Next year will be the year of reasoning
Posted: Tue Feb 18, 2025 2:54 am
Second, another headwind Nvidia faces is the high standard set by its astonishing performance growth over the past two years.
Analysts surveyed by FactSet generally believe that Nvidia's revenue will grow 55% to $191.45 billion in . In contrast, Nvidia's revenue is expected to more than double to $123.37 billion in .
The market's perception of the competitive landscape of AI chips is being revised, and there may be huge changes next year.
Broadcom's customized ASIC solution is emerging, which is somewhat usa phone number list similar to TSMC's challenge to traditional IDM back then.
More than two years after ChatGPT was launched, the main battlefield of AI is shifting from training to a larger and more competitive field: reasoning.
In recent months, many impressive events have occurred in the field of AI applications.
Open AI is currently holding 12 consecutive press conferences with shocking product updates. At the same time, Google has also recently launched a series of big moves, including Gemini 2.0, Veo 2, quantum chips... Google executives positioned Gemini 2.0 as an AI model for the era of intelligent agents.
Three months ago, Open AI released the o1 series, which took LLM's logical reasoning ability to a higher level through reinforcement learning. And, with more reinforcement learning training time computation and more thinking time test time computation, the performance of o1 continues to improve.
Analysts surveyed by FactSet generally believe that Nvidia's revenue will grow 55% to $191.45 billion in . In contrast, Nvidia's revenue is expected to more than double to $123.37 billion in .
The market's perception of the competitive landscape of AI chips is being revised, and there may be huge changes next year.
Broadcom's customized ASIC solution is emerging, which is somewhat usa phone number list similar to TSMC's challenge to traditional IDM back then.
More than two years after ChatGPT was launched, the main battlefield of AI is shifting from training to a larger and more competitive field: reasoning.
In recent months, many impressive events have occurred in the field of AI applications.
Open AI is currently holding 12 consecutive press conferences with shocking product updates. At the same time, Google has also recently launched a series of big moves, including Gemini 2.0, Veo 2, quantum chips... Google executives positioned Gemini 2.0 as an AI model for the era of intelligent agents.
Three months ago, Open AI released the o1 series, which took LLM's logical reasoning ability to a higher level through reinforcement learning. And, with more reinforcement learning training time computation and more thinking time test time computation, the performance of o1 continues to improve.