The emergence of DeepSeek-R1, an open-source AI model that promises both low cost and high performance, has sent shockwaves through Silicon Valley and Wall StreetInvestors across the globe are eagerly tapping into the software sector, betting heavily on the remarkable potential of software companies to generate unprecedented earnings as the AI revolution continues to unfoldThe anticipation is that those firms demonstrating rapid revenue growth might witness share price trajectories reminiscent of Nvidia's meteoric rise, which spiked by an astonishing 1000% since October 2022.

As we glance toward the future of AI applications, it's evident that the groundbreaking "extremely low-cost AI training and inference" paradigm championed by DeepSeek could considerably streamline the operational demands associated with deploying large AI modelsThis leap could unleash a wave of generative AI applications across various sectors, including both B2B and B2C markets, thus catalyzing a surge in productivity levels for human societyIt is no surprise, therefore, that global financial resources are flowing towards software stocks, with a particular focus on Chinese software giants that exhibit significant valuation advantages.

DeepSeek-R1 has essentially declared a new chapter in the open-source big model arena by introducing a model that aligns with the principles of "extreme compression, highly efficient training, and simplified inference architecture." The tech reports released recently from companies like Meta, Palantir, SAP, and AppLovin depict an exceptionally robust outlook for AI-driven revenuesMajor global investment banks, including Goldman Sachs, Bank of America, and UBS, are increasingly convinced of the "long-term bullish narrative" surrounding software stocks, creating upward momentum in share prices for colossal software firms.

A key facet of the AI application software market lies in its transformative capacity to integrate advanced AI models into everyday business practices

Advertisements

These applications not only enhance corporate productivity but also improve individual efficiency in various tasks and learning processesEncompassing all software products that leverage AI, particularly generative AI technologies, these applications are designed to deliver insights through natural language processing, predictive analysis, and automation of repetitive tasks, ultimately providing tangible benefits to both businesses and individuals.

The contemporary trajectory of AI application software appears to be honed in on "generative AI applications" such as those introduced by DeepSeek, ChatGPT, Sora, and Claude from AnthropicA notable shift is emerging within the industry, where the functionality of AI is evolving from mere text-based query-response systems to "AI agents" capable of executing complex tasks autonomouslyExamples include the Agentforce 2.0 by Saifle and the embedded AI agents within Microsoft Dynamics 365. These advancements mark significant milestones in AI technology and signify the forthcoming lines of evolution in the field.

With the support of DeepSeek, a global wave of AI applications is about to gain momentum as software giants enter an era of unparalleled opportunities.

DeepSeek's unique combination of low-cost infrastructure, super-efficient training, and performance akin to OpenAI’s models, stems from rigorous engineering practices and fine-tuning processes throughout the training stagesThis model prioritizes pure reinforcement learning over traditional supervised learning methods that rely heavily on extensive datasets, significantly lowering costs associated with model training and inferenceBy employing advanced strategies such as multi-layer attention mechanisms particularly targeting the Query endpoints for low-rank approximation, DeepSeek optimizes hardware resource utilization during training, thus minimizing performance wastageCoupled with innovations like FP8 mixed-precision training and dual-pipe parallel communication, the model achieves an extraordinary balance of performance and cost-efficiency.

It is remarkable that for an investment of less than $6 million, and under conditions utilizing performance subsiding well below chips like H100 or Blackwell, DeepSeek's team has managed to develop an open-source AI model that rivals OpenAI's offerings

Advertisements

By contrast, competitors like Anthropic and OpenAI spent upwards of $1 billion on training their modelsThe pricing strategy for DeepSeek’s token outputs is also aggressively competitive; while OpenAI’s GPT-4 demands as much as $60 per million tokens, DeepSeek is positioned at merely $2.19.

The global dissemination of DeepSeek's low-cost computing paradigm is not only driving down costs associated with AI training and application but also accelerating the integration of AI software—specifically generative AI and AI agents—across numerous industriesThis shift promises to revolutionize business operations significantly while boosting revenues considerablyMajor players like Meta, SAP, Palantir, Salesforce, AppLovin, as well as Chinese tech giants like Alibaba and Tencent, could see exponential increases in AI revenues, emerging as the leading beneficiaries of the remarkable rise of DeepSeek.

In an illustrative example of the industry's optimism, Bank of America has crowned AppLovin, a leader in generative AI for advertising, as a "preferred stock" for the year 2025. Their highly optimistic forecast sets a twelve-month target price for AppLovin at $580, reflecting substantial confidence in their AI-driven software growth modelsComparatively, AppLovin's shares recently closed at approximately $471.67, a 24% increase, indicating robust investor interest.

The management at SAP has similarly acknowledged the increasing demand for innovative AI solutions that can autonomously complete tasks, reinforcing the need for AI agents in their operating modelThere is a definitive push from businesses to enhance efficiency and reduce operational costs, which has led to a surge in the application of AI agents capable of automating mundane tasks and performing intricate data analyses with real-time insights, thus enhancing overall business efficiency.

The urgency for AI integration into business practices has propelled the rapid adoption of AI agents

Advertisements

Advertisements

Advertisements