On Monday, OpenAI announced that it has secured $40 billion in a new investment round, valuing the ChatGPT manufacturer at $300 billion—the most extensive capital-raising session ever for a company.
The funding comes from a collaboration with Japanese investment firm SoftBank Group and “enables us to push the frontiers of AI research even further,” according to a post on the San Francisco-based company’s website.
“Their support will help us continue building AI systems that drive scientific discovery, enable personalized education, enhance human creativity, and pave the way toward AGI (artificial general intelligence), benefiting all of humanity,” the company said.
AGI refers to a computational platform that possesses human-level intelligence.
The company intends to expand its infrastructure and deliver increasingly powerful tools to the 500 million people who use ChatGPT on a weekly basis.
Opening up?
The investment announcement coincided with OpenAI’s revelation that it was developing a more open generative AI model. In the open-source field, it faced more competition from Chinese rivals DeepSeek and Meta.
The move would represent a strategic shift for OpenAI, which has previously been a staunch supporter of closed, proprietary models that prevent developers from modifying the fundamental technology to tailor AI more closely to their specific needs.
OpenAI and supporters of closed models, such as Google, have frequently criticized open models as riskier and more prone to harmful applications by hostile actors or non-US governments.
OpenAI’s embrace of closed models has also been a bone of contention in its battles with former investor Elon Musk, the world’s wealthiest person, who has called on OpenAI to honor the spirit of the company’s name and “return to the open-source, safety-focused force for good it once was.”
In response to OpenAI’s efforts, many large companies and governments have proven reluctant to build their AI products or services on models over which they have no control, especially when data security is a concern.
The core selling point of Meta’s family of Llama models or DeepSeek’s models is that they address these concerns by allowing companies to download their models and have greater control over modifying the technology for their specific purposes while retaining control of their data.
Meta CEO Mark Zuckerberg announced earlier this month that Llama had reached one billion downloads. At the same time, the release of DeepSeek’s lower-cost R1 model in January shook the world of artificial intelligence.
We’ve been thinking about this for a long time, but other priorities have taken precedence. ” Now it feels important to do,” OpenAI Chief Executive Sam Altman said on X regarding the decision to build a more open model.
OpenAI has been riding on the success of its latest image-generation features in ChatGPT, the world-leading AI app and chatbot.
Altman posted on Monday that the tool helped add “one million users” in one hour.
That claim came days after Altman stated that the new image features were so popular that they were overloading the OpenAI graphics processing units that power the AI due to heavy use.
