billion

Africa’s 5G access is far below the global average of more than 20 percent, highlighting connectivity challenges.

Only 1.2 percent of Africans currently have access to 5G networks compared with a global average of more than 20 percent, a sign that the continent remains at an early stage in accessing next-generation mobile technology, according to a new report from the International Telecommunication Union and the United Nations Educational, Scientific and Cultural Organization (UNESCO).

The State of Broadband in Africa 2025 report says that while Africa has made strong progress in mobile connectivity, the newest wireless technology remains largely out of reach for the continent’s 1.24 billion inhabitants.

However, industry projections suggest this could change dramatically over the next decade, with 5G expected to reach 17 percent penetration by 2030.

The slow 5G rollout contrasts sharply with Africa’s mobile success story in previous generations of wireless technology. Currently, 3G networks reach 77 percent of the continent’s population, while 4G coverage extends to 44.3 percent of people.

“The mobile sector has proved especially successful, with strong growth in mobile broadband and the development of large mobile operators,” the report states, citing companies like MTN and Vodacom as key drivers of expansion.

“However, there is still a significant usage gap, with 710 million Africans not using the internet despite living in an area served by mobile broadband infrastructure,” it adds. Key barriers, it says, remain affordability of handsets and lack of digital skills.

Chinese companies like Huawei, with more competitively priced products, have been able to establish a strong presence as a result across Africa.

The technology mix across sub-Saharan Africa shows 3G connections still dominating at roughly 50 percent of all mobile connections, while 4G accounts for 33 percent. Legacy 2G networks maintain 10 percent of connections, with 5G making up the remaining fraction.

Some countries, such as Somalia, have seen mobile connectivity flourish, not despite a lack of central authority but largely because of it, as large telecom companies have established large networks that cover urban areas well but also remote parts of the country, leading one British researcher to unfavourably compare Manchester with Mogadishu.

Mobile operators have invested heavily in infrastructure development, spending $28bn over the past five years across sub-Saharan Africa. Looking ahead, the industry plans to invest an additional $62bn between 2023 and 2030, much of which will focus on 5G network rollout and expansion.

The mobile ecosystem already contributes significantly to African economies, generating 7.3 percent of gross domestic product (GDP) worth $140bn in economic value and supporting 3.7 million jobs across the region in 2023.

Regional disparities within Africa reveal stark contrasts in connectivity progress.

Internet usage in Africa rose from 25 percent to 38 percent between 2019 and 2024, but remains well below the 68 percent global average. Sub-Saharan Africa lags furthest behind at 38 percent connectivity, with regional variations from 35 percent in Eastern and Southern Africa to 39 percent in Western and Central Africa.

Africa had a stark rural-to-urban divide in internet connectivity globally too. Only 57 percent of people in Africa were using the internet in urban areas compared with an 83 percent global average, and only 23 percent in rural areas.

Rwanda emerges as a particular success story, with telecommunications transformation following market liberalisation in 2006. The country developed a wholesale open-access 4G LTE network through a public-private partnership with Korea Telecom, ranking ninth among 38 African countries for mobile broadband affordability in 2017.

Satellite connectivity is expanding rapidly, with Starlink already operating in 14 African countries, including Benin, Ghana, Kenya, and Nigeria, though South Africa notably lacks a confirmed launch date. The Democratic Republic of the Congo and Somalia became the latest African countries to gain access to Starlink this year.

CoreWeave (CRWV 2.77%) just announced that it has landed an order from AI chip giant Nvidia (NVDA 0.34%) worth at least $6.3 billion. The deal obligates Nvidia to purchase its residual unsold capacity until April 13, 2032, if it is not already purchased by CoreWeave’s customers.

CoreWeave is a longtime Nvidia client. It has purchased hundreds of thousands of GPUs, which CoreWeave rents out to clients. With the agreement, CoreWeave gains an additional revenue source.

Still, investors need to remember that the spare capacity was valuable enough to pay CoreWeave $6.3 billion, raising questions about which company will reap the greater benefit. Thus, investors should take a closer look at the deal to see whether CoreWeave or Nvidia benefits more.

Image source: Getty Images.

How CoreWeave benefits

CoreWeave obviously benefits from the deal since it has a guaranteed customer, and it will receive $6.3 billion in revenue over the length of the agreement.

But understanding the benefit fully means knowing what CoreWeave offers to the market. CoreWeave is an artificial intelligence (AI) based cloud computing company. It differs from cloud providers like Amazon‘s AWS as it builds servers specifically for workloads related to AI, machine learning, high-performance computing, and visual effects.

It also provides customers with the latest hardware, can handle heavy workloads at lower costs, and its per-instance pricing allows customers to manage their costs more closely. Such an approach makes CoreWeave valuable to the AI industry, but it can also leave it with unused capacity. Nvidia’s move to claim that spare capacity should therefore provide CoreWeave with some degree of stability.

Knowing that, investors should note where CoreWeave stands financially. Its $58 billion market cap is a tiny fraction of Nvidia’s $4.25 trillion size. Also, even though CoreWeave’s revenue of $2.2 billion in the first half of 2025 grew by 275% yearly, it still lost $605 million during that time. Such losses mean it will likely have to turn to capital markets to raise funds.

That need is even more acute because it pledged to spend between $20 billion and $23 billion in capital expenditures (capex) in 2025 alone. So it needs deals like the one with Nvidia so it can recoup its capex investments and eventually grow into a profitable operation.

Why Nvidia made this agreement

What may confuse investors is how the deal helps Nvidia. Even though Nvidia and CoreWeave are close partners, CoreWeave’s business requires it to purchase Nvidia’s latest AI accelerators. But investors may not understand why it chose to spend $6.3 billion to help this particular customer.

For one, Nvidia owns 24.3 million shares of CoreWeave as of June 30, about 5% of the outstanding shares. The agreement goes a long way toward solidifying Nvidia’s investment and the relationship between the two companies.

Intense demand for AI has led to a shortage of cloud capacity, so this deal also gives Nvidia a claim over a scarce commodity. Thanks to this deal, Nvidia will have to rely less on large cloud providers like Amazon and Microsoft, giving it more control over its destiny in this regard.

Furthermore, having CoreWeave in a more solid financial position means CoreWeave will more likely meet the aforementioned goals on capex spending. With that, it will spend more on Nvidia GPUs, helping to boost the size and expanse of the AI ecosystem over which it has considerable control.

What may look like charity to CoreWeave will likely serve Nvidia’s interests in the end.

Does CoreWeave or Nvidia benefit more from the deal?

This deal will probably benefit both companies, although CoreWeave is likely to derive more benefit, at least from an investor perspective.

Nvidia may benefit more in an abstract sense, as it gains influence over the AI ecosystem. Nvidia’s larger size also makes it a safer investment, a welcome relief to investors who feel uncomfortable buying into a money-losing company spending heavily on capex.

Still, the Nvidia deal offers a considerable boost to CoreWeave’s top line. This makes it more likely that CoreWeave’s massive spending will ultimately deliver positive returns for the company.

CoreWeave’s much smaller size also means it can attain higher percentage growth from a much smaller base. The doubling of CoreWeave’s value takes its market cap to $116 billion. The same percentage move would take Nvidia’s market cap to $8.5 trillion, a challenging feat in a market that has yet to see a company with a $5 trillion market cap.

In the end, CoreWeave comes with higher risks than Nvidia. However, if you are willing to take a chance, CoreWeave offers greater potential for higher-percentage returns.

For the first time since artificial intelligence (AI) captured Wall Street’s imagination, investors are beginning to broaden their scope beyond the “Magnificent Seven.” Two names that have attracted growing attention this year are Oracle and CoreWeave.

Unlike the tech titans that dominate headlines, Oracle and CoreWeave are carving out their niche at the infrastructure layer of the AI ecosystem. The opportunity they’ve identified is straightforward but also mission-critical: providing cloud-based access to GPUs. These chips — designed primarily by Nvidia and Advanced Micro Devices — remain supply constrained as they are largely absorbed by the world’s largest companies.

This supply imbalance has created an opportunity to enable AI model development by offering GPUs as a service — a business model that allows companies to rent chip capacity through cloud infrastructure. For businesses that cannot secure GPUs directly, infrastructure services are both time-saving and cost-efficient.

In the background, however, a small, albeit capable, company has been competing with Oracle and CoreWeave in the GPU-as-a-service landscape. Let’s explore how Nebius Group (NBIS 5.54%) is disrupting incumbents and why now is an interesting time to take a look at the stock for your portfolio.

17.4 billion reasons to pay close attention to Nebius

Last week, Nebius announced a five-year, $17.4 billion infrastructure agreement with Microsoft. For reference, up until this point, Nebius’ management had been guiding for $1.1 billion in run rate annual recurring revenue (ARR) by December. I point this out to underscore just how transformative this contract is in terms of scale and duration.

The Microsoft deal not only places Nebius firmly alongside peers like Oracle and CoreWeave in the AI infrastructure conversation, but it also serves as validation that its technology is robust enough to meet the standards of a hyperscaler.

For Microsoft, the partnership is equally strategic. With GPUs in chronically short supply and long lead times to expand data center capacity, this agreement allows Microsoft to secure adequate compute resources without stretching internal infrastructure or assuming the upfront capital expenditure (capex) budget and execution risks that come with it.

Image source: Getty Images.

Why this deal matters for investors

AI investment is not a cyclical trend — it’s a structural shift. Enterprises are deploying applications into production at unprecedented speed, workloads are scaling rapidly, and new use cases in areas like robotics and autonomous systems are emerging.

For companies that supply the compute underpinning this increasingly complex ecosystem, these dynamics create durable secular tailwinds. By securing Microsoft as a flagship customer, Nebius has established itself within this foundational layer of the AI infrastructure economy.

Is Nebius stock a buy right now?

Since announcing its partnership with Microsoft, Nebius shares have surged roughly 39% as of this writing (Sept. 16). With that kind of momentum, it’s natural to wonder whether the stock has become expensive. To answer that, it helps to put its valuation in context.

Prior to the Microsoft deal, Nebius was guiding for $1.1 billion in ARR by year-end. If I assume Microsoft’s $17.4 billion commitment is evenly spread across five years (2026 to 2031), that adds about $3.5 billion annually — bringing Nebius’ pro forma ARR closer to $4.6 billion.

Against its current market cap of $21.3 billion, Nebius stock trades at an implied forward price-to-sales (P/S) ratio of 4.6. On the surface, that looks meaningfully discounted to peers like Oracle and CoreWeave.

ORCL PS Ratio data by YCharts

That said, there are important caveats to consider. My analysis assumes no customer attrition over the next several years — this is unrealistic due to competitive pressures. While Nebius may continue winning large-scale contracts, it’s also reasonable to expect some customer churn.

Moreover, comparing Nebius’ future ARR to Oracle’s and CoreWeave’s current revenue base is not an apples-to-apples match. Oracle, for example, has reportedly inked a $300 billion cloud deal with OpenAI. Meanwhile, CoreWeave also has multiyear, multibillion-dollar commitments tied to OpenAI. The catch is that OpenAI itself doesn’t have the cash on its balance sheet to fully fund these agreements — leaving questions about their viability.

In short, Nebius appears attractively valued relative to its peers — but the landscape is evolving quickly and riddled with moving parts. The more important takeaway is that Nebius is now winning significant business alongside its brand-name peers.

In my eyes, this validation in combination with ongoing structural demand tailwinds makes Nebius a compelling buy and hold opportunity as the AI infrastructure narrative continues to unfold.

Nvidia (NVDA 3.52%) has built an artificial intelligence (AI) empire thanks to the dominance of its AI chips and its expansion into a wide variety of other related offerings. But the company isn’t isolating itself, and instead, has looked to work with others — even much smaller players — in this AI boom. One company in particular has become a key Nvidia ally, and that’s CoreWeave (CRWV 0.39%).

CoreWeave launched an initial public offering in March, and the stock has since surged about 195%, buoyed by the company’s soaring sales — and its relationship with Nvidia. The AI chip giant held a 7% stake in CoreWeave as of the end of the second quarter, and CoreWeave makes up 91% of Nvidia’s investment portfolio. And CoreWeave’s business relies heavily on Nvidia as the company’s specialty is the following: It rents out Nvidia’s high-powered graphics processing units (GPUs) to customers through its cloud platform.

Now, Nvidia’s latest move — a $6.3 billion deal with CoreWeave — signals something big for shareholders of both companies. Let’s take a closer look.

Image source: Getty Images.

A 1,300% gain

First, though, a quick summary of the businesses of Nvidia and CoreWeave. As mentioned, Nvidia is the AI chip leader, with its GPUs and related products delivering record revenue and earnings over the past few years. Nvidia’s chips offer the highest performance on the market, so tech giants, prioritizing AI success, have rushed to get in on these essential tools. All of this has helped Nvidia stock climb 1,300% over the past five years — and pushed market value past $4 trillion to make Nvidia the world’s biggest company.

CoreWeave, as mentioned, offers customers access to Nvidia compute through its cloud platform. Customers may rent GPUs by the hour or for the long term, and this offers them great flexibility. CoreWeave holds about 250,000 GPUs across 32 data centers and has been the first to make Nvidia’s latest innovations generally available. All of this has translated into outsized revenue growth, with sales tripling in the latest quarter. CoreWeave clearly depends on Nvidia’s success as demand for Nvidia GPUs power its revenue higher — if demand were to decline, not only would Nvidia suffer, but so would CoreWeave.

And this brings me to the latest deal between the two companies. Nvidia signed a $6.3 billion order with CoreWeave, ensuring that the chip leader will buy any cloud capacity that CoreWeave is unable to sell to customers. The deal, extending a 2023 agreement, covers the period through April 13, 2032.

Eliminating a risk

This order signals something different — but significant — for both companies and their shareholders. For CoreWeave, this removes the big risk of the company being stuck with excess capacity. Though the future of AI spending looks bright, any dip in spending, even over a short period, could be costly for the company. So, Nvidia’s agreement to potentially step in means that if any drop in demand happens, it won’t hurt CoreWeave’s sales. As a result, shareholders may breathe a sigh of relief, and cautious investors who have worried about this risk may consider getting in on CoreWeave.

As for Nvidia, this move suggests the company truly is confident about the demand for AI capacity over the next several years. It’s unlikely the tech giant would agree to such a deal if it saw a major slowdown on the horizon. This reinforces Nvidia’s prediction a few weeks ago that AI infrastructure spending may reach $4 trillion by the end of the decade. Nvidia has said in the past that its customers offer it visibility about their upcoming needs — so the chip designer has a good idea of how the demand situation will evolve.

All of this means this latest deal between Nvidia and CoreWeave is fantastic news for shareholders of both companies — for CoreWeave, the agreement lowers risk, and for Nvidia, the agreement confirms that demand for AI is going strong.

Considering this, both of these companies make great AI stocks to buy and hold onto as this AI growth story develops.

Billionaire Bill Ackman is one of the most widely followed investment managers on Wall Street. His Pershing Square Capital Management hedge fund has outperformed the S&P 500 in 2025. It’s up 22.9% as of the end of August, compared to a 10.8% gain in the benchmark index during that period.

Ackman’s outperformance stems from taking advantage of opportunities when the market temporarily undervalues certain stocks. He holds only a handful of positions in the fund, and he typically buys and holds them for a long time. Even better, he and his team are happy to share the details on social media and investor calls, making it relatively easy for average investors to follow along.

In May, Pershing Square disclosed that it had bought another member of the “Magnificent Seven” stocks. Its first Magnificent Seven stock, Alphabet (GOOG -0.72%) (GOOGL -0.73%), has been a longtime holding for the hedge fund, and represents one of its biggest holdings. While the new addition isn’t quite as large as its stake in Alphabet, it presents another great opportunity for those following Ackman’s investing style.

Image source: Getty Images.

A magnificent new position

The stock market saw some very big swings at the start of the year, which were exacerbated in early April by President Donald Trump’s tariff announcements. While the stock market was moving wildly, it presented several great opportunities for investors that could follow Warren Buffett’s timeless advice: “Be greedy when others are fearful.”

To that point, Ackman saw the chance to pick up one stock he’s been studying and has long admired. Amazon (AMZN -1.20%) shares fell on fears that tariffs would negatively affect its retail business, and that a slowing economy would produce less demand for its cloud computing services. Ackman and his team freed up capital by selling Pershing Square’s entire position in Canadian Pacific Kansas City to buy the stock.

Ackman got a steal of a deal. He said he bought shares at 25 times forward earnings estimates. While there was a lot of uncertainty at the time about whether those earnings estimates would need to be revised downward, Ackman had confidence that Amazon was well worth the price. In fact, he thinks the stock is still undervalued. “Although the company’s share price has appreciated meaningfully from our initial purchase, we believe substantial upside remains given its ability to drive a high level of earnings growth for a very long time,” he wrote in his letter to shareholders last month.

Here’s why Ackman may continue to hold Amazon shares for a very long time.

Two great category-defining businesses

Amazon essentially has two businesses: Its retail operations and its cloud computing platform. Ackman believes both still have room to benefit from long-term growth trends and opportunities for margin expansion.

On the cloud computing side, Amazon Web Services (AWS) is the largest public cloud provider in the world. It now sports a $120 billion run rate, and it’s about 50% bigger than its next-closest rival. It’s also tremendously profitable already. The segment sports a 37% operating margin over the past 12 months. To put that in perspective, Alphabet’s Google Cloud has an operating margin of less than half that (although it’s gaining leverage as it scales).

Despite Amazon’s large run rate, there’s still ample room for growth in both the near term and long term, according to Ackman. Amazon’s management has struggled to build out capacity fast enough to meet the surging demand from artificial intelligence customers. It’s spending over $100 billion on capital expenditures this year (some of that related to its logistics network), and management says that demand continues to outstrip supply growth. That situation is echoed by Alphabet’s management and other hyperscale cloud providers.

In the long run, Ackman expects more enterprises to move from on-premise computing to the cloud. He points out that just 20% of IT workloads are currently using cloud computing, but he expects that to invert over time, to 80% of workloads being in the cloud.

On the retail side of the business, Ackman points out that Amazon isn’t the only retailer affected by tariffs. In fact, it may be better suited to navigate the environment, as it sports a wide selection of goods. Amazon’s ability to offer reliable and convenient delivery on a growing number of items gives it an advantage over competitors.

That advantage is only improving as it continues to build out its logistics network and warehouse technology, and reduce costs. That allows it to get more items to more customers faster, all while decreasing its fulfillment expenses. Ackman points out that Amazon’s logistics improvements led to a 5% reduction in per-unit shipping costs last quarter. He thinks further improvements could lead it to double its retail profit margin from 5%. That’s a huge profit on a $550 billion business.

While Amazon shares have climbed significantly since Ackman established Pershing Square’s position, investors shouldn’t shy away from the stock at this higher price. The long-term trends favor Amazon’s businesses, and it’s a leading player in both.