Yazılar

Surge AI Eyes Up to $1 Billion Capital Raise Amid Growth and Competition with Scale AI

/in /tarafından

Surge AI, a fast-growing data-labeling company competing directly with Scale AI, is reportedly preparing to raise as much as $1 billion in its first-ever capital fundraising, according to sources cited by Reuters. Founded by former Google and Meta engineer Edwin Chen, Surge AI aims for a valuation exceeding $15 billion, although talks remain in the early stages and the final amount could be higher. The planned funding round would include both primary capital to fuel growth and secondary capital to provide liquidity for employees.

Surge AI has achieved profitability and has been bootstrapped since its 2020 founding. It generated over $1 billion in revenue last year, surpassing Scale AI’s $870 million revenue for the same period. By comparison, Scale AI was last valued at $14 billion in a funding round last year, and more recently at nearly $29 billion following Meta’s strategic investment, which included hiring Scale’s CEO Alexandr Wang to lead Meta’s Superintelligence Labs.

The surge in interest for Surge AI coincides with a shift among some major AI customers, such as Google and OpenAI, who are reportedly moving away from Scale AI due to concerns about sharing sensitive research priorities with Meta, Scale’s largest investor. Despite this, Scale AI maintains its business remains strong and reassures clients about data protection.

Surge AI has grown quietly but rapidly, becoming a major player in the data labeling space, distinguished by its use of a network of highly skilled contractors rather than large pools of low-cost labor. Its premium services cater to leading AI labs including Google, OpenAI, and Anthropic.

As reinforcement learning from human feedback (RLHF) becomes critical for training advanced AI, the need for precise, nuanced data labeling has soared, benefiting companies like Surge AI. However, some investors remain cautious about the sector due to its traditionally low margins and reliance on human labor, which could face automation pressures as AI technologies advance.

Nvidia’s New AI Chips Slash Training Times for Massive AI Models

/in /tarafından

Nvidia’s latest generation of AI chips is making significant advances in training some of the world’s largest artificial intelligence systems, according to new benchmark data released on Wednesday by MLCommons, a nonprofit organization that tracks AI system performance.

The results show a dramatic drop in the number of chips required to train large language models (LLMs), highlighting Nvidia’s growing technological lead in this critical area of AI development. While much of the financial market’s current focus is on the booming sector of AI inference—where AI models answer user queries—training remains a core competitive battleground, especially for developing next-generation models with trillions of parameters.

Blackwell Chips Outperform Previous Generations

Nvidia’s new Blackwell chips demonstrated superior performance over its previous Hopper generation. In tests involving Meta Platforms’ open-source Llama 3.1 405B model, which is complex enough to simulate some of the most demanding AI training workloads, Nvidia’s Blackwell chips completed training tasks with more than double the speed per chip compared to Hopper.

In one benchmark, a system using 2,496 Blackwell chips completed the training run in just 27 minutes. By comparison, even though more than three times as many Hopper chips were used in previous tests, they only achieved faster results due to sheer scale rather than efficiency.

Nvidia and its partners were the only ones to submit data for models of this size, giving Nvidia a clear demonstration of its leadership in training capabilities for multi-trillion parameter models.

Changing Industry Trends in AI Training

Chetan Kapoor, chief product officer of CoreWeave, which collaborated with Nvidia on the results, noted that AI companies are moving away from building vast, homogenous data centers with 100,000 or more identical chips. Instead, they are increasingly assembling smaller, specialized subsystems that handle different aspects of the training process. This modular approach allows companies to speed up training times and manage extremely large model sizes more efficiently.

“Using a methodology like that, they’re able to continue to accelerate or reduce the time to train some of these crazy, multi-trillion parameter model sizes,” Kapoor explained at a press briefing.

Global Competition Also Heating Up

While Nvidia maintains a dominant position, competitors around the world are also pushing for breakthroughs. For example, China’s DeepSeek has recently claimed it can create competitive chatbots while using far fewer chips than many U.S. rivals, adding to the growing international race for AI supremacy.

MLCommons’ report also included results from Advanced Micro Devices (AMD) and others, though Nvidia’s Blackwell system stood out in the training category.

LinkedIn Lawsuit Over Customer Data Use for AI Models Dismissed

/in /tarafından

A class action lawsuit against Microsoft’s LinkedIn, which accused the platform of using customers’ private messages to train artificial intelligence models, has been dismissed. The case was dropped by plaintiff Alessandro De La Torre on Thursday in the U.S. federal court in San Jose, California, just days after the suit was filed. LinkedIn had argued that the allegations were unfounded.

De La Torre’s lawsuit claimed that LinkedIn violated the privacy of its Premium users by disclosing their private messages to third parties involved in developing AI. He accused the platform of breaching its promise to use customer data only to enhance its services, not for external uses like AI training.

The issue came to light when LinkedIn updated its privacy policy in September, revealing that a new account setting would not affect data used in previous AI training. This disclosure sparked concerns among users about how their data was being handled.

However, LinkedIn clarified that it had not shared private messages with third parties for AI training. In a LinkedIn post, Sarah Wight, the company’s vice president and legal counsel, confirmed, “We never did that.” De La Torre’s legal team acknowledged the clarification, stating that users could take comfort in knowing their private messages had not been used for AI purposes.