Deepseek-R1 is a blessing for corporate Apps cheaper, easier to build and innovative

The publication of the Deepseek R1 argumentation model has caused Shockwaves in the entire Tech industrysuddenly with the most obvious sign Sale of large AI shares. The advantage of well-financed AI laboratories such as Openaai and Anthropic no longer appears very solid, since Deepseek reports that their O1 competitors were able to develop to a fraction of the costs.

While there are currently some AI laboratories Crisis modeWith regard to the company sector, there are mostly good news.

Cheaper applications, more applications

As we had already said here, one of the Trends that are worth seeing in 2025 is the ongoing decline in the costs for the use of AI models. Companies should experiment and create prototypes with the latest AI models, regardless of the price, since they know that the continued price reduction enables them to finally use their applications on a scale.

This trend line was changing a big step. Openai O1 costs 60 USD per million output token against 2.19 per million for Deepseek R1 R1. And if you are worried Send your data to Chinese serverYou can access US providers on R1 Together.ai And Fireworks AiWhere it is at a price of $ 8 or 9 per million token – still a large bargain compared to O1.

To be fair, O1 still has the advantage over R1, but not so much that they justify such a large price difference. In addition, R1’s skills will be sufficient for most corporate applications. And we can expect that more progressive and capable models will be published in the coming months.

We can also expect the second order on the overall AI market. For example, Sam Altman, CEO von Openaai, announced that free Chatgpt users will soon have access to O3-Mini. Although he did not expressly mentioned R1 as the reason, the fact that the announcement was published shortly after the publication of R1 was published.

More innovation

R1 still leaves many questions unanswered – for example, there are several reports that Deepseek has trained the model on editions of Openai -Groß language models (LLMS). If his paper and the technical report are correct, Deepseek was able to create a model that almost corresponds to the state -of -the -art artist, while reducing the costs and removing some of the technical steps that require a lot of manual work.

If others can reproduce the results of Deepseek, it can be good news for AI laboratories and companies that were held by the financial obstacles to innovation in this area. Companies can expect faster innovation and more AI products to supply their applications with electricity.

What will happen to the billions of dollars that large technology companies have spent on the acquisition of hardware accelerators? We still have not reached the ceiling of what is possible with AI, so that leading technology companies will be able to do more with their resources. In fact, affordable AI will increase demand in medium to long-term increase.

However, it is even more important that R1 is proof that not everything is tied to larger computing clusters and data records. With the right technical chops and good talents, you can exceed the limits.

Open source for victory

In order to be clear, R1 is not completely open source, since Deepseek only published the weights, but not the code or the complete details of the training data. Nevertheless, it is a big victory for them Open Source Community. Since the publication of Deepseek R1, more than 500 derivatives have been published on the hug face, and the model has been downloaded millions of times.

There will also be more flexibility where they should run their models. Apart from the full model of 671 billion parameters, there are distilled versions of R1 that are between 1.5 and 70 billion parameters so that companies can operate the model for a variety of hardware. Over and beyond, In contrast to O1R1 shows his complete chain of thought and gives developers a better understanding of the behavior of the model and the ability to steer it in the desired direction.

With open source that obtain closed models, we can hope for renewing the commitment to exchange knowledge and research so that everyone can benefit from progress in AI.