根据IT之家的独家报道，科技媒体The Information昨日披露了OpenAI公司在2024年面临的严峻财务挑战。据称，OpenAI公司当前在AI推理计算上的需求，预估相当于35万台搭载英伟达A100芯片的服务器，其中约29万台用于运行ChatGPT，且这些服务器几乎处于满负荷运行状态。至今年3月，OpenAI公司已经花费近40亿美元的巨额资金用于租用微软的服务器进行ChatGPT的推理工作负载。此外，预计今年OpenAI在训练ChatGPT和新AI模型上的开销将高达30亿美元，从而导致训练和推理AI的总成本可能超过70亿美元（当前约509.02亿元人民币）。
这一系列的高额支出对OpenAI的财务状况构成了巨大压力。据称，OpenAI公司目前的员工规模约为1500人，员工成本预估达到15亿美元（当前约109.08亿元人民币）。尽管如此，OpenAI通过ChatGPT获得了约20亿美元的年收入，并可能从语言模型LLMs的收费访问中获得近10亿美元的额外收入。然而，考虑到OpenAI最近每月的总收入为2.83亿美元，其全年销售额估计在35亿至45亿美元之间，这与高昂的支出相比，显示出OpenAI在2024年面临着50亿美元的资金缺口。
面对这一困境，有消息称OpenAI很可能在未来12个月内启动新一轮的融资，以填补巨大的资金缺口并维持公司的运营和发展。这一系列的披露不仅揭示了AI领域尤其是大型语言模型训练所需资源的惊人消耗，同时也反映了AI公司如何在技术创新与财务可持续性之间寻求平衡的挑战。
英语如下：
News Title: “OpenAI Faces $5 Billion Funding Gap in 2024, Likely to Launch Fresh Round of Financing”
Keywords: OpenAI, Funding Gap, AI Costs
Content: According to an exclusive report by IT House, the technology media The Information disclosed yesterday the serious financial challenges that OpenAI, a leading AI company, is expected to face in 2024. It is reported that OpenAI’s current demand for AI inference computing is estimated to be equivalent to 350,000 servers equipped with Nvidia A100 chips, with approximately 290,000 of them running the ChatGPT model. These servers are running at nearly full capacity. By March, OpenAI had spent nearly $4 billion in huge funds to rent Microsoft’s servers for the inference workload of ChatGPT. In addition, it is expected that this year, OpenAI will incur costs of up to $3 billion for training ChatGPT and new AI models, leading to a total cost of training and inferring AI that may exceed $7 billion (approximately 509.02 billion RMB).
These high expenditures have put a significant strain on OpenAI’s financial situation. It is reported that OpenAI currently has an estimated 1,500 employees, with employee costs reaching $15 billion (approximately 109.08 billion RMB). However, despite this, OpenAI has generated approximately $2 billion in annual revenue from ChatGPT and may earn an additional $1 billion from the subscription fees for accessing language models like LLMs. Nonetheless, considering that OpenAI’s monthly total revenue recently stood at $2.83 billion, its estimated annual sales range between $35 billion and $45 billion, this contrasts sharply with the high expenditures, indicating that OpenAI is facing a $5 billion funding gap in 2024.
In the face of this predicament, there are reports suggesting that OpenAI is likely to initiate a new round of financing within the next 12 months to fill the significant funding gap and sustain its operations and growth. These revelations not only shed light on the astonishing resource consumption in the AI field, particularly in the training of large language models, but also highlight the challenge AI companies face in balancing innovation with financial sustainability.
【来源】https://www.ithome.com/0/784/152.htm
Views: 1