SAP: DeepSeek – What companies need to know about AI tools

DeepSeek has really shaken up the AI landscape with its new R1 model. According to the media, the Chinese artificial intelligence start-up’s developments could “revolutionize” the AI industry. As a result, the company’s chatbot app topped the list of the most downloaded free apps in the iOS App Store – even relegating ChatGPT to the ranks.

Meanwhile, the technology-heavy NASDAQ¬100 experienced a slump: investors reassessed the need for investment in AI technology and the index fell by almost three percent. With a dramatic slump of 17%, companies such as NVIDIA experienced one of the biggest one-day losses in stock market history. This was due to concerns about falling demand for high-performance microchips.

The DeepSeek strategy

DeepSeek is pursuing a special AI strategy that focuses primarily on cost efficiency. The company’s R1 model has been put on a par with established models such as OpenAI’s ChatGPT, but the development costs for R1 amounted to just 5 to 6 million US dollars. This is in stark contrast to the billions spent by competitors. This casts doubt on the prevailing assumption that powerful AI requires significant investment in advanced hardware.

As a result, investors have begun to rethink their strategies, leading to a significant drop in share prices for companies like NVIDIA and AMD that rely on high-performance microchips.

Cost-effective AI – what’s the catch?

As AI technology grows, companies need to manage the complexity of a variety of platforms while ensuring they maintain basic security measures.

“As technology evolves, cyber threats are also becoming more sophisticated – and with that comes increased responsibility for organizations,” says Gabriele Fiata, Head of Cybersecurity Market Strategy at SAP.

While the prospects of cost savings are tempting, there are several sensitive areas where companies should exercise caution to avoid higher costs down the line: 

1. Privacy and security

The privacy policies of many free AI tools state that they are trained with the information we provide them. This means that anything we enter into a tool can potentially be used by others, including our competitors. But it’s not just about sharing data: Data breaches or criminal activity can also pose a risk to data security.

“In 2025, this technology will be the focus of many organizations to ensure business continuity and the protection of sensitive data. It’s about using technology to provide the best possible protection,” emphasizes Fiata.

Companies need to strengthen their security measures to protect sensitive data from data breaches. Employees need to be sensitized to be extremely careful with the information they enter into these tools to avoid costly legal consequences and potential reputational damage.

2. Data storage concepts

Newer applications may not be compatible with older storage systems. Outdated storage systems can lead to high latency and inefficient operations, making it difficult to gain timely insights that are essential for decision-making. Another common problem is that a high number of users can lead to system overload or the tool not being available when it is needed.

3. Data quality

AI systems are only as good as the data used to train them. Poor quality data can lead to inaccurate results and operational problems. For example, if a retail company uses faulty customer data and therefore misinterprets buying trends, this could lead to poor inventory management and lost sales opportunities.

4. Strategic challenges

In some cases, companies are more interested in short-term savings than long-term growth potential. Choosing an AI approach that doesn’t take scaling opportunities into account could lead to problems and require a costly rethink when the organization’s business needs change.

5. Compliance challenges

“In an era of tightening data privacy regulations and growing geopolitical tensions, the ability to maintain control over data will become a critical factor for 2025,” Fiata explains. Companies will need to invest in compliance strategies to manage regulatory risks associated with the use of AI. And this can result in additional costs.

Customers need to be aware that while DeepSeek offers opportunities for innovation and cost savings in AI development, it also has some vulnerabilities. For organizations looking to innovate and save money, cost efficiency may be an attractive prospect, but it’s important to keep in mind the potential pitfalls associated with it as well.

New technologies are rapidly becoming an integral part of our everyday lives. Leveraging these advances effectively while reducing the associated risks will present many industries with major challenges in the next phase of AI development.

– – – – – –

Further links

👉 www.sap.com  
👉 KI in 2025: Five dominant topics 

Photo: pixabay