AI inference and its significance in technological progress
Inference is defined by the dictionary as “a conclusion derived from evidence and reasoning.” In the realm of artificial intelligence (AI), however, inference acquires a different significance. It describes the mechanism by which an AI model generates predictions or conclusions derived from data.
As per cybersecurity firm Cloudflare, an AI model that can perform inferences does so without requiring instances of the expected outcome. Essentially, inference represents an AI model in operation. A notable illustration of this would be a self-driving vehicle identifying a stop sign on a route it has never traversed previously.
AI inference plays a vital role in the evolution of AI technologies and serves as the foundation for its most thrilling applications, such as generative AI, which fuels tools like OpenAI’s ChatGPT. IBM has observed that AI models depend on inference to simulate human cognition, reasoning, and reactions to queries. Nevertheless, as a rapidly developing technology, AI inference also encounters hurdles, such as regulatory matters related to data sovereignty and a deficit of skilled personnel within the industry.
IBM further emphasized the global dependence on Taiwan for semiconductor production, where 60% of the world’s semiconductors and 90% of advanced chips, including AI accelerators, are manufactured. This dependence presents a potential threat to the AI sector, particularly in light of the geopolitical strains in the area.
Broadcom’s approach to AI and market predictions
Broadcom has been strategically positioning itself to leverage the AI surge, with a specific emphasis on AI inference. The company’s partnership with OpenAI to create a new AI chip is a clear indication of its desire to lessen dependence on Nvidia’s GPUs, which currently lead in the AI training sector. This strategy could potentially transform the competitive environment, as Broadcom aims to secure a larger footprint in the AI hardware market.
CEO Hock Tan has been optimistic about Broadcom’s AI potential, predicting AI-chip sales to reach billion in fiscal 2024, up from a prior estimate of billion. This enthusiasm is mirrored in the company’s stock performance, with shares soaring 55% year-to-date. Investors appear to have faith in Broadcom’s AI strategy, especially as the company continues to attract high-profile clients like OpenAI.
Nevertheless, Broadcom faces its share of hurdles. Analysts have pointed out potential issues ahead, including seasonal variances and a product transition concerning Google’s Tensor Processing Units (TPUs). Furthermore, alterations in Apple’s Wi-Fi sockets could affect Broadcom’s revenue, as the company is a principal supplier of chips that link Apple devices to mobile networks. Bank of America Securities recently reasserted its buy rating for Broadcom but adjusted its fiscal 2025 earnings forecasts downward in light of these challenges.
In spite of these difficulties, Broadcom’s AI networking segment is anticipated to benefit from Nvidia’s upcoming next-generation Blackwell GPU, which is poised to increase production in the next few years. This could offer a substantial boost to Broadcom’s financial performance, as the company continues to broaden its array of custom AI clients.
Looking forward, Broadcom intends to transition from annual to quarterly guidance, a shift that may introduce greater volatility into its earnings reports. However, analysts believe the company will persist in conveying its strong AI positioning, whether by disclosing AI as a percentage of sales or presenting a defined range for fiscal 2025. This openness could alleviate some of the uncertainty surrounding the company’s future outcomes.
While the comprehensive effect of Broadcom’s collaboration with OpenAI may not become apparent until fiscal 2026, analysts view it as a potential benefit to their predictions. As AI continues to transform industries, Broadcom is well-equipped to take advantage of the increasing demand for AI chips and infrastructure. For Australian investors, this offers a chance to gain exposure to the AI transformation through a company that is not only surfing the wave but is also actively influencing the future of the sector.