Yazılar

SoftBank in Talks to Invest Up to $25 Billion in OpenAI

/in /tarafından

SoftBank is in discussions to invest between $15 billion and $25 billion in OpenAI, the owner of ChatGPT, according to a source familiar with the matter. The move is part of the Japanese conglomerate’s ongoing efforts to expand its footprint in the AI sector. Some of the investment may go towards funding OpenAI’s commitment to Stargate, a joint venture between Oracle, OpenAI, and SoftBank.

The Stargate project aims to secure up to $500 billion in investments to keep the U.S. ahead of China and other global competitors in the AI race. SoftBank has already committed $15 billion to Stargate and is exploring additional financing options for the project. The investment talks with OpenAI are still in early stages, according to the source.

Earlier reports indicated that SoftBank is considering a total investment of $40 billion into both Stargate and OpenAI, including borrowing up to $18.5 billion in financing, secured by its publicly-listed assets. The venture was unveiled at a White House event last week, attended by U.S. President Donald Trump, SoftBank CEO Masayoshi Son, OpenAI CEO Sam Altman, and Oracle Chairman Larry Ellison.

However, SoftBank’s stock price has taken a hit following a market selloff triggered by the rise of DeepSeek, a Chinese AI startup that introduced a free AI assistant developed at a fraction of the cost of its U.S. competitors. This shift has raised concerns about the sustainability of business models that rely on expensive AI technology and contributed to a decline in SoftBank’s shares, which were down 1% in morning trading on Thursday.

Despite this, SoftBank remains committed to strengthening its position in the AI sector. The company already holds a $1.5 billion stake in OpenAI, which was valued at $157 billion in its most recent funding round. SoftBank and OpenAI have yet to comment on the reported investment discussions.

 

Meta Reports Strong Q4 Sales But Forecasts Muted Outlook Amid AI Investments

/in /tarafından

Meta Platforms (META.O) exceeded Wall Street expectations for its fourth-quarter revenue, reporting $48.4 billion, which outpaced analysts’ predictions of $47 billion. Despite this strong performance, the parent company of Facebook and Instagram issued a cautious outlook for the first quarter of 2025, with expected revenue between $39.5 billion and $41.8 billion, slightly below the analysts’ consensus of $41.72 billion.

Meta’s CEO, Mark Zuckerberg, struck an optimistic tone during a conference call, focusing on the company’s AI initiatives. He expressed confidence in the open-source AI strategy, commenting that the emergence of Chinese startup DeepSeek’s AI models reinforced his belief in this direction. “There’s going to be an open-source standard globally,” Zuckerberg remarked, emphasizing that he sees it as an American standard.

However, Meta’s outlook raised questions about the company’s capital spending, as the company relies on its core social media advertising business to finance its significant AI and “metaverse” investments, such as smart glasses and augmented reality systems. Meta has already committed up to $65 billion in capital expenditures for 2025 to expand its AI infrastructure, alongside increasing AI-related hires. The company also anticipates total expenses for 2025 to be between $114 billion and $119 billion, up from $95 billion in 2024.

Jeremy Goldman, principal analyst at eMarketer, noted that while Meta’s fourth-quarter results were strong, the key issue going into 2025 is whether its substantial AI investments will pay off. He highlighted that Meta’s family daily active people (DAP) metric, which tracks unique users who open any of its apps in a day, grew by 5% year-over-year to 3.35 billion.

Meta’s results came after DeepSeek, a Chinese AI company, launched new models that outperformed top U.S. competitors at a fraction of the cost, fueling concerns about the sustainability of U.S. AI business models and triggering a tech stock selloff. Zuckerberg acknowledged the challenges posed by DeepSeek but said it was too early to determine how its global impact would affect Meta’s AI strategy. He added that Meta’s AI teams were already incorporating insights from DeepSeek’s models into their work.

Meta continues to face financial losses in its metaverse-focused Reality Labs unit, which, while exceeding sales expectations, still reported a $5 billion loss in Q4. Despite these losses, Zuckerberg believes the long-term business potential of AI will become evident after 2025.

Meta’s hefty investment in AI includes plans to acquire more of Nvidia’s AI chips and develop custom silicon to train AI systems for better feed recommendations by next year. CFO Susan Li confirmed the company’s goal to use its own chips for these AI tasks. Meta’s continued investment in AI is part of its broader strategy to enhance user experiences across its platforms, while also aiming for cost reductions in AI model development.

 

DeepSeek’s Global Success Sparkles with National Pride in China

/in /tarafından

Chinese bloggers, state media, and citizens are celebrating the rapid success of DeepSeek, the homegrown AI startup, viewing its rise as a symbol of China’s resilience against Western efforts to constrain its tech industry. Last week, DeepSeek launched a free AI assistant that claims to use less data at a fraction of the cost of competing services. By Monday, it had surpassed U.S. rival ChatGPT in downloads on Apple’s App Store, prompting a significant selloff in tech shares globally.

DeepSeek’s ability to rival the capabilities of OpenAI while offering a more affordable alternative has raised concerns about the sustainability of profit margins and business models of U.S. AI giants such as Nvidia and Microsoft. In China, the startup’s success has been seen as a victory against U.S. efforts to block access to advanced semiconductors, which are critical for AI development.

“This symbolizes that U.S. containment, persecution, and sanctions in advanced technology against China have completely failed,” wrote military commentator Chen Xi on his WeChat account. This sentiment aligns with statements from former U.S. President Donald Trump, who suggested that DeepSeek’s achievements should spur American firms to innovate and that it was beneficial for Chinese companies to introduce cheaper, faster AI technology.

In Zhejiang, where DeepSeek is based, the provincial government’s media office published a widely shared essay celebrating the company’s success. The article, read more than 100,000 times, declared, “The moon overseas is not actually more round. Whatever others can do, we can also do—and even do it better.” The essay pushed back against both overly optimistic and overly pessimistic views of China’s technological progress.

This wave of pride surrounding DeepSeek mirrors the response to Huawei’s surprise launch of the Mate 60 Pro smartphone in 2023, which came despite U.S. sanctions. At that time, the state-backed Global Times praised Huawei’s ability to produce high-end smartphones, arguing that the U.S. crackdown had failed.

The reaction from the Chinese public has been equally supportive. Chen Jianuo, a 38-year-old Beijing resident, expressed pride over DeepSeek’s international success, reflecting on the positive global attention the company has garnered. “China has made great progress in AI development, and I hope our technological growth continues,” she shared.

Leo Li, a 24-year-old student, also voiced his pride, saying, “It’s exciting that a Chinese company is on par with Meta and OpenAI. As a Chinese citizen, it feels great to see our AI research becoming a global sensation.”