Yazılar

CoreWeave Slashes U.S. IPO Size Amid Lukewarm Investor Interest

CoreWeave has significantly reduced the size of its U.S. initial public offering (IPO) and priced shares lower than expected, signaling weaker investor confidence in the AI infrastructure sector.

The Nvidia-backed company will now offer 37.5 million shares at $40 each, a 23.5% reduction from its initial plan. This is well below the previously indicated range of $47 to $55 per share. The offering is expected to raise approximately $1.5 billion, valuing CoreWeave at around $23 billion on a fully diluted basis.

Investor Concerns and Market Challenges
CoreWeave’s IPO roadshow received a weaker-than-expected response due to investor concerns over its long-term growth, financial risks, and capital-intensive business model. The company’s heavy reliance on Microsoft, whose AI datacenter strategy is evolving, has also raised uncertainty about future demand for CoreWeave’s GPU-based services.

Additionally, CoreWeave’s substantial debt—approximately $8 billion—and its leasing model for 32 data centers and equipment have heightened investor caution. The company plans to use about $1 billion of the IPO proceeds to pay down debt but has indicated it will continue borrowing.

Despite being a key Nvidia customer with over 250,000 Nvidia GPUs deployed, CoreWeave has yet to turn a profit, making investors wary of its long-term viability.

AI Market Volatility and IPO Climate
The tepid reception to CoreWeave’s IPO raises concerns about the strength of the AI infrastructure market. Analysts suggest investors are recalibrating AI valuations amid uncertainty about data center spending. Additionally, competition from China’s DeepSeek, a low-cost AI rival, adds pressure to the sector.

CoreWeave had initially planned to sell 49 million shares to raise up to $2.7 billion, which would have valued it at $32 billion. However, due to mounting investor concerns, the company opted to scale back its IPO ambitions.

Despite the reduced offering, CoreWeave has secured significant partnerships, including an $11.9 billion infrastructure deal with OpenAI. The company will also issue $350 million in shares to OpenAI through a private placement.

Morgan Stanley, J.P. Morgan, and Goldman Sachs are leading the IPO underwriting.

Meituan to Invest “Billions” in AI Chips, Joining China’s AI Spending Surge

Meituan, China’s leading food delivery company, has announced plans to invest “billions” in chips for training artificial intelligence (AI) models, according to CEO Wang Xing during the company’s post-earnings call on Friday. This move aligns with the growing trend of significant AI investments by China’s tech giants.

The announcement comes as the company seeks to bolster its capabilities in AI, a sector rapidly gaining traction across industries. Meituan’s investment is part of a larger wave of funding within the tech sector, with companies like Alibaba also ramping up their AI-related expenditures. In February, Alibaba revealed plans to allocate at least 380 billion yuan toward cloud computing and AI infrastructure over the next three years.

These investments underscore the increasing competition in China’s tech industry as firms race to develop the next generation of AI technologies. Meituan’s focus on AI chip development signals its commitment to maintaining its leadership in food delivery and expanding into other AI-driven sectors.

Microsoft to Invest $300 Million in South Africa’s AI Infrastructure Expansion

Microsoft has announced plans to invest an additional 5.4 billion rand ($296.81 million) in South Africa by 2027 to expand its cloud and artificial intelligence (AI) infrastructure, catering to the increasing demand for Azure services in the region.

At a Johannesburg event on Thursday, Microsoft Vice Chair and President Brad Smith revealed the company’s strategy to support digital skills development. Microsoft will cover the cost of technical certification exams for 50,000 individuals in areas of high demand, including cloud architecture, AI, and cybersecurity.

This new investment builds on Microsoft’s previous expenditure of 20.4 billion rand, which was used to establish South Africa’s first enterprise-grade data centres in Johannesburg and Cape Town. These facilities have positioned the country as a critical hub for data centres to meet the growing computational needs of AI as businesses look to integrate the technology into their services.

Looking ahead, Microsoft plans to spend approximately $80 billion globally in fiscal 2025 to advance data centre infrastructure, with a focus on training AI models and deploying AI-powered applications and cloud services.