Yazılar

Microsoft Unveils Copilot Chat to Accelerate AI Adoption for Businesses

Microsoft launched a new AI service, Copilot Chat, on Wednesday, aiming to help businesses integrate on-demand AI agents for routine tasks. The service, which uses OpenAI’s GPT-4, is designed to make AI more accessible with a pay-as-you-go model that could drive widespread adoption. Copilot Chat allows users to create AI agents in natural languages, such as English and Mandarin, for tasks including market research, writing strategy documents, and meeting preparation.

While the basic service is free, additional features like summarizing and transcribing Teams calls and creating PowerPoint slides are available through a $30 monthly Microsoft 365 Copilot subscription. Microsoft, which has invested heavily in AI infrastructure, including a projected $80 billion this fiscal year on data centers and AI-related projects, is under pressure to demonstrate the financial return on these investments.

Despite some initial skepticism about Copilot’s uptake, as highlighted by a Gartner report, Microsoft has continued to promote the service. In November, the company started enabling customers to create autonomous AI agents requiring minimal human oversight, a move that some analysts see as a potential way for tech companies to monetize AI more efficiently.

 

Emerging AI Investment Opportunities Beyond Big Tech

The ongoing artificial intelligence (AI) revolution, described as the “biggest platform shift since electricity,” is predicted to create lucrative opportunities for smaller tech firms, according to Clare Pleydell-Bouverie, co-lead fund manager at Liontrust Asset Management.

In an interview with CNBC, Pleydell-Bouverie emphasized that the dominant players of the last tech cycle—referred to as the “Magnificent Seven,” which includes Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla—may not necessarily lead in this new phase of technological transformation. Instead, emerging firms focused on AI applications and infrastructure are poised to become significant players.

“This year, we’ve concentrated on the AI infrastructure layer, which is essential for scaling this technology,” Pleydell-Bouverie stated. She highlighted sectors like silicon chip production, semiconductor equipment manufacturing, and network infrastructure as critical to enabling AI growth. Companies like Broadcom, Amphenol, and Arista Networks are vital in building the foundational layers of AI’s technological stack.

The “stack,” as described by Pleydell-Bouverie, includes several layers:

  1. AI Infrastructure: Firms providing hardware and connectivity, such as chips, cables, and networks.
  2. Foundation Model Providers: Companies creating large-scale machine-learning models, which she characterized as highly competitive and commoditized.
  3. AI Engineering Firms: Those enabling businesses to integrate AI into their operations and services.

While the infrastructure layer currently holds the most value, Pleydell-Bouverie foresees this shifting toward application and integration in the near future.

Nvidia’s Strategic Position in AI
Nvidia remains a standout in the AI space, which Pleydell-Bouverie compares to Apple’s dominance during the smartphone revolution. However, she argues that Nvidia is often misunderstood as merely a chip provider.

“Nvidia is positioning itself as the operating system for the next generation of AI-infused software,” she noted, pointing to the company’s strategic shift toward integrating software and hardware to power AI applications. Nvidia’s shares have surged by over 180% in 2024, fueled by demand for its advanced AI chips like Blackwell.

Pleydell-Bouverie sees Nvidia as the primary beneficiary of the AI boom in 2025, likening its current trajectory to Apple’s rise under Steve Jobs, who combined hardware innovation with software integration to dominate the tech landscape.

As AI continues to redefine industries, investors are encouraged to look beyond traditional Big Tech giants and explore opportunities in emerging firms that are reshaping the AI ecosystem.

 

Cerebras IPO Faces Challenges as AI Chipmaker Tries to Rival Nvidia

Cerebras Systems, an artificial intelligence (AI) chipmaker, is aiming to be the first major tech company in the U.S. to go public since April 2024. As the AI infrastructure market continues to boom, driven by the astronomical success of Nvidia, Cerebras hopes to ride the wave of investor enthusiasm for AI technologies. However, the company faces significant hurdles, particularly its heavy reliance on a single customer and the risks associated with foreign investment, which could hinder its IPO.

The Cerebras Opportunity

Founded in 2016 and headquartered in Sunnyvale, California, Cerebras has positioned itself as an innovative player in the AI hardware market. The company claims its chips outperform Nvidia’s GPUs in training large language models—an essential task in AI. In fact, Cerebras has seen impressive growth, with sales tripling to $78.7 million in 2023, and revenue reaching $136.4 million in the first half of 2024. The company has secured contracts to sell $1.43 billion in AI systems and services, a significant boost for its growth trajectory.

Cerebras’ most advanced chip, the WSE-3 (Wafer Scale Engine 3), is hailed as the fastest AI processor globally, boasting 4 trillion transistors. It’s designed to handle massive computational tasks with greater efficiency than competing chips, including Nvidia’s.

Customer Concentration and Foreign Investment Concerns

Despite its technological advances, Cerebras’ reliance on a single client raises red flags for investors. G42, a company based in Abu Dhabi, accounted for 87% of Cerebras’ revenue in the first half of 2024. G42, backed by Microsoft, has committed to purchasing $1.43 billion worth of systems from Cerebras. However, this heavy dependence on one customer leaves the company vulnerable to revenue fluctuations if G42 fails to meet its obligations.

Moreover, G42 is also a significant investor in Cerebras. The Middle Eastern company has agreed to purchase a $335 million stake by April 2024, which would make it Cerebras’ largest shareholder. G42 can buy an additional $500 million in shares if it commits to spending $5 billion on Cerebras’ computing clusters. This foreign investment has raised concerns with U.S. regulators. Cerebras is currently seeking approval from the Treasury Department’s Committee on Foreign Investment in the U.S. (CFIUS) to ensure there are no national security risks associated with G42’s involvement.

The CFIUS review could potentially delay Cerebras’ IPO, as reported by sources familiar with the matter. Adding to the uncertainty, U.S. lawmakers have expressed concerns about G42’s previous business ties with Chinese companies, raising further scrutiny of its involvement.

Skepticism Among Wall Street Banks

Major Wall Street banks have been noticeably absent from Cerebras’ IPO deal, which is being led by Citigroup and Barclays. Banks such as Goldman Sachs, JPMorgan Chase, and Morgan Stanley, which typically dominate tech IPOs, have stayed away, reportedly due to the risks tied to Cerebras’ customer concentration and foreign investment concerns. The absence of these banks may further erode investor confidence.

In addition, Cerebras’ auditor, BDO, is not part of the Big Four accounting firms, which typically audit the largest tech IPOs. This adds another layer of uncertainty for potential investors.

Competitive Landscape and Financial Struggles

Cerebras is entering a highly competitive space dominated by Nvidia, which controls 95% of the AI chip market, according to Mizuho Securities. Nvidia’s chips are critical to training models like OpenAI’s GPT-4, and the company is seen as a cash machine in the AI boom. Nvidia’s strong market position and profitability present a significant challenge for Cerebras, which is still not profitable.

In the second quarter of 2024, Cerebras posted a net loss of $51 million. Although it is approaching breakeven on an operating basis (excluding stock-based compensation), the company still faces financial headwinds as it tries to scale up.

Investor Outlook

Despite these challenges, some investors are intrigued by Cerebras’ potential. For instance, Jim Fitch, a retired investor, sees promise in the company’s AI chip technology. Fitch, who sold his Nvidia stock years ago, is excited about the WSE-3 chip, believing it could revolutionize AI computing.

While Cerebras has hurdles to overcome—including regulatory scrutiny, customer dependence, and competition from Nvidia—the excitement surrounding AI technologies keeps the door open for a successful IPO. If the company can navigate these challenges and tap into the growing demand for AI infrastructure, it may still find a path to Wall Street, despite the risks involved.