➀ Chuan expresses doubts about the success of OpenAI's hardware project due to its negative cash flow of nearly $5 billion per year;

➁ Ming-Chi Kuo's industry investigation suggests that the new AI hardware device developed by Jony Ive and OpenAI is expected to be mass-produced in 2027;

➂ The device will be assembled and shipped outside of China to reduce geopolitical risks, with Vietnam as the current planned location;

➃ The current prototype is slightly larger than the AI Pin, with an exterior design reminiscent of the iPod Shuffle, but designs and specifications may change before mass production.

➀ In Q1'25, global venture funding reached $121 billion, driven by OpenAI's $40 billion raise, valuing the company at $300 billion;

➁ Despite a decline in deal count, AI startups remained the primary beneficiaries of this capital concentration, capturing more than half of the quarter's funding;

➂ The AI dealmaking landscape is evolving, with a shift from infrastructure investments to vertical solutions and application-layer platforms.

➀ The dilemma facing Big Tech is the cost of building AI, with estimates ranging from tens of millions to tens of billions;

➁ Some investments in cheap, effective Chinese products have raised concerns about the wisdom of some investments;

➂ Microsoft is withdrawing from data center projects in various countries, and expectations for GenAI are declining due to high failure rates and dissatisfaction with current results;

➃ OpenAI is not yet profitable and expects positive cash flow only in 2029, despite spending $644 billion on AI programs and data centers.

OpenAI's release of GPT-4.5 has been delayed due to a shortage of GPUs. CEO Sam Altman mentioned on X that the rollout will be staggered and will involve adding tens of thousands of GPUs next week. This shortage is prompting OpenAI to develop its own AI silicon in partnership with Broadcom.

The high cost of GPT-4.5, at $75 per million input tokens and $150 per million output tokens, is a concern, but Altman emphasizes its unique intelligence.

NVIDIA's GPUs are in high demand, with Blackwell GPUs sold out until October. The expansion of AI infrastructure, including massive AI supercomputers and data centers, is driving this demand.

➀ The National Supercomputing Internet Platform announced the 'AI Ecosystem Partner Acceleration Program', offering free 3-month use of DeepSeek API and a 10 million card-hour computing power pool.

➁ OpenAI CEO Sam Altman emphasized that the company does not sell and expressed willingness to acquire Twitter if Musk is willing to talk.

➂ Altman also expressed willingness to cooperate with China in the field of artificial intelligence and said it is very important.

➀ OpenAI plans to tape-out its first chip on TSMC's 3nm process later this year;

➁ The chip is designed for both training and inference and is being developed by an in-house team;

➂ The objective is to reduce dependency on Nvidia's processors, which are expensive and in high demand.

➀ OpenAI's o3-mini and DeepSeek R1 are leading AI inference models, differing in design, performance, cost, and methods.

➁ o3-mini offers advanced reasoning, quick response times, and is integrated into ChatGPT, with a cost of about $1.10 per million input tokens and $4.40 per million output tokens.

➂ DeepSeek R1 is open-source, cost-effective, and shows its reasoning process, with a cost of about $0.14 per million input tokens for cache hits and $2.19 per million output tokens.

➀ On January 22, 2025, President Donald Trump announced a $500 billion investment in AI infrastructure at the White House.

➁ This project, called The Stargate Project, aims to build new AI infrastructure in the United States over the next four years, led by SoftBank and OpenAI.

➂ The initiative is expected to secure American leadership in AI, create jobs, and enhance national security.

➀ NVIDIA将为Stargate AI超级计算机提供64,000个新的GB200 AI GPU，以支持OpenAI和Oracle；

➁ 这些GPU预计将在2026年交付，总价值22.4亿美元；

➂ Stargate AI超级计算机位于德克萨斯州阿比林，将与Elon Musk的xAI超级计算机竞争。

➀ OpenAI CEO Sam Altman宣布GPT-4.5已准备好；

➁ GPT-4.5是OpenAI迄今为止最大的聊天模型，具有更高的复杂性和改进；

➂ GPT-4.5在无监督学习、模式识别、建立联系和生成创意见解方面表现出色；

➃ GPT-4.5模型庞大且运行成本高，目前无法对所有用户开放。

➀ Meta plans to launch a standalone AI app this year to establish itself as a leader in AI, prompting a reaction from OpenAI.

➁ The app is part of CEO Mark Zuckerberg's strategy to position Meta at the forefront of the AI industry.

➂ OpenAI, the creator of ChatGPT, has noticed Meta's plans and CEO Sam Altman hinted at a possible new social app from OpenAI.

➀ Sam Altman 宣布 OpenAI 取消了 o3 模型的全面发布，理由是模型和产品复杂性的增加，并希望简化产品线。

➁ OpenAI 将所有精力集中在 GPT-5 的发布上，该模型将融合 o3 的元素。

➂ Altman 上个月警告 AI 粉丝，不要对 OpenAI 向超级智能发展的进展抱有过高期望。

➀ DeepSeek is rising to a leading position in global AI chatbot rankings, as customers shift away from OpenAI's ChatGPT.

➁ The AI market is experiencing significant changes due to the introduction of DeepSeek, causing fluctuations in stock values and probes into potential trade violations.

➂ Data shows that ChatGPT's user base in the US has decreased while DeepSeek's has surged, indicating a shift in user preference.

➀ OpenAI has released a new AI-powered tool called Deep Research on February 2nd, which can generate comprehensive reports within 30 minutes.

➁ Deep Research is available to ChatGPT Pro users and is designed for complex inquiries in fields such as finance, science, policy, and engineering.

➂ The tool emphasizes precision and reliability, and aims to limit the hallucinations common in conventional language models.

➀ OpenAI has launched the o3-mini, its most cost-efficient and powerful model in the reasoning series.

➁ The new model responds 24% faster than o1-mini and provides more accurate answers.

➂ o3-mini supports function calling, structured outputs, and developer messages, making it production-ready.

➀ Microsoft provides unlimited free access to OpenAI's o1 model, which was released in December and costs up to $200 per month.

➁ The o1 model is designed to be more thoughtful and implement a step-by-step chain-of-thought process, making it excel in specific tasks and categories of questioning.

➂ Microsoft's Copilot users can now access this new functionality by selecting the 'Think Deeper' toggle option.

➀ OpenAI CEO Sam Altman commented on the market hype around the new DeepSeek AI model.

➁ The DeepSeek R1 model, created for $6 million using old NVIDIA GPUs, has raised questions about the billions of dollars invested in AI development.

➂ OpenAI CEO Sam Altman acknowledged the impressive nature of the DeepSeek R1 model and expressed excitement about continuing their research roadmap.