Startup Founders: As Nigeria passes through a difficult fundraising season, what many describe as funding paralysis, I want you to remember a simple but enduring truth: nations rarely fail. Economies bend, currencies weaken, systems stall, but societies recover. History is remarkably consistent on this point. So, that your growth fund is not coming does not mean you have to give up. Go back to the drawing board and update your business model to see if you can modulate on that scaling to preserve cash.

For me, there are two men I often study in moments like this: Franklin Templeton, arguably the greatest stock picker of the twentieth century, and Carlos Slim, the Mexican billionaire. Both built enduring empires at moments when their countries appeared broken. While others saw ruin, they saw an unbounded future, and they won.

As I write, I am investing big in Nigeria. I expect to have about 100  NEW people in a new business in Nigeria by the end of 2026.

Franklin Templeton founded his investment firm in 1947, at the wreckage of World War II. Europe was in ashes, confidence was low, and capital was scarce. Yet he trusted humanity. He bought what others dismissed as “useless” stocks, betting not on balance sheets alone but on the human instinct to rebuild.

Carlos Slim did something similar decades later. At one of the lowest points in Mexican history, when the peso collapsed and markets were in disarray, he bought aggressively. His father had taught him a powerful lesson: countries do not die; they reset. Slim believed that when stability returned, value would follow.

We have seen this pattern repeat. Uber was founded during the Great Recession. Airbnb was also born in that same crisis. Had both been conceived during a time of abundance, they likely would have failed. Yes, too much comfort dulls imagination. Scarcity, on the other hand, sharpens thinking.

Success is not about being busy. It is about understanding context and making sense of moments. Today, Nigeria has acres of diamonds, scattered across sectors and markets. They are not obvious. They are buried in constraints, inefficiencies, and unmet needs. You must believe in people because if you do not trust human aspiration, you will never see opportunity.

This challenging funding moment will pass. Like cryolite hidden inside periwinkle, beauty often emerges only after pressure. Unless the shell is cracked, the gem remains unseen. Do not lose confidence. Abundance is still ahead. Yes, funding will return but we must survive for the moment when it does.

The task is productive exploration. Think deeply. Build deliberately. Nigeria still has vast “diamond fields” waiting to be mined and the best playbook is to understand our long gestation period and then retool business models to accommodate that reality. Drop the Silicon Valley playbook and build an African model, accounting for the realities we have on ground on funding.

Good People, the question is now: Who can thrive in a “funding recession”? Yes, The Nigerian equivalents of Uber and Airbnb which rose out of the miry clay of great recession. #build.

Nvidia Corp. has reportedly pressed Samsung Electronics to hasten deliveries of its sixth-generation high-bandwidth memory (HBM4) chips, even before completing full reliability and quality evaluations, according to Chosun.

The move signals a high-stakes scramble for advanced memory that underscores the shifting power dynamics in the global AI ecosystem.

Industry sources indicate Samsung is finalizing inspections for mass production and shipments starting in February 2026, but Nvidia’s request to bypass detailed testing reflects urgency driven by intensifying competition from rivals like AMD and Google in AI accelerator design.

This marks a notable role reversal from the HBM3E era, where Samsung’s supply hinged on passing Nvidia’s rigorous qualifications; now, within a single generation, Nvidia is prioritizing speed over exhaustive verification, viewing HBM4 as sufficiently vetted. The company plans to integrate HBM4 into its next-generation Rubin AI accelerators, which demand unprecedented bandwidth and capacity for handling massive AI workloads.

The collaboration extends beyond basic supply: Nvidia and Samsung are synchronizing production timelines, with HBM4 modules slated for immediate use in Rubin performance demonstrations ahead of the official GTC 2026 unveiling. This partnership tightens Korea-U.S. supply loops for top-tier AI silicon, but rushing shipments risks exposing issues in reliability, thermal management, or yield consistency—challenges that have plagued prior HBM ramps.

The urgency highlights a profound shift in the AI supply chain, where Korean memory giants Samsung and SK Hynix now command bottleneck control. Once viewed as subordinate suppliers, these firms have ascended to “super subcontractor” status, dictating terms in a market where HBM scarcity directly impacts AI chip launches, data center expansions, and competitive edges.

Without sufficient HBM, even Nvidia’s advanced GPUs falter, as the memory supports the computational intensity required for frontier AI models.

Market forecasts underscore this dominance. Counterpoint Research projects SK Hynix capturing 54% of the global HBM4 market in 2026, with Samsung at 28%—together holding over 80% share.  UBS anticipates SK Hynix securing approximately 70% of Nvidia’s HBM4 needs for the Rubin platform, while Samsung aims for over 30%. In the broader HBM market for Q3 2025, SK Hynix led with 53% revenue share, followed by Samsung at 35% and Micron at 11%.

Financial projections reflect the “memory supercycle” boom. Morgan Stanley forecasts Samsung’s 2026 operating profit at 245 trillion won ($180 billion)—nearly six times its 43.6 trillion won in 2025—while SK Hynix is expected to hit 179 trillion won, up from 47.2 trillion won.

Combined, the duo could exceed 200 trillion won in profits, per some estimates. SK Hynix’s Q4 2025 operating profit surged 137% to 19.2 trillion won ($13.5 billion), beating forecasts, while Samsung’s memory division recorded 24.9 trillion won for FY2025. Both firms have sold out their 2026 memory inventory, entering a phase of severe supply constraints and elevated margins.

Capital expenditures are ramping up accordingly. SK Hynix plans over 30 trillion won in 2026 (up from mid-20 trillion in 2025), with 90% allocated to DRAM and HBM; Samsung anticipates exceeding 40 trillion won, focusing on HBM output, Pyeongtaek expansion, and its Texas fab. This investment surge responds to insatiable AI demand, with Nvidia’s Jensen Huang warning of massive memory needs straining supply chains.

Broader shortages in commodity DRAM and NAND—exacerbated by HBM prioritization—have driven explosive price hikes: consumer DRAM up 750% to $11.50 from $1.35 in January 2025, NAND to $9.46 from $2.18. These dynamics enhance Korean firms’ bargaining power, likened by KB Securities’ Kim Dong-won to TSMC’s foundry dominance.

KAIST professor Kim Jeong-ho noted memory’s evolution toward customized products amplifies their influence: “In the future AI era, memory will dominate the industry.”

SK Hynix shares surged 23% in a week amid speculation of Nvidia HBM4 breakthroughs.

This shift elevates Korean memory makers to “super subcontractor” status, controlling the AI bottleneck. As demand outpaces supply into 2027, their leverage is expected to reshape alliances, pricing, and innovation timelines.

Intel is preparing to enter the graphics processing unit (GPU) market in a more serious way, a move that would pit the once-dominant chipmaker more directly against Nvidia in data centers.

This comes as the chipmaker’s chief executive, Lip-Bu Tan, sounds a note of urgency about China’s rapid progress in chip design despite U.S. restrictions.

The renewed ambition marks a significant strategic pivot for a company that once defined the cutting edge of the global semiconductor industry but has spent much of the past decade playing catch-up. Comments from Lip-Bu Tan on Tuesday underline both the scale of Intel’s aspirations and the competitive pressures driving them, from Nvidia’s dominance in AI chips to China’s accelerating capabilities under export restrictions.

Speaking at the Cisco AI Summit, Tan confirmed that Intel is actively developing GPUs, the class of chips that has become central to artificial intelligence workloads and the primary engine of Nvidia’s extraordinary growth. GPUs, originally designed for graphics rendering, are now the backbone of data centers used for training and running large AI models, an area where Intel has been notably underrepresented.

“I just hired the chief GPU architect, and he’s very good. I’m very delighted he joined me,” Tan said, adding that it took some effort to persuade him to come on board. While Tan did not disclose timelines or product specifics, the hire signals that Intel’s GPU push is no longer experimental but structural, anchored by senior technical leadership.

In an interview with Reuters on the sidelines of the event, Tan made clear that Intel’s GPU ambitions are tightly integrated with its data center strategy rather than aimed at consumer graphics or gaming, where Nvidia and AMD are also strong.

“It’s tied in with the data center,” Tan said. “We’re working with customers, and will then define what the customer needs.”

That customer-first framing reflects lessons Intel has learned the hard way. Nvidia’s success has been driven not just by powerful chips but by tight collaboration with hyperscalers and AI developers, alongside a mature software ecosystem. Intel, by contrast, has often been accused by customers and analysts of designing products in isolation and arriving late to fast-moving markets.

A key figure in Intel’s renewed effort is Eric Demmers, a senior executive who joined from Qualcomm last month. Demmers will report to Kevork Kechichian, Intel’s head of data center chips, reinforcing the message that GPUs are now a core part of Intel’s data center roadmap. Demmers’ move is notable because Qualcomm itself has been expanding its ambitions beyond mobile chips, and Intel’s ability to attract senior talent from rivals suggests a degree of renewed confidence under Tan’s leadership.

Beyond product development, Tan also highlighted early momentum at Intel Foundry, the contract manufacturing arm that sits at the heart of the company’s turnaround plan. Intel has bet heavily on becoming a leading alternative to Taiwan Semiconductor Manufacturing Company (TSMC), offering advanced manufacturing to external customers while also supplying its own chips.

Tan said “a couple of customers are engaging heavily” with Intel Foundry, particularly around its upcoming 14A process technology. He indicated that volume manufacturing on 14A could ramp up later this year, though he stressed that firm commitments depend on customers specifying volumes and products well in advance.

“In order to have a customer … they have to let us know what is the volume and which product, so that we can plan and take time to build the capacity,” Tan said.

This cautious tone reflects the reality that Intel’s foundry ambitions are still unproven at scale. While interest from customers is encouraging, Intel must demonstrate that it can deliver leading-edge manufacturing on time and with competitive yields, something it has struggled with internally in recent years.

Perhaps the most striking part of Tan’s remarks, however, was his warning about China’s progress, particularly at Huawei Technologies. Tan said he was “shocked” to discover during a recent hiring drive that Huawei had recruited around 100 highly skilled chip designers, despite U.S. restrictions that have cut the company off from advanced manufacturing tools and key design software.

When Tan asked Huawei engineers why they chose to work for a firm operating under such constraints, their answer was telling.

“They said, ‘Even though we don’t have access to the best tools, like electronic design automation tools from Cadence and Synopsys, we have the poor man way to do it, and we can do it,’” Tan said.

His conclusion was stark. “To me, they are just shortly behind us, and if you’re not careful, they will just leap forward ahead of us.”

The warning cuts to the heart of the global semiconductor race. U.S. export controls were designed to slow China’s technological progress, but Tan’s comments suggest that talent, ingenuity, and scale can partially offset restrictions on tools. This adds urgency to Intel’s efforts to rebuild leadership in advanced chips, not just as a commercial imperative but as part of a broader geopolitical competition.