 2023, Google is building Android XR, and Apple will be entering the niche market as well. Samsung, along with eyewear biggie Gentle Monster, is working on its first AI-capable smart glasses. China’s Huawei, Alibaba, and Xiaomi are not behind either.

The shipments of smart glasses, in 2025, stood at 8.7 million units, up by over 300% from 2024. As per the analysts, the sales data will cross the 15 million mark in 2026. And that’s where the South Korean start-up LetinAR, which has spent the last decade building the optical technology, the technology that powers the wearables, is looking to lead the ecosystem of suppliers and component makers of the niche technology.

The LG Electronics-backed start-up hit the headlines recently by securing funding worth USD 18.5 million from a consortium consisting of the likes of the Korea Development Bank and Lotte Ventures. LetinAR is also eyeing filing an IPO in South Korea by 2027.

In today’s episode of the “Start-up of the Week,” International Finance will talk in detail about LetinAR, whose CEO Jae-hyeok Kim and CTO Jeonghun Ha, also high school friends, are looking to redefine the innovation game at the smart glasses segment.

Powering the heart behind smart glasses

LetinAR’s expertise lies in making the optical module that makes the smart glasses work. The module, which has an appearance like a tiny lens component, projects images into the smart glass user’s field of vision. The lighter, thinner, and more power-efficient the optical module is, the better it becomes in terms of delivering a sharp and clear image.

So, the critical engineering task here is to get that particular single component right by making it small enough to fit inside a normal-looking frame. Kim, while interacting with TechCrunch, stated that his start-up’s specialization lies in executing the tough engineering challenge perfectly.

“We see AI glasses as that next platform. And the optical module is the hardest part to get right, as AI glasses makers will need a lens that is thinner, lighter, and more power-efficient than what exists today,” he said further.

If AI is going to redefine how the next generation of smart glasses is going to work, then LetinAR’s goal is simple here: be the company that the glass makers will rush to during the R&D phase.

Talking about the company’s key technology, we have PinTILT. The start-up defines the technology through these following words: “The hybrid structure consisting of advantages of both waveguide and birdbath AR optics provides better form factor, power consumption, and image quality, enabling next-level AR smart glass experiences.”

To remain useful throughout the day, an augmented reality (AR) device must possess specific characteristics, such as being lightweight, having a long battery life, providing a clear exterior view when not in use, and displaying a wide and clear virtual image that draws users in.

The PinTILT technology, as per LetinAR, has been tailored to meet exactly those requirements. The technology offers high light efficiency, resulting in lower power consumption for the display. This significantly reduces the battery’s weight, making the smart glasses lighter and allowing for longer wear.

PinTILT’s structure prevents information from leaking to external observers. At the same time, its high transmittance allows users to maintain eye contact naturally. Another plus point of PinTILT is the simplicity in its design. In AR optical lenses, enhancing one key specification often means a reduction in another. To increase the FOV (Field of View) or eye box, birdbath designs grow in form factor.

“EPE (Exit Pupil Expander) designs, on the other hand, can suffer from reduced optical efficiency, leading to dimmer displays or requiring larger batteries. However, our B-Type PinTILT technology stands apart. Beyond the minimal thickness required by the display’s vertical dimension, it presents virtually no other significant trade-offs,” the start-up remarked.

EPE (Exit Pupil Expander) technology also requires high-refractive-index materials for expanded FOV, leading to higher production costs and manufacturing challenges that impact the end product’s pricing aspects. PinTILT has reportedly solved the issue by utilizing a “Wide Total Internal Reflection Angle,” enabling a large FOV using only conventional refractive indices. The B-type PinTILT structure has further been tailored for eyeglass forms so that the end product scores significantly on the wearability front, making it ideal for everyday AR and AI glasses.

“PinTILT lenses are produced using a high-efficiency injection moulding process. Unlike birdbath and waveguide designs, which require precise assembly of multiple components, PinTILT lenses are assembled with minimal parts, offering high manufacturability. These lenses are made with an injection-moulded plastic structure that is solid inside, offering strong durability. As a result, there is no need for additional protective covers or rigid frames. The outer shaping is simple, making it easy to adapt to various designs depending on the desired eyeglass form,” the start-up continued further.

By making the technology available through A-Type and B-Type, LetinAR has unlocked state-of-the-art performance across a wide range of displays and application fields. The A-Type PinTILT for Maximum FOV has emerged as suitable for wide FOV applications by providing two to four times the field of view compared to the same size and same components.

“While birdbath designs increase in volume to achieve a wide FOV, waveguides are limited to using only the area below half of the total internal reflection angle for multiple internal reflections. PinTILT offers a wider total internal reflection angle, allowing for a significantly larger FOV,” the start-up said.

The PinTILT structure has been specifically designed to cater to the large set of consumer demands. These include the basic type (suitable for car HUDs), the A-type (perfect for goggles), and the B-type (well-matched for glasses).

Through PinTILT, LetinAR has focused only on the light that can actually enter the eye. The technology has been carefully engineered, keeping in mind the angle of tiny elements inside the lens. The start-up wants to ensure that the AR device using the PinTILT technology can produce a brighter image in a thinner, lighter form factor, using less power.

Backing PinTILT, we have Pin Mirror, a device smaller than a pupil, that makes things increasingly transparent for AR as it approaches the eye, much like how eyelashes do not obstruct vision. Additionally, adjusting the size, spacing, transparency, and other factors of Pin Mirrors can offer users a sharper and more precise view.

Rewriting the AR playbook

Using PinTILT and Pin Mirror, LetinAR has developed Letinus (formerly B30), a compact lens module that features a lightweight design and is ideal for the development of smart glasses that can be worn for extended periods of time, thanks to its 22-degree field of view (FOV) and low power consumption.

T-Glasses, on the other hand, is an EV-Kit in the form of glasses that utilizes the Letinus lens module produced through standardized processes. Then we have FrontiAR, the module, Letinus’ successor, that has been customized for developing wide-angle smart glasses with a wide field of view of 45 degrees, low power consumption, and a lightweight design.

Last but not least, KeplAR is an EV kit in the form of glasses that incorporates the FrontiAR, offering a wide field of view of 45 degrees. With low power consumption and a lightweight design, FrontiAR is suitable for developing applications in everyday and industrial environments.

Talking about the commercial success of LetinAR’s products, Japan’s NTT QONOQ Devices and Dynabook, formerly known as Toshiba Client Solutions, have emerged as the start-up’s customers, giving the company real manufacturing experience at scale. However, it is the deal with Aegis Rider, a Swiss deep tech company spun out of ETH Zurich’s Computer Vision Lab, that has presented the deep tech company its “make-or-break” moment.

Aegis Rider is building an AI-powered AR helmet that displays navigation, speed, and safety alerts directly in a motorcycle rider’s field of vision, not floating on the visor but anchored to the road itself. Using LetinAR’s module inside the helmet, Aegis Rider is targeting a 2026 launch date for the European Union and Swiss markets.

The post Start-up of the Week: Meet LetinAR, the venture powering AI glasses appeared first on International Finance.

Even the words we use (like cloud computing, artificial intelligence, and wireless networks) suggest that they exist in a world untethered from physical constraints.

In actuality, the global internet is real and physical. It is anchored to the ocean floor by a web of fibre-optic cables, which function much like the human central nervous system.

These subsea systems transport about 95% of all international digital communication and carry around 64,000 terabytes per second of global data. They also make possible $10 trillion worth of daily financial transactions.

Due to the tremendous demand for AI and the geopolitical rivalry between the US and China, the submarine cable industry has become a highly contested geostrategic front. The global subsea market is projected to reach $32.8 billion by 2026, with some estimates suggesting it could grow to $60.5 billion by 2036. This expansion is closely linked to the generative AI boom, which is driving significant growth; consequently, required bandwidth is projected to triple between 2022 and 2027, while overall demand for international bandwidth is expected to double every two years.

Satellite internet is expensive and can cause higher latency. However, submarine cables provide high-throughput, low-latency connectivity, which AI and training workloads demand. Therefore, the industry is predicting $13 billion in new undersea cable investments between 2025 and 2027. The projected capital expenditure has doubled from that of the preceding three years.

The hyperscaler era

There has been a shift within the subsea market as it’s changed from a carrier-led consortium to private ownership dominated by a few technology hyperscalers, such as Google, Meta, Amazon, and Microsoft.

These companies are constructing proprietary networks to control their data pathways directly. They already control the vast majority of bandwidth demand on the core trans-Pacific, trans-Atlantic, and intra-Asia routes, and are responsible for almost half of all new cables built since 2021.

The vertical integration into physical infrastructure helps hyperscalers reduce costs and expand capacity. They can dictate landing points to bypass congested or geopolitically sensitive nodes and implement bespoke optical switching technologies to optimise their global data centre interconnects.

Meta launched Project Waterworth, which is set to be the world’s largest private subsea cable system. At 50,000 kilometres in length, it is designed to connect the United States, India, Brazil, and South Africa.

Waterworth is rapidly expanding digital economies across the global South by utilising up to 24 fibre pairs to maximise data throughput via spatial division multiplexing. Interestingly, the project sidelines Europe-centric corridors, signalling that Meta is betting big on the global South as a core growth region for AI-driven connectivity.

FASTNET, the flagship cable of Amazon Web Services (AWS), is a transatlantic system from Maryland to County Cork, Ireland. Scheduled for 2028, it is designed to deliver over 320 Tbps of capacity. This is a bid to meet the unique demands of modern AI, which must balance the asymmetry between US-based model training and European data residency laws. Achieving this requires high capacity, low latency, and an ability to route around traditional choke points.

FASTNET avoids legacy cable clusters in the US Northeast and the UK. It also incorporates advanced optical switching technology that allows AWS to redirect data to future landing points as AI workloads evolve.

Then there is the whale among them. As the heaviest investor in privately owned submarine cable systems and the largest owner of submarine cable networks, Google owns significant subsea cables such as Dunant, which provides 250 Tbps, Grace Hopper at 352 Tbps, and Equiano. Additional systems like TPU, Nuvem, and Firmina are scheduled to come online in 2026. By creating alternate routes that isolate infrastructure from carrier-dependent nodes, Google has established a gold standard for hyperscaler infrastructure planning.

Infrastructure fragility and the Red Sea chokepoint

Hyperscalers can avoid certain geographic hurdles, but all global data traffic has a few specific maritime choke points. There is nothing more vulnerable and more critical to the internet than the Red Sea.

Around 70% of all global internet traffic and 90% of Europe-Asia data communication is believed to pass through a narrow corridor called the Bab al-Mandab Strait, which is just 26 kilometres (16 miles) wide. This concentration, this choke point, can become a catastrophic point of failure if left undefended.

In March 2024, the region’s fragility was highlighted when four undersea cable systems were severed, disrupting an estimated 25% of all data traffic between Asia, Europe, and the Middle East.

This issue re-emerged when the SMW4 and IMEWE systems near Jeddah in Saudi Arabia failed simultaneously, leading to significant latency spikes and degraded connectivity across the Middle East and South Asia. While Microsoft acknowledged the resulting traffic latency, Pakistan, India, and the UAE were forced to scramble to reroute data through secondary paths.

Repairing these essential assets in conflict-affected waters presents significant challenges. Insurance premiums for repair vessels have skyrocketed due to the presence of Houthi militants and Somalian pirates, leading to situations where cable breaks remain unrepaired for several months. A notable example of this occurred in March 2025, when the PEACE cable broke and remained offline for months on end.

These chronic vulnerabilities are now incentivising the development of new routes to ensure better stability. Key initiatives include Google’s Blue-Raman cable, which is designed to travel overland through Israel, and projects like Africa-1, which aim to bypass the Middle East altogether by utilising a route around South Africa.

China’s Digital Silk Road and the fibre-optic cold war

On one side, there are physical vulnerabilities for these subsea networks, while on the other, a macro-level geopolitical struggle exists for control over the cables. In 2015, China launched the Digital Silk Road (DSR) to export Chinese digital infrastructure to the Indo-Pacific and the Global South, leveraging financing to secure diplomatic alignment. The initial foray was conducted by Huawei Marine Networks, which successfully captured approximately 15% of the global market by 2019.

Following US sanctions, the entity was rebranded as HMN Technologies, and Hengtong Optic-Electric acquired an 81% stake in the company by late 2025. Hengtong has since become one of the top three global optical fibre manufacturers, controlling over 25% of the domestic Chinese market and 15% of the international market. The dominant market position is further strengthened by end-to-end vertical integration and a portfolio of over 5,000 patents.

The Chinese claim that they are trying to ensure stability so that trade happens seamlessly. However, a team of marine engineers from Lishui University (in Zhenjiang province across the coast of Taiwan) applied for a patent for ’a dragging type submarine cable cutting device’ in 2020. According to Newsweek, which inspected the patent, it was described as an ’ocean towing type cutting device’.

The Lishui University authors explicitly wrote: “With the development of science and technology, more and more submarine cables and communication cables are laid on the seabed in all parts of the world, and in some emergency situations, the cables need to be cut.”

Scientists Zhang Shusen, Dai Ying, Fu Changrong, Gao Zikun, Li Xuping and Ji Guangyao co-authored the document.

The patent was either rejected or retracted later, without providing an explanation.

The US responded to Chinese advancements through a comprehensive campaign to excise Chinese state-linked firms from the global subsea ecosystem.

Marsha Blackburn, a Republican Senator, said, “Undersea cables are a critical component of our digital economy and national security. If we let hostile actors control or threaten that infrastructure, we are effectively surrendering a key lever of global influence.”

Washington systematically dismissed Sino-American cable partnerships. The Pacific Light Cable Network (PLCN) was forced to drop its Hong Kong-linked leg and reroute capacity via Taiwan and the Philippines. Regulators in the US had warned that the original configuration could place sensitive data under Chinese jurisdiction. The GAP-1 system, a trans-Pacific cable involving Amazon, Meta, and China Mobile, was effectively shelved in 2023. This left hundreds of millions in construction costs stranded after China Mobile withdrew amid geopolitical pressure.

The confrontation with the most consequences happened with the SeaMeWe-6, a 19,000-kilometre system which links Western Europe and Southeast Asia. HMN Technologies made a bid that was three times cheaper than Western competitors.

However, the US State Department used diplomatic pressure and warned of serious sanctions and a ban on American purchasing capacity on the line if HMN won the contract. The US Trade and Development Agency offered financial incentives to steer countries towards American suppliers.

The pressure succeeded, and the contract was awarded to SubCom for $600 million (about $130 million more than HMN’s bid).

China retaliated and withdrew its 20% funding from SMW6, and announced a parallel Europe-Middle East Asia cable that mirrored the same route but was built exclusively by Huawei. Now there is a structural bifurcation with the US suppressing Huawei’s share of planned global cable contracts to around 10%. This is far behind France’s Alcatel Submarine Networks at 41% and SubCom at 21%.

Huawei has been excluded from many Western consortia, and Chinese firms are systematically building parallel networks with China, Russia, Pakistan, and allied African and Middle Eastern states. Analysts have found that this fragmentation and decoupling between great powers is leading to the fragmentation of the internet into eastern and western blocs, creating a techno-nationalist paradigm where political alliances are more important than network efficiency.
India’s ascension and the Question of resilience

The biggest winner in this geopolitical conflict is India, which is slowly becoming a global hub for data. All the rerouted traffic coming away from the South China Sea and the Red Sea is finding itself in India.

India has 950 million internet users, and its digital economy is expected to reach 20% of GDP by 2027. Anil Kumar Lahoti, Chairman, Telecom Regulatory Authority of India (TRAI), said, “India’s data transmission capacity is set to quadruple with new undersea systems, turning the country into a critical junction between Europe, the Middle East, and Asia.”

Without this cable overhaul, India cannot anchor the AI driven workloads of the next decade.

Reliance Jio’s India-Asia-Express (IAX) and India-Europe-Express (IEX) systems are contributing over 200 Tbps to this growth. Mumbai hosts at least 14 cable landing stations.

Despite all the money that’s been poured into these projects, they are still dangerously vulnerable. Over 80% of all cable faults occur in shallow waters because of commercial fishing trawlers and anchoring, or mundane accidents with devastating consequences.

In an era of heightened tensions, accidents and deliberate attacks are in a grey zone. Anchor dragging by state-aligned vessels has disrupted cables in the Baltic Sea and near Taiwan, though it is hard to prove that it was done deliberately.

There is also a shortage of dedicated repair vessels around the world, which makes the threat even worse. Most repair vessels are Chinese-owned, and there are fewer trusted Western ships, which makes Americans and Europeans wary.

Security analysts worry that passive data extraction devices could theoretically be inserted into cables. Even if such an event were to happen, no one would know.

The great bifurcation

The global submarine cable market is now a logistical necessity for the telecommunications industry. Hyperscalers are internalising the internet’s physical layer to feed the insatiable data requirements of artificial intelligence, fundamentally altering the economics of global bandwidth.

In 21st-century geopolitics, physical choke points (like the Red Sea) have demonstrated how asymmetric threats are and how ill-equipped repair fleets can be, to the detriment of intercontinental connectivity. Washington has been campaigning to block Chinese firms from Western networks, and has been trying its best to stop the Digital Silk Road.

But this has structural costs. As the bifurcation of the subsea architecture into politically aligned spheres continues. The physical cables that once supported the globe are now being instrumentalised to divide it.

The post Fight for control of the internet that we cannot sea appeared first on International Finance.

The Fourth batch successfully raised its 108 million Series A. Laqua said that by accumulating funding worth USD 268 million to date, Corgi has become Y Combinator’s latest unicorn.

Corgi’s co-founders, Laqua and Emily Yuan, were part of YC’s Spring 2024 batch. The start-up, established in the same year, offers coverage for general liability, cyber liability, tech and AI liability. The venture now wants to use the funding raised from its Series B to expand into more lines of insurance while building a generational company.

Building Thoughtful Insurance Products

According to Corgi, most insurance providers don’t sell their own policies or coverages, which means they also do not pay out their customers’ claims out of the companies’ pockets. This type of insurance is slower because they have to connect with a third-party financial entity to pay out claims. The start-up has disrupted the practice by emerging as an AI-native, full-stack insurance company that creates and sells its own policies, in addition to paying out customers’ claims when an issue arises. These elements, leveraged with AI, allow us to provide cheaper, better, and faster insurance for everyone.

“Our goal is to rebuild the USD 1 trillion-plus insurance industry from the ground up. We’re not just making a company; we’re engineering financial infrastructure for the next hundred years,” the start-up said.

As an AI-native, full-stack insurance company, Corgi has specialised itself in providing insurance to start-ups through scalable packages and customisable coverage solutions.

Every start-up, at its pre-revenue or seed stage, requires basic compliance to hire or lease space. Corgi’s answer to this is “Pre-Seed & Seed”, which addresses expenses related to office leases and initial hiring. This package provides foundational “General Liability and Business Property” coverage to protect against everyday operational risks. Tailored for business founders looking to meet landlord requirements and protect their initial assets while staying lean.

Next is “Series A”, which has been designed for high-growth tech companies. The package, apart from scaling with the start-ups’ funding rounds, also includes Directors & Officers (D&O) Insurance (a requirement for most Series A boards) and Tech E&O / Cyber Liability to satisfy enterprise vendor contracts and SOC 2 compliance standards.

For the growth stage, Corgi’s most comprehensive commercial insurance bundle offers high-limit protection, including Employment Practices Liability (EPLI) for large-scale hiring and fiduciary liability for employee benefit plans. The plan mitigates complex risks across multiple jurisdictions and ensures the companies remain “IPO-ready” with advanced management liability.

The “Custom Package” allows a start-up to tailor its coverage limits and policy types to specific industry risks. Whether the business needs specialised “AI Liability”, “Media Liability”, or specific international coverage, Corgi’s advisors help them build a bespoke risk management strategy that fits their exact contract requirements.

Solutions For Industries

Among its prominent insurance products for industries, let’s start with “Industry for SaaS”, tailored for companies facing digital risks that require software-oriented modern liability coverage. Through the mechanism, SaaS (Software as a Service) can meet insurance requirements in enterprise MSAs (Master Service Agreements), apart from gaining investor trust by providing the protection required for the company’s Series A term sheet and board. The insurance also protects its customers from incidents like system-wide outages, bugs, or implementation failures by reducing the monetary fallout.

For AI start-ups, Corgi’s insurance mechanism provides protection against scenarios like model performance and hallucination (liability for when an LLM provides false, defamatory, or harmful information that causes a third party loss), algorithmic bias (protection against claims of discriminatory outcomes in hiring, lending, or healthcare AI) and training data disputes (legal defence for intellectual property disputes related to the data used to train your proprietary models).

Next in line is “Fintech Insurance”, which blends risks like technology performance, multiple partner and stakeholder expectations, and regulatory scrutiny. A bug in your API prevents users from accessing features, balances, or transactions—partners and customers demand answers fast. Fintechs face constant worries like mistakes in payment routing, reporting, or product configuration, leading to allegations of customer harm or operational losses, with either customers, banks, processors or vendors knocking on companies’ doors with their monetary claims. Corgi acts like a shield in these situations by providing coverage against scenarios like professional liability arising from technology products or services; claims arising due to hacking, ransomware, and data privacy-related incidents; and risks associated with content, advertising, and intellectual property.

Also, “Marketplace Startup Insurance” from Corgi acts as the comprehensive insurance for two-sided marketplaces and platforms by protecting client platforms and their users and ensuring company growth from seed to IPO.

The insurance has three coverage areas: protection for claims involving bodily injury or property damage connected to platform-facilitated transactions, liability allegations tied to screening, verification, onboarding, or moderation workflows and claims tied to outages, ranking changes, listing errors, payment flow issues, or other product failures that impact users.

The post Start-up of the Week: Through tailored products, Corgi powers new businesses appeared first on International Finance.

The legacy mechanisms of the logistics industry are known for having tedious back-office processes and siloed data, along with a lack of operational and financial visibility, factors that as a combination, drain companies’ profits and increases consumer costs. Loop wants to ease up the pressure, through the technology’s helping hand.

Loop’s founders, Matt McKinney and Shaosu Liu, are Uber veterans. In Uber Freight, McKinney tackled the problem of building a new structure for the legacy transportation payment system, while Liu handled the building process of a software that can optimize a network of networks to build a more collaborative supply chain.

The assignments gave the duo a first-hand experience of the “legacy issues” bothering the global supply chain ecosystem, giving them the inspiration to come up with Loop, about which International Finance, in today’s episode of the “Start-up of the Week,” will talk in detail.

Redefining the next-generation supply chains

As per the Loop, enterprises require a more dependable data foundation to sustain high operational performance, free up working capital, and take strategic decisions with confidence. To enable this, the start-up, through its “Logistics Data Platform,” has automated 99% of freight and parcel audits to deliver efficient, accurate outcomes.

Loop stores important supply chain and spend documents (rate tables, manual calculations and invoice) in one platform, before automatically auditing every line-item on the invoice and identifying discrepancies in hours.

Another headache logistics players face is the “black box,” the event in which shipment, after leaving the dock, leaves businesses powerless to prevent delays or manage customer expectations. To address this, Loop, using the “click-to-porch visibility” approach, lets its online platform absorb over 20 billion scans annually across all modes and carriers, standardizing messy data into a single, accurate view.

With a wide range of custom dashboards and reports, logistics players get to track every package in real-time to monitor network velocity, confirm delivery status, and ensure their shipments are moving as promised. Loop’s predictive “at-risk” logic also flags shipments that are stalling, dwelling, or projected to miss their commit date, way before the carrier even reports a delay.

The start-up’s customizable dashboards provide a real-time command center for logistics operations, tracking critical parameters like on-time delivery percentages, carrier service levels, and dwell times.

When it comes to logistics companies securing the best possible terms and rates, Loop prepares them with data-driven insights and veteran carrier intelligence services. The start-up’s “parcel contract optimization” combines deep shipping data analysis with decades of industry expertise to secure better rates and maintains optimal contract performance over time.

Loop’s contract intelligence dashboards provide real-time visibility into a logistics company’s contract performance, breaking down costs by service type, zone, and accessorial charges to reveal optimization opportunities at a glance. Interactive analytics let the business drill down into specific contract terms, track realized versus contracted discounts, and monitor minimum charge utilization to ensure the business is maximizing every negotiated benefit. The start-up claims its contract intelligence and industry expertise delivering up to 50% post-audit savings for logistics players by optimizing their rates.

Loop’s contract intelligence dashboards come equipped with built-in tools, that show the client companies how proposed discount changes will directly impact their real-world financials. The “what-if” scenario modelling allows the businesses to simulate the precise impact of different discount structures or minimum charge adjustments against their historical shipping data. They get to see the potential savings or costs of the proposals, before agreeing to the terms and conditions.

Loop’s “Carrier Invoice Pay” streamlines global AP (accounts payable) process with accurate, seamless freight payments, with flexible (including custodial) options reducing administrative costs and improving working capital. No more of logistics professionals sorting invoices, chasing approvals and answering “where is my money?” calls from carriers, as the start-up takes over these repetitive functions, through its AI-native audit, which guarantees every invoice is 100% accurate before it’s approved. From there, Loop’s flexible payment solutions provide logistics companies the control, flexibility, and automation they want.

A full stack of cutting-edge tech

Loop’s “Control Suite” puts logistics players in command, with tailored policies assisting the industry players to get a real-time visibility into their financial health, network performance, and policy compliance. We are talking about a business automation an audit-related engine, that helps its users to seamlessly review their organizational rules for financial approvals, service-level enforcement and required documentation.

The heart of Loop’s “Control Suite” is the “Control Tower,” an operational and financial intelligence platform, that monitors shipment volumes, spend trends, and compliance performance in real-time, before identifying the root cause behind the failure of the company policies.

Another brilliant piece of Loop’s innovation has been its AI-native platform, “Loop AI,” that has been customised with various AI models and agents, to fix messy logistics data. This approach delivers what legacy systems can’t: comprehensive data capture, leading data quality, and a path to end-to-end automation.

“DUX,” Loop’s another AI-based solution for logistics businesses, uses high-quality data, creating a complete and accurate shipment data, helping supply chain teams to move from reactive exception management to proactive optimization method.

The post Start-up of the Week: Through AI, Loop redefines logistics industry appeared first on International Finance.

The post Start-up of the Week: Mariana Minerals brings automation to mining sector appeared first on International Finance.

Artisanal mining has been a job generator on a global level, and in Africa, it often operates informally on or near company-run mines, hitting their profits, spreading pollution and depriving nations of revenue. To address this, the proposed “stakeholder prosperity bond,” according to Rob Karpati, Veridicor’s finance director, aims to professionalise artisanal mining.

The instrument will link investor returns to predefined social and environmental outcomes for workers, communities and host economies rather than output.

“The debut issuance would raise between USD 100 ‌million and USD 200 million by ⁠year-end to help ⁠Metalex Commodities integrate artisanal and small-scale miners through regulated offtake agreements as well as shared infrastructure and equipment investment,” Reuters reported.

“Potential investors include European sustainability bond funds, impact and ‌mining investors, banks and wealthy individuals focused on sustainability,” ⁠the firms said.

“Zambia, Africa’s second-largest copper producer, hosts tens of thousands of artisanal miners, including around Metalex’s northwestern permit. Large mines tend to be the anchor of these because it’s got to go on someone’s balance sheet,” Karpati said.

“They end up gaining financially because they get offtake from it, and the artisanal miners gain financially because it’s a fair price, not some predatory intermediate,” the official added.

“Industrial mines would sit at the centre of each bond structure to support repayment, while sustainability-linked terms would adjust interest rates based ‌on social and environmental performance,” Karpati remarked.

Metalex founder and chief executive Ayo Sopitan said the bond would allow the company to run large programmes that integrate artisanal miners into its supply chain.

“We plan to source around 30% of our ore from trained, licensed local miners. The bond lets us do that at a much ‌larger scale than our balance sheet alone would allow,” he said.

The post Africa’s first stakeholder prosperity bond: All you need to know appeared first on International Finance.

The investment in the Canadian company will allow the Italian energy group to enter the critical minerals value chain, as European companies seek to reduce their dependence on China-sourced materials.

“The capital increase is aimed at supporting the development of the Matawinie Mine of high-quality natural graphite, a key feedstock for the battery sector, as well as other high-margin industrial segments. The Matawinie Mine forms the asset base for a vertically integrated project of natural graphite extraction and its refining at NMG’s Becancour Battery Material Plant,” ENI said.

“The investment in NMG is consistent with Eni’s strategy to diversify its supply chains. In particular, it enables Eni to enter the critical minerals value chain through a partnership with a leading company in the sector, while leveraging its distinctive technological know-how,” it added.

According to Eni, the transaction will also give it the option to negotiate exclusive supply agreements for graphite and other materials to support its gigafactory initiative to produce stationary lithium batteries in southern Italy.

The investment forms part of a USD 297 million capital increase at Nouveau Monde Graphite, alongside Canadian institutional investors ‌Canada Growth Fund and Investissement Quebec, as well as a public equity raise.

Following completion of the transaction, Eni will hold about 11.5% of Nouveau Monde Graphite’s share capital, in addition to gaining a seat on the company’s board.

The acquisition also serves one of the critical bedrocks for Eni’s goal to become a player in battery cell manufacturing for stationary energy storage systems. In September 2025, it started the development activities for a new lithium iron phosphate (LFP) battery cell factory through a joint venture (JV) called Eni Storage Systems.

Eni Storage Systems will construct a manufacturing hub at existing Eni facilities in Brindisi, southern Italy, with over 8GWh annual production capacity. The pre-construction work, which includes an initial engineering phase, along with the economic, financial and permitting assessments, is expected to be completed this year, after which the project will enter the execution stage.

Eni’s JV partner is FIB, a subsidiary of Italy’s Seri Industrial group, which owns several battery and materials companies. In October 2024, both ventures decided to cooperate in developing an industrial supply chain for LFP batteries, targeting the stationary battery energy storage system (BESS), along with the commercial and industrial (C&I) electric mobility sectors. The joint venture wants to capture more than a 10% share of the European stationary energy storage market.

The post Eni to acquire stake in Nouveau Monde Graphite appeared first on International Finance.

Other significant players were Embark Ventures, The Grantham Foundation for the Protection of the Environment, Massachusetts Clean Energy Centre (MassCEC), Safar Partners, Climate Investment and Marathon Petroleum Corporation. The venture will now use the fresh capital to scale its business, apart from financing the deployment of a modular filtration platform into the refining and chemical sectors.

Via Separations’ expertise lies in providing filtration systems designed to lower energy use in industrial separation processes by up to 90%, apart from significantly reducing emissions in industrial processing.

Electrifying Heat-based Separation

“Via Separations is a membrane technology company driving the transition from expensive, energy-intensive thermal separations to efficient, lower-cost filtration. Via is operating at commercial scale, delivering value to industrial customers today with no green premium. Via Separations is a membrane technology company driving the transition from expensive, energy-intensive thermal separations to efficient, lower-cost filtration. Via is operating at commercial scale, delivering value to industrial customers today with no green premium,” the start-up explained itself through these words.

Via electrifies heat-based separation with modular filtration systems that integrate directly with existing industrial equipment, reducing the energy required for chemical separations in the process.

“These thermal separation steps account for roughly 12% of global energy use, driving significant fuel and steam demand across industrial separations. By replacing them with a mechanically driven membrane process, Via’s system can reduce energy use at the separation step by up to 90%, delivering lower operating costs, higher uptime, and a more flexible pathway to energy efficiency and electrification,” Via Separations remarked.

Via Separations’ membrane technology uses up to 90% less energy than traditional evaporation or distillation because it is a physical separation, rather than a thermal process. In the start-up’s language, “these membranes work like a coffee filter or pasta strainer, but for chemicals.”

These materials are made of a unique source called graphene oxide (GO), which is extremely stable in nature, ensuring the membranes withstand high temperature and corrosive process conditions, where typical polymer membranes don’t work at all. Via Separations has already proven the membrane technology at commercial scale in the pulp and paper sector, approaching two years of continuous operation at a Canadian pulp mill. The start-up is now expanding the technology’s deployment into refining and chemicals, with hundreds of millions of dollars of capital projects in the commercial pipeline. In 2025, the company also completed a pilot at a major Gulf Coast refinery.

The Via Ecosystem

Through its filtration systems, the start-up is driving both bottom- and top-line improvements for its industrial customers, reducing energy costs, apart from providing operational flexibility and de-bottlenecking production.

The start-up has tuned its durable graphene oxide membranes to perform challenging industrial separations across markets. The start-up continued, “Complete system integrates Via membranes into customers’ existing processes in a compact footprint. Via systems deliver quality separations while reducing costs, energy, and production bottlenecks.”

A very good example of Via Separations’ industrial innovation has been its first commercial-scale Black Liquor Concentration System (BLCS), which is operational at the International Paper site in Grande Prairie, Alberta, Canada. The global pulp and paper industry, known for annually producing over 400 million tonnes of paper, packaging, and tissue from wood fibres and recycled materials, faces one challenge: black liquor recovery. We are talking about the most capital and energy-intensive component of a Kraft pulp mill, which burns waste cooking liquor (12%-15% solids) in a specialised boiler to produce energy (steam/electricity) and recover inorganic cooking chemicals.

To address the challenge, Via has created Black Liquor Concentration System (BLCS), which concentrates weak black liquor before evaporation, translating to cost savings and operational benefits for pulp mills.

“Via’s Black Liquor Concentration System (BLCS) displaces steam use in evaporators using a reverse-osmosis-like process to directly remove hot, clean water from weak black liquor (WBL). The compact BLCS integrates into existing mill footprints to concentrate WBL up to 40% solids,” the start-up said.

Via Separations’ efficient concentration process eliminates production bottlenecks and reduces the energy consumption of black liquor concentration by up to 50%. The technology, if widely deployed, will help pulp mills realise additional free cash flow in the millions of dollars per year.

In the petrochemical industry, despite heavy crudes becoming more prevalent, refineries are facing limited capabilities when it comes to processing the fuel due to the size of the vacuum distillation unit (VDU). To solve this, Via filtration system is increasing the VDU capacity, reducing costs and energy consumption while unlocking additional heavy crude processing capacity.

Via’s innovations have been expanding in the chemical manufacturing industry as well. Acid processing is a large market that touches both the chemicals and refining sectors. Knowing the potential, Via is looking to provide on-site processing of sulfuric acid, reducing costs and emissions while enabling recovery of acid-soluble oils (ASOs).

The post Start-up of the Week: Via Separations secures funding to deploy modular filtration appeared first on International Finance.

To integrate Islamic finance into its domestic economy, the country will introduce several key instruments like Murabaha (financing customers through instalment credit sales), Mudaraba (profit-sharing investments or fund attraction), Wakala (providing or attracting funds via agency agreements), Musharaka (financing clients through joint business activities), Salam and Istisna (financing through advance payments for goods) and Islamic leasing (Ijara), which will provide property under Sharia-compliant lease terms.

To support the adoption of these tools, the government will implement specific tax exemptions. While value-added tax (VAT) will not be applied to the markup on goods sold via Murabaha (Sharia-compliant financing structure, often called ‘cost-plus financing’), income generated from investment deposits will be tax-exempt as well. Furthermore, Islamic leasing agreements will be legally equivalent to financial leasing and traditional leasing.

To ensure systemic management and compliance with Sharia standards, the Central Bank of Uzbekistan will have its Islamic finance council. Additionally, banks providing these services will be required to form their own internal councils.

The panel, while operating under the Central Bank of Uzbekistan, will be tasked to develop industry standards, draft regulatory legal acts, provide clarifications on disputed issues, review contracts and internal documentation and ensure overall compliance with Islamic financial principles.

The latest policy move follows the Central Asian country’s Senate’s approval of the law on the introduction of Islamic banking activities in February 2026, marking a significant step toward modernising the nation’s banking sector.

Officials also proposed additional plans, such as establishing bodies like the Tashkent International Financial Centre and the International Centre for Digital Technologies. These will infuse Islamic finance mechanisms into the country and help Uzbekistan position itself more competitively in the global economy amid rising geopolitical uncertainty and intensifying competition for foreign investment. Officials see the country’s natural resources, economic potential, and ongoing reforms as the main engines for creating favourable conditions to attract international companies exploring new markets.

Tashkent International Financial Centre will likely serve as a platform for new investment flows. By 2030, it is projected to attract an additional USD 20-25 billion, contributing up to 1% of Uzbekistan’s annual GDP growth, in addition to creating as many as 15,000 jobs.

The centre will operate under a special legal regime, incorporating elements of the common law system of England and Wales, thereby allowing its governing bodies to adopt independent regulations. The platform will also have a Tashkent International Commercial Court and an International Arbitration Centre to handle disputes, while providing investors with benefits like tax incentives, simplified visa procedures, the capability of freely moving and repatriating capital, and access to modern financial instruments, including digital assets.

The International Centre for Digital Technologies, on the other hand, will operate under the “Enterprise Uzbekistan Brand.” The centre will function under a special legal framework, expected to remain in place until 2100. Within a regulatory sandbox, companies will be able to test new technologies, pay salaries in foreign currency, and operate under international labour and data standards.

The digital centre will also focus on AI, data processing, research and development, and startup support. By 2030, it is expected to attract up to 1,000 companies, create over 300,000 jobs and generate export revenues of up to USD 5 billion.

The post Uzbekistan’s Islamic financial framework: All you need to know appeared first on International Finance.