DD
MM
YYYY

PAGES

DD
MM
YYYY

spot_img

PAGES

Home Blog Page 9

Jensen Huang Tips Marvel as The Next Trillion-Dollar Company, Shares Rise

0

Marvell Technology’s rise as one of the biggest beneficiaries of the artificial intelligence boom received a powerful endorsement this week from Nvidia chief executive officer Jensen Huang, who declared that the semiconductor company could become the next trillion-dollar enterprise.

The remarks, delivered during an onstage discussion with Matthew Murphy at the Computex technology conference in Taipei, immediately boosted investor confidence in Marvell’s long-term prospects. Shares of the company climbed sharply in premarket trading after Huang highlighted Marvell’s growing importance in the infrastructure powering the AI revolution.

Huang’s comments carry significant weight in the semiconductor industry. Nvidia has become the dominant force in AI computing, and its chief executive is widely regarded as one of the most influential voices shaping the future direction of the sector. His endorsement suggests that the next phase of AI growth will not be driven solely by companies building processors, but also by firms providing the networking technologies that connect vast computing systems together.

“When you take a computing problem, and you disaggregate it into a lot of parts, and you distribute it across the entire data center, what’s necessary is connectivity,” Huang said. “That’s the reason why Matt’s doing so well. That’s the reason why Marvell is so essential.”

He added: “We’ve distributed and disaggregated computing so that it runs across these enormous clusters, so that we could get aggregating the total compute, the total memory, the total bandwidth that we have, and what makes it possible is connectivity.”

While attention has largely focused on Nvidia’s graphics processing units, powerful AI models are placing enormous demands on the networks that link thousands, and sometimes tens of thousands, of chips together inside modern data centers.

Why Huang is Betting on Marvel

As AI workloads scale, moving data efficiently between processors has become almost as important as the processors themselves. Industry executives describe networking as one of the most critical bottlenecks in the development of next-generation AI systems.

That trend has placed Marvell in an enviable position.

The company designs a range of high-performance semiconductors that enable data to move rapidly across cloud computing systems, AI clusters, enterprise networks, telecommunications infrastructure, and automotive platforms. Its portfolio includes networking chips, custom silicon, optical interconnects, and data-center connectivity solutions that are becoming increasingly essential as AI systems grow larger and more complex.

Marvell has emerged as one of the leading suppliers of custom AI chips and networking technology at a time when hyperscale cloud providers are pouring unprecedented sums into AI infrastructure.

The scale of that spending is staggering. Technology giants, including Microsoft, Amazon, Alphabet, and Meta, are collectively investing hundreds of billions of dollars to build AI data centers. Nvidia executives have repeatedly estimated that annual AI infrastructure spending could approach or exceed $1 trillion in the coming years.

For Marvell, that spending cycle is creating a substantial growth opportunity. The company delivered stronger-than-expected results in its fiscal 2027 first quarter, reporting revenue of $2.4 billion and projecting continued growth driven largely by its data-center business. Demand for AI-related products has become the primary engine of expansion, helping offset weakness in some traditional semiconductor markets.

Another significant vote of confidence has come from Nvidia itself. The AI chip leader recently committed a $2 billion investment in Marvell, deepening ties between the two companies as they seek to address one of the industry’s most pressing challenges: moving ever-larger volumes of data across AI systems.

The investment also aligns with a broader industry push toward photonics, a technology that uses light rather than electrical signals to transmit information. Photonic systems are viewed as a potential breakthrough for AI infrastructure because they can deliver higher speeds, lower power consumption, and greater efficiency than conventional networking technologies.

As AI models become more sophisticated and computationally demanding, photonics is increasingly seen as a critical technology for overcoming bandwidth and energy constraints inside future data centers.

Marvell’s networking and connectivity products are already widely deployed across cloud and telecommunications infrastructure. As data-center operators search for ways to improve performance while managing soaring energy consumption, Marvell’s expertise in high-speed interconnect technologies could become increasingly valuable.

Huang’s trillion-dollar prediction may sound ambitious, but it reflects a broader market belief that AI infrastructure winners will extend well beyond chip manufacturers themselves.

Nvidia has already crossed the trillion-dollar threshold and remains one of the world’s most valuable companies. Other firms tied to AI infrastructure, including memory-chip producers, networking specialists, and data-center equipment suppliers, have also seen their valuations surge as investors bet on years of sustained AI spending.

Marvell is now increasingly viewed as part of that group.

The company occupies a strategic position between AI processors and the networks that connect them, giving it exposure to one of the fastest-growing segments of the semiconductor industry. If AI investment continues at its current pace, demand for advanced networking, connectivity, and custom silicon solutions could rise alongside demand for computing power itself.

That is the future Huang was pointing to in Taipei: a world where the success of AI depends not only on the chips performing calculations, but also on the technologies enabling those chips to communicate at unprecedented speed and scale.

In that environment, Marvell’s role becomes much larger than supplying components. It becomes part of the foundational infrastructure of the AI economy, a position that explains why Nvidia’s chief executive believes the company could eventually join the trillion-dollar club.

The Diamond Bank’s Lecture on Win-Win and Co-opetition

0

When communities come together, they outperform.

Usain Bolt’s individual world-record performance in the 100 meters was extraordinary. Yet, when he ran the 4x100m relay with fellow Jamaicans, the team often delivered a faster average pace per runner than their individual performances suggested. There is something powerful about collective purpose. When people align around a shared mission, the whole becomes greater than the sum of its parts.

I learned a version of that lesson many years ago as a young banker in Lagos at Diamond Bank; yes, that bank treasured in the heart of Ndubuisi. During our training school, one of the bank’s distinguished leaders, Ben Oviosu, introduced us to the concept of Win-Win. The following day, another respected leader, Chinedu Uzoho, took us through Stephen Covey’s 7 Habits of Highly Effective People. By the time the program ended, we were no longer simply engineers, accountants, economists, or scientists. We had been transformed into bankers, equipped with a broader understanding of institutions, markets, relationships, and value creation.

To this day, Diamond Bank Training School remains one of the most consequential professional experiences of my life. It was a transduction process, a movement from one state to another, intellectually and professionally. The treasury lectures delivered by Ohis Ohiwerei were legendary. For hours, he deconstructed the mysteries of finance and markets with uncommon clarity. I asked so many questions that he eventually gave me a nickname: “Prof.” The name stuck throughout my years in what I still consider one of the finest banks Nigeria ever produced, even if it did not ultimately live forever.

Yet beyond the technical lessons, there was a deeper philosophy running through the training: Win-Win and Co-opetition. The message was simple. Build not only for your institution but also for your industry. Before joining the bank, many of us viewed Zenith, GTBank, STB, and other competitors as adversaries. By the end of the training, we understood something more sophisticated. Yes, we competed. But we also shared a collective responsibility to advance the banking industry. Competition and collaboration were not mutually exclusive; they were complementary forces.

That is why I find recent developments in the American technology sector fascinating. When Intel began struggling, many stakeholders across the semiconductor ecosystem recognized that the decline of Intel would have implications far beyond one company. Intel was not merely a corporation; it was part of the strategic infrastructure of American technology leadership. So the ecosystem responded. Capital flowed. Partnerships emerged. Support arrived. The industry effectively performed a Win-Win exercise to preserve a critical pillar of its technological architecture.

But co-opetition is not charity. Once Intel regained stability, competition resumed. Nvidia, one of the companies that helped strengthen the broader semiconductor ecosystem, is now entering markets traditionally dominated by Intel, including Windows-based computing platforms. The same companies that cooperate to strengthen an industry can compete vigorously within it. That is the essence of co-opetition: collaborate where collective success matters and compete where innovation demands it.

Good People, Africa must learn this lesson. When an important company begins to fade, the strongest players should sometimes ask a bigger question: what happens to the ecosystem if this entity disappears? There are moments when preserving the platform creates more value than maximizing immediate competitive advantage.

The future belongs to ecosystems, not isolated firms. And the most successful ecosystems understand a powerful truth: cooperation and competition are not opposites. They are partners in the advancement of industries, nations, and civilizations.

Sure, creative destruction must happen but creative re-construction must not be overlooked!

Brass to Shut Down Independent Operations, Migrate Customers to Paystack MFB

0

Nigerian fintech Brass, has announced that its business banking operations will transition into Paystack Microfinance Bank (MFB), marking a significant new chapter in its mission to serve African businesses with stronger financial infrastructure and expanded capabilities.

As part of the transition, the fintech will cease operating as an independent entity, with its business banking services being integrated into Paystack MFB, a licensed microfinance bank.

The move is expected to provide customers with access to a more robust banking platform equipped with deeper infrastructure and a broader range of financial products.

According to Brass, Paystack MFB offers many of the services that its customers currently depend on, including account management, operational support, payouts and expense management, fast and reliable money transfers, and financial insights.

In addition, customers will gain access to a wider suite of capabilities designed to support business growth and efficiency.

Founded in 2020, Brass set out to address a fundamental challenge facing African businesses: the difficulty of accessing simple, efficient banking services.

The company was built around a straightforward mission to make it easier for businesses to open accounts, move money, pay employees, and gain visibility into their finances without the bureaucracy and inefficiencies that often characterized traditional banking.

Recognizing that many Nigerian businesses struggled with excessive paperwork, high fees, and outdated banking systems, Brass introduced a modern business banking platform that allowed users to open accounts within minutes from any device, eliminating the need for physical branch visits.

The platform also offered tools for payments, payroll management, expense tracking, and cash flow monitoring, tailored to the needs of startups and small and medium-sized enterprises (SMEs).

Over the years, Brass earned the trust of thousands of founders and businesses across Nigeria, becoming a key financial partner for companies seeking a more seamless banking experience.

However, like many fintech startups operating in Africa’s challenging business environment, the company faced significant hurdles that tested both its operations and team.

In May 2024, Brass entered a new phase of its journey when a consortium led by Paystack, alongside PiggyVest, Ventures Platform, and P1 Ventures, acquired the company.

Speaking on the acquisition, the company wrote,

Over the months following the acquisition, our team, led by Philip Obosi and Yvonne Obike, entered a focused phase of rebuilding. We overhauled our internal systems and operational processes from the ground up, making sure that businesses that relied on us could manage their accounts, run recurring payments, organise team spending, and track their finances with greater clarity and reliability.

“As we rebuilt and as our platform became more mature, something became increasingly clear. The next phase of our growth could not be achieved alone. It required deeper infrastructure, broader capabilities, and a platform already trusted by businesses across Africa”.

The acquisition was aimed at strengthening Brass’ foundation while exploring opportunities to integrate its offerings within the broader Paystack ecosystem and expand its impact on African businesses.

The merger of Brass into the Paystack system, comes as there are calls for struggling startups to consider mergers instead of shutting down.

By combining resources, capabilities, and market strengths, struggling companies can preserve value, improve operational efficiency, and enhance their chances of surviving the current economic headwinds.

The migration of Brass customers to  Paystack Microfinance Bank will take place gradually between now and July 31, 2026, with Brass aiming to move interested merchants to Paystack MFB in the least disruptive manner possible.

The company stated that all Brass customers will be contacted directly and provided with clear guidance on the transition and the steps required to ensure a smooth migration.

Notably, for Brass customers, the transition is expected to unlock access to a more comprehensive suite of financial services backed by Paystack’s growing ecosystem and banking capabilities.

The move also positions Paystack MFB to deepen its presence in the business banking segment, where demand for reliable financial tools among startups and SMEs continues to grow

How Continuous Yield Accrual and Intraday Ownership Are Reshaping Institutional Cash Management

0

In conventional money market funds and most traditional cash-equivalent instruments, yield accrual is tied to discrete accounting boundaries. The day is treated as an indivisible unit, and beneficial ownership is typically determined at a specific cutoff—often the end of the trading session or a defined valuation point.

Under this framework, interest is effectively binary: you either own the fund at the snapshot time and receive the full day’s accrual, or you do not and receive nothing. This system is operationally efficient for legacy financial infrastructure, where batch processing, reconciliation cycles, and end-of-day NAV calculations dominate.

However, this model introduces an inherent discontinuity between economic reality and financial representation. Capital does not move in discrete daily blocks.

Corporate treasurers, hedge funds, and institutional liquidity managers increasingly operate on intraday horizons, shifting large pools of capital multiple times within a single day to optimize yield, manage risk, or respond to liquidity needs. In such an environment, the binary nature of traditional accrual systems can create distortions—effectively penalizing intraday participation and favoring static, overnight positioning.

Tokenized money market funds, by contrast, propose a continuous-time alternative. Instead of anchoring yield distribution to a single daily snapshot, they track ownership on a real-time or near-real-time basis. Every second of ownership contributes proportionally to accrued interest. If an investor holds the asset for half a day, they earn half a day’s yield. If they hold it for two hours, they earn two hours’ worth. The underlying principle is simple but powerful: time-weighted ownership replaces point-in-time ownership.

This shift has profound implications for liquidity management. For a corporate treasurer managing working capital across global accounts, the ability to deploy funds intraday without sacrificing yield efficiency changes the calculus of idle cash. Capital no longer needs to be parked overnight simply to capture the day. Instead, it can be dynamically allocated across instruments, strategies, or counterparties while still accruing proportional yield in real time.

Hedge funds and proprietary trading desks—entities that frequently move dry powder capital in response to short-term market dislocations—gain the ability to optimize both opportunity capture and yield retention simultaneously. Under legacy systems, the friction of losing a full day’s interest can discourage intraday reallocation, subtly encouraging inefficiency. Tokenized systems remove that friction by aligning yield precisely with holding duration.

This is not merely a UX improvement layered onto existing infrastructure; it reflects a different financial abstraction. Traditional systems are built around batch settlement and periodic reconciliation, while tokenized systems operate on continuous ledger states.

In the former, time is discretized into accounting periods. In the latter, time becomes a variable directly embedded into asset behavior. Of course, this model introduces its own complexities. Continuous accrual requires robust oracle infrastructure, precise time-weighted accounting mechanisms, and careful handling of edge cases such as chain latency, forks, or reorganization risks. It also raises regulatory questions: if yield is continuously streamed, how should it be classified for reporting, taxation, and compliance purposes?

These are non-trivial challenges that determine whether such systems can scale beyond niche adoption. Still, the core innovation is clear: by removing the artificial boundary of the trading day, tokenized money market funds transform interest from a daily event into a continuous function of ownership. For high-frequency capital allocators, this is not just an incremental efficiency gain—it is a structural redefinition of how cash yield is measured, earned, and optimized in modern financial markets.

Why AI Investors are Missing this Massive Blind Spot

0

AI investment today is increasingly defined by a narrow consensus: capital flows overwhelmingly into foundation models, GPU infrastructure, and a handful of hyperscaler ecosystems. Yet beneath this surface-level momentum sits a structural blind spot that many investors are systematically underpricing—the shift from model-centric value creation to workflow-native, constraint-driven, and distribution-anchored AI systems.

This misalignment between where capital is deployed and where durable value accrues is becoming more pronounced as the AI stack matures. The dominant investment thesis assumes that the primary bottleneck in AI remains model capability. As a result, funding continues to concentrate on scaling parameter counts, securing compute supply, and training ever-larger multimodal systems. However, marginal gains from raw model scaling are showing diminishing returns in real-world enterprise adoption.

In practice, most organizations are not constrained by the absence of frontier intelligence, but by integration friction, workflow redesign costs, and governance constraints. The bottleneck has quietly shifted from intelligence creation to intelligence deployment.

This creates a critical blind spot: investors are overweight exposure to upstream AI infrastructure while underestimating the economic gravity of downstream application layers. The highest long-term margins are increasingly emerging not from model ownership, but from control over decision loops—systems that embed AI into repetitive, high-frequency, economically consequential workflows.

These include compliance automation, procurement optimization, industrial scheduling, fraud detection pipelines, and enterprise knowledge systems. These domains do not reward general intelligence; they reward specificity, latency reduction, and institutional embedding. Another dimension of this blind spot lies in distribution asymmetry.

Many AI companies assume that superior model performance will naturally translate into adoption. In reality, distribution—not capability—is becoming the binding constraint. Enterprises are not choosing tools based on benchmark superiority but on integration depth into existing software ecosystems. This favors incumbents with established user bases and proprietary workflow lock-in.

It also favors vertically integrated platforms that can bundle AI functionality into existing SaaS layers rather than standalone model providers. Compounding this is the underestimated cost structure of AI deployment at scale. While training costs dominate headlines, inference economics are becoming the true margin determinant.

Enterprises are discovering that AI systems that are technically superior can be economically nonviable at scale due to inference latency, token consumption, and orchestration overhead.

This is pushing value toward architectures that prioritize efficiency, caching, and hybrid symbolic-neural systems—areas that are underfunded relative to pure deep learning scaling bets. There is also a cognitive blind spot among investors regarding substitution risk. Many assume AI adoption is purely additive—new tools replacing manual labor without disrupting existing software incumbents.

In practice, AI is reshaping entire software categories by collapsing multi-step workflows into single-agent operations. This creates nonlinear disruption risk for traditional SaaS models, while simultaneously opening space for new categories of agent-native software that do not resemble conventional applications at all.

Finally, capital markets are underpricing the importance of regulatory and organizational friction. AI adoption is not just a technological problem; it is an institutional coordination problem. Legal liability, auditability requirements, and data governance frameworks significantly slow enterprise deployment. Companies that solve these constraints—rather than simply improving model accuracy—will capture disproportionate value.

The result is a widening divergence between narrative-driven AI investment and constraint-driven AI adoption. Investors focused solely on compute scaling and frontier model development risk overlooking where compounding returns actually emerge: at the intersection of workflows, distribution, and operational integration.

The next phase of AI value creation will not be defined by who builds the largest models, but by who embeds intelligence most deeply into the economic machinery of organizations.