As 2025 dawns, a transformative wave triggered by AI models has swept across various industries, igniting a remarkable change in productivity levelsWhat began as a whimsical online exploration with Deepseek has evolved into serious endeavors by leading Chinese tech giants, aiming to transition the “national fortune” level tool from a mere internet sensation to a widely used AI application that serves a collective public good.

From cloud computing services to smartphones and automobiles, companies ranging from Huawei, Tencent, and Alibaba to ByteDance have plunged into the frayThe arrival of AI into the realm of commercial innovation has sparked a fierce battle for supremacy in improving productivity, showcasing the sweeping influence of this transformative technology.

In early February, while most of the nation was still basking in the festive post-New Year glow, cloud service providers were scrambling to catch up with the burgeoning popularity of DeepseekOn February 1, Huawei Cloud led the charge by announcing the launch of DeepseekR1/V3 inference services based on its Ascend Cloud serviceThe very next day, Tencent Cloud revealed a one-click deployment feature for the Deepseek-R1 model, making it available to developers in as little as three minutes.

Just a few days later, on February 3, Baidu’s Intelligent Cloud Platform officially integrated Deepseek-R1 and Deepseek-V3 modelsOn the same day, Alibaba Cloud announced support for the one-click deployment of these models on their PAI Model GalleryBy February 4, Volcano Engine also announced it would support various sizes of the Deepseek open-source models, including the full-powered R1 version.

For casual internet users still engaged in playful interactions with Deepseek, the rapid response from cloud service providers might have seemed less compelling initiallyHowever, as smartphone manufacturers and automakers joined the fray, the scale of the “Deepseek integration wave” began to expand exponentially.

On February 5, the first workday after the holiday, Huawei announced that its native Hongmeng operating system’s Xiaoyi assistant app had integrated Deepseek

Advertisements

Just three days later, Honor's YOYO intelligent assistant store launched a trial version of Deepseek-R1. On the same day, OPPO disclosed that their upcoming foldable flagship, the OPPO Find N5, would incorporate Deepseek-R1. Within a week, a slew of other smartphone makers like Nubia, Lenovo’s Moto, and Xiaomi followed suit, announcing their integration plans.

Automakers, always eager to inject themselves into technological trends, soon joined the initiative alongside smartphone companiesOn February 6, Geely announced that its self-developed Star Vision model had achieved a deep integration with Deepseek-R1. Following this, several other car manufacturers—including Zeekr, Baojun, Lantu, ZhiMi, Leapmotor, Great Wall, GAC, Changan, and Chery—quickly confirmed that “Deepseek is now in our vehicles.”

In just two weeks, Deepseek had established itself as an “accelerator” for intelligent transformation, boasting low-cost inference capabilities and on-device deployment advantages, overwhelming demand from well-known Chinese enterprisesHowever, it wasn’t until the appearance of a line stating “AI Search” below the WeChat search box that many users realized the implications of this “AI revolution.”

On February 15, users discovered that WeChat Search had launched the AI Search feature, with Tencent responding that WeChat’s search was employing the Deepseek model during its gray testing phaseSoon after, Baidu announced its search platform was also engaged in such testing and officially launched AI Search at midnight on February 18.

When compared to the AI frenzy ignited by ChatGPT in November 2022, the current excitement surrounding “Deepseek integration” evokes a feeling of AI arriving en masse, ready to integrate into daily lifeThe incorporation of Deepseek into essential internet tools signifies a shift where AI is no longer a high-tech marvel but rather an accessible feature that users can deploy effortlessly.

Commercial motives drive every business action, and big firms are eager to deliver tangible experiences of how AI transforms lives

Advertisements

By participating in the Deepseek trend, these corporations aim to achieve their strategic goals rapidly.

The elevated status of Deepseek largely stems from its unique advantages in cost and functionalityFollowing the emergence of ChatGPT, people marveled at AI's capabilities but also hesitated due to high expensesIn contrast, Deepseek leaped ahead by offering strong performance at a fraction of the cost.

Research from Guotai Junan Securities indicates that training the full version of Deepseek V3 (with 671 billion parameters) costs about $5.576 million—only 7% of the cost of Meta’s Llama 3. In comparison, OpenAI's ChatGPT-4o could cost $78 million or more, requiring thousands of high-performance chipsNotably, during training, Deepseek used a performance-restricted custom AI chip, the Nvidia H800.

By using less than 5% of the resources required for GPT models, Deepseek-R1 achieves performance comparable to OpenAI's top inference modelsThis “low-cost AI model paradigm" not only reduces operational costs but also makes AI applications more feasible, especially in real-time interactive scenarios where inference power is essential to enhance user experience.

For instance, reports from PConline’s travel column highlight how Geely channels Deepseek-R1’s cognitive capabilities into its Star Vision vehicle control function, resulting in a 40% boost in interactive response speed and a 98% accuracy rate in intent recognition while costing merely a third of traditional solutions.

Deepseek's diverse range of parameter versions, from 1.5 billion to 70 billion, also enables flexibility for various hardware capacitiesThis means that devices like cars, personal computers, and smartphones can feasibly implement local deployments of Deepseek, fostering a more customized and sizable AI experience at the device level.

Moreover, the open-source nature of Deepseek dramatically lowers the barriers to commercializing AI solutions

Advertisements

In industries like management, finance, and healthcare, businesses can leverage Deepseek’s robust logical reasoning capabilities to train AI agents tailored to their specific sectors.

Larger companies can also utilize open-source models for private deployments, addressing data privacy concerns while meeting custom needs and filling technological gapsFor instance, by incorporating Deepseek, WeChat and Baidu have enhanced existing search tools with deep analysis capability, elevating their product value without significantly increasing costs.

The proliferation of on-device AI fostered by Deepseek could spur notable innovations in the hardware marketGiven the expansion of model parameter requirements, future devices necessitating localized deployment will require enhanced memory capacityCurrently, smartphones equipped with at least 16 GB of DRAM are considered the minimum for AI functionality, as 16 GB also serves as the entry point for Microsoft's AIPC memory offerings.

The trend led by Deepseek showcases how lower training costs can diminish commercial barriers, while the ensuing wave of big companies engaging with the technology can incite widespread AI adoptionThis chain reaction contributes to infrastructure enhancements across sectors and creates new markets, products, and growth avenues, akin to Prometheus from Greek mythology bringing the fire of knowledge to humanity.

The race to integrate Deepseek reflects not only recognition of its potential to revolutionize industry dynamics but also a strategic move to establish an ecosystemFor instance, within the cloud services market, Deepseek’s arrival has disrupted the status quo, paving the way for competitive pricingThe quicker companies can deploy Deepseek, the more they can cement their market foothold and attract new clients.

In response, Tencent Cloud has launched developer incentives, providing one-click deployment for all Deepseek models along with free trials

Similarly, Baidu has offered incredibly low inference rates for the models, as much as 70-80% off the official priceVolcano Engine also provided a limited-time 50% discount for firms eager to integrate Deepseek-R1 fully.

Conversely, smartphone and automobile manufacturers benefit from Deepseek’s capabilities by upgrading functions while filling competency gaps, thus enabling differentiated competitionNoteworthy advancements like "voice activation + localized AI" significantly enhance user interaction and become unique selling points for existing products.

For example, Geely capitalizes on Deepseek-R1's capabilities to develop a synergy between the dual cognitive systems, precisely interpreting vague commands, hence seamlessly adjusting the air conditioning based on passive cues like "a bit stuffy," promoting an experience of intuitive care over simple interaction.

Within the content realm, firms can apply AI to interconnect their ecosystems and generate substantial valueFor example, Zhihu's AI search product has integrated the full capability of Deepseek-R1, ensuring a more accurate reflection of user intent, while enhancing data utility and knowledge management overall.

Furthermore, in content creation, Yuanwen Group has utilized the independently executed Deepseek-R1 model to aid writers in generating online literature with greater efficiency.

For large organizations with rich operational ecosystems, merely adding AI search functionality may be enough to leverage ecological resources, enhancing collaborative efficiency and fostering a competitive edgeTake, for instance, Tencent’s Yuanbao, which melds Deepseek with various information sources from its expansive ecosystem to deliver real-time, comprehensive answers, thereby redefining its own ecological value.

On a developmental level, the cost-effective, efficient performance of Deepseek has birthed a wave of integration, facilitating equal access to AI, ultimately fostering value reconstruction across various sectors

Advertisements

Advertisements