Sometimes the artificial intelligence in the world, the money flows like water. But even according to the high standards of the industry, the message of Openai On $ 40 billion in raising led by Softbank The Japanese, she is unprecedented. The company, which brought into the world ChatGPT The popular, $ 300 billion is currently valued – which places it as one of the world’s most expensive private companies, more than old grants such as Coca -Cola and Chevron.
The timing of the huge recruitment is not accidental. Just a few months ago, Chinese Deepseek Company has a earthquake in the world of technology when introducing an open artificial intelligence model called R1, which the fish advanced capabilities with significantly lower development than its American competitors. The response in Wall Street was acute and immediate – billions of dollars were deleted from the market value of the major technology companies, and the hysteria of the American industry realized – which feared the losing premiere in the strategic field.
Prominent Paradox – On the one hand, Chinese discontinuation has shown that advanced artificial intelligence models can be developed at a relatively low cost, and on the other, OpenAI raises $ 40 billion. So what is behind him? Bloomberg is exposed to a challenging economic image – despite $ 12.7 billion in revenue this year, the company does not expect positive cash flow by 2029, which explains the need for a lot of capital. However, the real story is the infrastructure.
Lack of training processors and to run models
Much of the money, almost half of it for accurate, about $ 18 billion, is designed for the “Stargate” venture – a huge $ 500 billion project declared last January, aimed at setting up huge databases for AI in the US, which served the needs of the future. Dabi.
This is a definite move for Depth’s ‘Practice’ approach – an approach in which “distiller” is a large model for a smaller and more economical model that preserves most capabilities. While the Chinese show tactical efficiency through the same “smart recycling” of existing technologies, OpenAi gains a huge investment in long -term infrastructure.
“This investment helps us to push the boundaries and make artificial intelligence more useful in everyday life,” wrote Sam Altman, OpenAi CEO, in the official announcement of recruitment. Missimoshi San, founder and CEO of Softbank, called the same message as “the power of the” Definite “, Ours to release the full potential of this technology. ”
Uri Eliaviev, a consultant in the field of artificial intelligence, also believes that the recruitment of OpenAi no paradox is also not in recruitment, and he explains this through Altman’s latest tweets on the X, after the huge success recorded by the new photo generator in Chatgpt.
“You can relax with the photo creation? It’s crazy, our team needs to sleep,” Etman wrote in the context of Studio Jibli Anime Style. Another tweet wrote that “it is fun to see that people love the photos in Chatgpt, but our processors are melting. We are about to temporarily bring rhythm limitations as we work on efficiency. Hope it won’t take a long time.”
Eliaviev analyzes the situation this way – despite the progress of competitors such as Detah, OpenAi is still dealing with a significant shortage of GPU processors, both for training and to want the models. Already since the chat was launched, he broke the Internet and became a product with the world’s fast growth rate in no time. Therefore, the company still needs a huge capital to acquire severity, recruit the leading researchers in the field, purchase companies, and the development of its independent chipset. Eliaviev explains that “there is no paradox here. You might argue if the esteemed value is too high and if the company really needs all this money, but in fact its financing needs are indeed enormous – the purchase of GPU processors, excellent recruitment and leading researchers or the acquisition of companies and development of independent chip.”
While giants recruit billions and invest huge sums in computing infrastructure, small companies and startups in the field find themselves for a broken trough. OpenAI’s recruitment marks a leap in market concentration, when Gil Lurie said an analyst at DA Davidson & Co, “there are not many companies that can raise $ 40 billion, or even a tenth of that amount.” He adds that “the list of investors who want to support such a scope has shrunk and may be mostly limited to the Spatfank, which may not have the necessary capital.”
This means that small companies in the field should search and find unique niches and benefits to continue surviving.
On Eliaviev’s, “Openai herself was a small company that came to Microsoft giant and asked her for funding. However, he adds that “on the other hand, it is important to remember that less than a year ago, OpenAi was almost broken up following the failed impeachment attempt of Sam Altman. In the business world, it is impossible to predict for sure what will happen in the future.”
Innovation that may change the rules of the game
However, and in the face of replica recruits – the same technological innovation called “Distillation” may, as in the case of Detah, change the rules of the game and provide an opening for hope for small companies. In simple terms, the refinery technique is parallel to the transfer of an experienced teacher to a student – take a large, expensive artificial intelligence model (the teacher) and use it to train a smaller and more economical model (the student).
That is, instead of investing hundreds of millions of dollars in model training from the beginning – which only huge companies can afford – small companies can use the refinery technique to create much lower cost models. This is exactly what the Chinese discontinuation did, when it was built strong open sources based on its competitors. “Whenever you can lower costs and the result still gives you the desired performance, there are very few reasons not to do so,” said David Cox, Vice President of Artificial Intelligence Models in IBM Research, to Finnent Times. He added that “most businesses do not need a huge model to operate their products, and distilled models are strong enough for purposes such as customer service chasshots or running on smaller devices like phones.”
This method is concerned about the big companies. “Openai has been trying to protect itself from refinery for a long time, but it is very difficult to avoid it completely,” indicates Daha Kyila, CEO of Contactual AI, a startup that builds information recovery tools for organizations, in a conversation with the Financial Times. Data for training competing models.
On the other hand, those who bless this development are supportive in open models. “We are going to use ‘refining’ and put it in our products immediately,” Jan Lakon recently declared, the Meta’s main artificial intelligence scientist. He also said that “this is the whole idea of open source. You earn from everyone’s progress and everyone else, as long as these processes are open.”
Company Access Change: Not a closed code anymore
As mentioned, the surprise that has not yet been made from the Chinese model led Altman to admit to a Berdit public forum that “Openai was on the wrong side of history” when it comes to open source. Along with the recruitment announcement, Altman declared in a network X that the company will soon release a language model with “visible weights”, which means it will reveal to the public the “finished product” of the model – the final weights (a kind of “ready -made recipe”) – but leave the technology and processes that created it confidential, which is not open.
Developers will be able to use the model and even adjust it to some extent, but will not be able to build just as it is from scratch. The move represents a significant turnaround from the traditional and closed access of OpenAI – which charges payment of users and organizations through subscribers to its products, as well as payment from developers through access to its programming interface.
At the same time, competition in the artificial intelligence market already affects business decisions of leading technology companies. Cyber Company Palo Alto Networks, for example, decided to move from dear OpenAI models to the open code model of the Chinese Detah itself – a model that provides similar performance at only 5% of the price offered by Openai. As the company’s CEO stated is a damn “for our tasks, I don’t need a particularly high IQ model. I’d rather pay five cents instead of dollars. ”
Please note: The Globes system strives for a diverse, matter -of -fact and respectful discourse according to the ethical code that appears in the trust report by which we act. Expression of violence, racism, incitement or any other inappropriate discourse is filtered in a way Automatically And they will not be published on the site.