Microsoft is intensifying its search for artificial-intelligence startups as the technology giant moves to reduce its dependence on longtime partner OpenAI and position itself to compete more directly at the frontier of advanced AI model development, Reuters reports, citing people familiar with the matter.

The acquisition discussions reveal a major shift inside Microsoft, which for years relied heavily on OpenAI’s technology to power its AI ambitions across products, including Azure, Office, and GitHub Copilot.

Now, after investing more than $13 billion into OpenAI and helping transform the startup into one of the most valuable companies in the world, Microsoft is increasingly preparing for a future in which it may need to stand on its own in the race to build cutting-edge artificial intelligence systems.

According to people familiar with the matter, Microsoft is exploring deals that could help it secure elite AI talent, gain access to new model architectures, and accelerate efforts to develop a world-class proprietary foundation model by next year.

The company’s interest comes as competition for AI startups and researchers reaches unprecedented levels across Silicon Valley, with major technology firms spending billions of dollars to secure scarce expertise and computing advantages.

Among the companies Microsoft considered acquiring this year was Cursor, the rapidly growing code-generation startup whose AI-assisted programming tools have become popular among software developers. People familiar with the discussions said Microsoft ultimately backed away from pursuing a deal because of internal concerns that regulators would likely scrutinize the acquisition heavily, given Microsoft’s ownership of GitHub and its AI coding platform GitHub Copilot.

Regulators in the United States and Europe have intensified scrutiny of large technology companies seeking to consolidate power in artificial intelligence, particularly in areas involving developer tools, cloud infrastructure, and foundation models.

Microsoft’s dominance in enterprise software and cloud computing already places the company under close regulatory watch, making major AI acquisitions more politically sensitive.

Still, Microsoft continues to pursue other opportunities. The company is currently in discussions with Inception, a small AI startup founded by researchers from Stanford University that is developing an alternative approach to large language models, according to people familiar with the talks.

Microsoft’s venture arm, M12, previously participated in Inception’s $50 million seed funding round in late 2025. The discussions remain ongoing and may not result in a transaction. Inception has drawn interest because of its work on diffusion-based language models, an emerging AI architecture that differs significantly from the autoregressive systems used by most leading AI developers today.

Traditional large language models generate text sequentially, producing one token at a time. Diffusion models instead generate and refine multiple tokens simultaneously, a technique more commonly associated with AI image and video generation.

Researchers believe the approach could dramatically improve inference speed and computational efficiency, two of the biggest constraints facing the AI industry as models become larger and more expensive to operate.

However, the technology remains experimental.

AI researchers caution that diffusion-based language systems can behave unpredictably, and it remains unclear whether they can scale effectively to frontier-level models containing trillions of parameters.

Microsoft’s Post-OpenAI Strategy Emerges

For years, the company’s partnership with OpenAI defined its position in the generative AI race. Microsoft first invested $1 billion into OpenAI in 2019 when the research lab remained relatively obscure. That partnership later became one of the most consequential alliances in modern technology after OpenAI launched ChatGPT in late 2022, triggering the global AI boom.

The relationship gave Microsoft privileged access to OpenAI’s models while helping drive enormous growth for its Azure cloud business. According to a securities filing released in April, Microsoft has already provided $11.8 billion of its promised $13 billion investment into OpenAI.

The company’s broader AI spending has become even larger.

Michael Wetter, who oversees Microsoft’s corporate development operations, testified in court Wednesday that Microsoft has spent more than $100 billion on OpenAI investments as well as the infrastructure and hosting systems required to support AI development. Those expenditures underscore the staggering financial scale of the AI arms race now unfolding across the technology sector.

Yet the Microsoft-OpenAI relationship has also become increasingly complicated. The original partnership granted Microsoft exclusive access to OpenAI technology while giving OpenAI guaranteed access to enormous computing resources for research and model training.

Over time, however, tensions reportedly emerged as OpenAI’s ambitions expanded. OpenAI executives concluded that their infrastructure needs were growing faster than Microsoft could supply, according to people familiar with the relationship. At the same time, Microsoft faced contractual restrictions that limited its ability to build foundation models competing directly with OpenAI’s offerings.

The companies have repeatedly renegotiated their agreements over the years as the balance of power shifted. One major change came in late 2025 when the partnership was amended to allow Microsoft to pursue development of artificial general intelligence, or AGI, a still-theoretical form of AI capable of outperforming humans across a broad range of complex cognitive tasks.

Another revision arrived in late April, when Microsoft and OpenAI reached an agreement allowing OpenAI greater flexibility to build products with rival cloud providers, including Amazon.

Those changes signaled a gradual loosening of what had once been one of Silicon Valley’s most tightly integrated AI partnerships.

The New AI Talent War

Microsoft’s acquisition search also underpins the escalating battle for elite AI talent. Compensation packages for top researchers now routinely reach tens of millions of dollars, while startup valuations are soaring as investors rush to secure exposure to promising AI technologies.

The market has become so competitive that even Microsoft is facing mounting pressure from rivals.

People familiar with the matter said Elon Musk’s SpaceX has emerged as one of Microsoft’s biggest competitors for AI startup deals following SpaceX’s acquisition of Musk’s AI company xAI earlier this year.

Shortly after Microsoft abandoned talks involving Cursor, SpaceX announced a partnership with the startup. SpaceX also reportedly courted Inception, according to people.

One source said Inception recently hired an investment bank to assist with negotiations and is seeking a valuation exceeding $1 billion.

The bidding competition shows that technology companies are no longer merely buying products or intellectual property. Increasingly, they are attempting to secure scarce clusters of researchers capable of designing next-generation AI architectures before competitors do.

Building an Independent AI Stack

Microsoft’s acquisition push is also believed to be an indication of a growing realization across the industry that long-term AI leadership may require tighter vertical integration. Companies increasingly want control over models, infrastructure, cloud services, and developer ecosystems simultaneously.

That helps explain why Microsoft is investing so aggressively in internal AI development alongside its OpenAI partnership.

Part of that effort is being led by teams under Mustafa Suleyman, the co-founder of DeepMind, who joined Microsoft last year to strengthen its consumer AI operations and long-term research capabilities.

The company’s broader strategy appears aimed at reducing reliance on any single external AI provider while preserving flexibility in an industry evolving at extraordinary speed. That transition is becoming increasingly important as frontier AI systems grow more expensive and strategically important.

Some leading AI laboratories are now building models containing roughly 10 trillion parameters, according to researchers, up from about 1 trillion parameters only three years ago. Those enormous systems require unprecedented levels of computing infrastructure, engineering talent, and capital investment.

For Microsoft, the stakes extend beyond competition with OpenAI. Artificial intelligence is rapidly becoming central to cloud computing, enterprise software, search, cybersecurity, and productivity applications, all areas critical to Microsoft’s long-term growth.

The company, therefore, appears determined to ensure it remains one of the industry’s dominant AI platforms even if its relationship with OpenAI eventually becomes less central to that ambition.

Aliko Dangote has revealed that his refinery business deliberately blocked the Nigerian National Petroleum Company Limited (NNPCL) from increasing its stake in the $20 billion Dangote Petroleum Refinery to make room for broader investor participation ahead of planned public listings across Africa.

Speaking during an interview with Nicolai Tangen, the head of the Norwegian Sovereign Wealth Fund, Dangote said the refinery rejected attempts by the state oil company to acquire additional equity because the group wants to spread ownership beyond a concentrated set of shareholders.

“The national oil company already owns 7.25%, and they are trying to buy more. We are the ones that said no; we want to now spread it and have everybody be part of it,” Dangote said.

The comments provide fresh insight into the evolving ownership structure of one of Africa’s most strategically important industrial projects and suggest Dangote is increasingly positioning the refinery as a pan-African investment vehicle rather than an asset dominated by either the founder or the Nigerian state.

The refinery, located in Lekki, Lagos, is already regarded as one of the largest single-train refining facilities globally and sits at the center of Nigeria’s efforts to reduce dependence on imported petroleum products.

In 2021, the NNPC agreed to purchase a 20% stake in the refinery for approximately $2.76 billion. However, the state-owned oil company ultimately completed payment for only 7.25% equity valued at around $1 billion. By 2024, Dangote disclosed publicly that the NNPC failed to complete payment for the remaining shares despite receiving an extension until June of that year.

The latest remarks suggest the original arrangement has now effectively been superseded by a broader capital market strategy aimed at distributing ownership more widely through planned stock exchange listings.

The move carries important financial and political implications. By widening ownership, Dangote may reduce concerns about excessive concentration of infrastructure under either private monopoly control or direct state dominance. A broader shareholder structure could also improve transparency, governance standards, and long-term capital access as the refinery expands operations.

Dangote linked the ownership strategy directly to concerns about policy instability in Nigeria.

“The other biggest risk is government inconsistencies in policies, and we are addressing that one,” he said.

That statement, which has been corroborated by several Nigerian entrepreneurs, reflects longstanding concerns among major investors about regulatory unpredictability, foreign exchange controls, subsidy shifts, and policy reversals in Nigeria’s energy sector. The refinery itself has repeatedly been drawn into disputes involving crude supply agreements, pricing frameworks, and fuel import dynamics.

By diversifying ownership across a wider investor base, Dangote may also be seeking stronger market-based protection against political risk and future regulatory pressure. The billionaire industrialist also made a notable pledge aimed at attracting both local and international investors at a time when currency instability continues to undermine confidence in Nigerian assets.

Dangote said future investors in the group’s businesses, including cement, fertilizer, petrochemicals, and refining operations, would receive dividends in foreign currency.

“What we are announcing is that when you invest in any of our businesses going forward, in cement or in the refinery, in petrochemicals, in fertilizer, we guarantee to pay you a dividend in dollars because we are very well into exports. Eighty per cent of our revenue will be in dollars,” he said.

The promise is significant in Nigeria’s current macroeconomic environment, where persistent naira volatility and foreign exchange shortages have weakened investor appetite for naira-denominated assets. Dollar-linked dividend commitments could make Dangote Group companies particularly attractive to foreign institutional investors and Nigerian investors seeking protection against currency depreciation.

The strategy also highlights how the Dangote conglomerate is increasingly evolving into an export-driven industrial platform rather than a business focused primarily on Nigeria’s domestic market.

The refinery, fertilizer operations, and petrochemical businesses are all expected to generate substantial foreign exchange earnings through regional and international exports. That export capacity has become especially important as Nigeria seeks to improve dollar inflows and stabilize external reserves.

Dangote also used the interview to discuss the personal sacrifices behind his industrial expansion strategy, presenting his decision-making as rooted in a long-term commitment to domestic industrialization.

He revealed that he sold his luxury homes in the United States and the United Kingdom in order to focus entirely on building businesses in Nigeria.

“When I decided to go into the industry, you know what I did? I sold all my properties in the US. I had two houses in the US, big mansions, and I had a house in the UK. I wanted to really sit in Nigeria and concentrate,” he said.

Dangote added that he now prefers staying in hotels while travelling abroad instead of maintaining foreign residences, arguing that permanent overseas assets can create distractions and divided attention.

The comments fit into a broader narrative Dangote has consistently projected over the years: that industrial transformation in Africa requires long-term capital commitment, operational patience, and local execution rather than dependence on imports or short-term speculative returns.

His refinery project itself became one of the most ambitious industrial bets in African history, facing repeated delays, cost overruns, logistical hurdles, and skepticism from investors and industry observers before eventually commencing operations.

The project has already altered dynamics within Nigeria’s downstream oil market by reducing fuel import dependence and increasing local refining capacity. Analysts say the refinery could eventually reshape fuel trade patterns across West Africa if it consistently operates near full utilization.

Dangote recently disclosed plans to double the refinery’s capacity to 1.4 million barrels per day, which would make it the world’s largest refining complex by capacity.

Such an expansion would dramatically increase the refinery’s importance not only to Nigeria but to global fuel markets, particularly as Europe and parts of Africa continue restructuring energy supply chains following disruptions in international refining capacity.

Cerebras Systems delivered one of the most explosive technology market debuts in years on Thursday, surging in its Nasdaq opening after raising $5.55 billion in a blockbuster initial public offering.

Shares of the Silicon Valley-based AI chipmaker opened at $350 after pricing at $185 late Wednesday, giving the company a valuation exceeding $100 billion within minutes of trading. The performance underscores how aggressively investors are betting on artificial intelligence infrastructure and the companies attempting to challenge Nvidia’s dominance.

The offering instantly became the largest U.S. technology IPO since Uber Technologies debuted in 2019 and marked one of the clearest signs yet that the long-frozen tech IPO market is reopening around artificial intelligence.

Cerebras sold 30 million shares in the offering, raising $5.55 billion. If underwriters exercise their option to purchase an additional 4.5 million shares, proceeds could climb to approximately $6.38 billion.

The scale of investor demand reflected a market increasingly consumed by the AI infrastructure race.

For much of the past two years, public equity investors remained cautious toward technology listings after inflation, rising interest rates, and weakening valuations triggered a severe IPO slowdown beginning in 2022.

But AI has dramatically altered investor sentiment.

The rise of generative AI systems and autonomous AI agents capable of performing increasingly complex tasks has unleashed one of the largest capital expenditure cycles in modern technology history, benefiting semiconductor firms, cloud providers, and data-center operators.

The semiconductor sector has become one of the biggest winners. Shares of Intel, Advanced Micro Devices, and Micron Technology have all recorded triple-digit gains this year, while the VanEck Semiconductor ETF has surged 58% in 2026 alone.

Against that backdrop, Cerebras entered public markets as perhaps the purest AI infrastructure bet available to investors outside Nvidia itself.

A Direct Challenge to Nvidia’s AI Dominance

Cerebras has spent years positioning itself as an alternative to Nvidia’s GPU-centric AI computing model.

While Nvidia dominates the AI accelerator market with graphics processing units originally designed for gaming applications, Cerebras built its business around radically different chip architecture optimized specifically for large-scale AI workloads. The company is best known for its wafer-scale processors, enormous chips that are physically much larger than conventional semiconductors and designed to handle massive AI computations with lower latency and higher efficiency.

Cerebras argues that its systems can outperform traditional GPU clusters in both speed and cost for certain AI inference and training tasks because its architecture reduces the need for complex inter-chip communication.

That positioning has become increasingly important as AI models grow larger and more computationally expensive.

The company’s emergence as a credible AI hardware competitor has drawn attention from across Silicon Valley, including Nvidia itself.

In December, Nvidia paid $20 billion for assets from startup Groq, whose chip architecture more closely resembles Cerebras’ approach than traditional GPU systems. Nvidia later announced plans for Groq-based products, signaling how seriously it views alternative AI computing architectures. The Cerebras listing therefore, represents more than another semiconductor IPO. It is effectively a public-market referendum on whether the AI infrastructure ecosystem can sustain major challengers to Nvidia’s near-monopoly position.

The AI IPO Window Reopens

Cerebras’ debut could also reshape the broader technology IPO market. Only 31 tech companies went public in 2025, according to data from University of Florida IPO expert Jay Ritter, down sharply from 121 offerings four years earlier. The collapse in listings followed the post-pandemic technology correction, which wiped out trillions of dollars in market value across growth stocks and made investors more cautious toward unprofitable technology firms.

Artificial intelligence is now reversing that trend.

Wall Street increasingly views AI-related companies as the next major infrastructure cycle, similar in significance to the rise of cloud computing or the internet itself. That shift is fueling expectations for a new wave of major AI listings. Elon Musk’s SpaceX, which merged with AI company xAI earlier this year, is reportedly preparing for a share sale. Meanwhile, OpenAI and Anthropic are both viewed as potential public-market candidates later this year.

Cerebras’ strong debut is likely to strengthen confidence among bankers and investors that the market can absorb massive AI-focused offerings again.

Revenue Growth and the UAE Connection

Cerebras’ financial performance helped fuel enthusiasm for the IPO. Revenue rose 76% last year to $510 million, while the company generated net income of $88 million after recording a loss of $481.6 million a year earlier. The profitability improvement distinguished Cerebras from many AI startups still heavily dependent on external financing.

Still, the company’s IPO process was unusually complicated. Cerebras first filed to go public in September 2024, but later withdrew the filing after regulatory and investor scrutiny intensified around its heavy dependence on customers linked to the United Arab Emirates.

At the time, a substantial portion of company revenue came from G42, the Microsoft-backed UAE artificial intelligence company that has attracted attention because of geopolitical concerns surrounding technology transfers and AI partnerships in the Gulf region.

Cerebras refiled in April with updated disclosures showing progress in diversifying its customer base. According to the revised prospectus, G42 accounted for 24% of revenue last year, down sharply from 85% in 2024. However, another UAE-linked institution, Mohamed bin Zayed University of Artificial Intelligence, represented 62% of revenue last year.

The disclosures highlighted one of the defining characteristics of the AI infrastructure market, where a small number of customers are spending extraordinary amounts of money on computing capacity.

“There’s some whales out there, there’s some really big customers,” Cerebras CEO Andrew Feldman told CNBC on Thursday. “That is one of the characteristics of this market.”

Discussing the company’s work with the UAE university, Feldman said: “We’re training models together,” adding that they are “English-Arabic models.”

“They are the first university set up and dedicated to training AI practitioners,” he said.

Governments and sovereign-backed institutions worldwide are investing heavily in AI capabilities as they seek influence over future technological development.

From Hardware Company to AI Cloud Provider

Another major transformation inside Cerebras is its shift away from pure hardware sales toward cloud-based AI services. Rather than simply selling AI systems, the company is increasingly offering access to computing capacity through cloud infrastructure built around its chips. That transition places Cerebras into more direct competition with major cloud providers, including Google, Microsoft, Oracle, and CoreWeave.

The strategy also reflects broader changes across the semiconductor industry. As AI computing becomes more centralized inside hyperscale data centers, chip companies are increasingly moving beyond hardware manufacturing into vertically integrated infrastructure services.

Cerebras has already signed major partnerships to support that transition. In January, the company announced a cloud agreement with OpenAI worth more than $20 billion through 2028. In March, Amazon Web Services said it would deploy Cerebras chips in its data centers, allowing developers to run AI models using Cerebras hardware through AWS infrastructure.

Both Amazon and OpenAI also hold warrants to purchase Cerebras shares, further embedding the company within the rapidly consolidating AI ecosystem.

Tekedia Institute is delighted to congratulate Dr. Julius Omodayo-Owotuga on his appointment as Executive Director of First Bank Nigeria, effective May 13, 2026. Good People, one of Nigeria’s finest professionals has ascended to yet another important leadership position, and from all of us at Tekedia, we celebrate JB Omodayo-Owotuga, FCA, CFA, DBA on this remarkable milestone.

Beyond professional excellence, JB and the Omodayo-Owotuga family have demonstrated a deep commitment to investing in people. Through a multi-year endowment established in memory of Late Most Supreme Apostle Matthew Omodayo Owotuga, hundreds of young people have attended Tekedia Mini-MBA through scholarships provided at no cost to the beneficiaries.

Even in our upcoming edition beginning in June, another cohort of scholars from the Owotuga Foundation will join the program, continuing a legacy of empowering and developing future leaders.

Good People, leadership is not only measured by positions attained but also by lives impacted. Through the Owotuga Family initiative, many young people have gained access to knowledge and opportunities which have shaped their careers. We know the number of recommendation letters we complete monthly as scholars and learners ascend to higher levels.

For First Bank, I expect the elephant to discover new dance moves. We wish JB and the entire First Bank Team continued success and many more wins ahead. And on a lighter note, perhaps now is the right time to open a First Bank branch in Ovim, Abia State. Ovim sons and daughters will ensure the bank has adequate deposits to make this branch a profit-center.