Winston Churchill once said, "The farther back you can look, the farther forward you are likely to see." In recent years, the race for supremacy in the AI domain has dramatically shifted the technological landscape, prompting experts to reconsider past estimates of potential developmentsTwo years ago, Baidu claimed that the gap between its AI model, Wenxin Yiyan, and ChatGPT was merely a month or twoDissenting voices suggested that the divide could stretch up to a decadeHowever, the emergence of DeepSeek has thrown these projections into disarrayFindings show that Chinese AI models are rivaling their international counterparts in terms of performance, often at a more compelling cost.

Looking ahead, the next two years promise a transformation of the playing field, as AI giants rapidly iterate through technological innovations and applicationsThe paradigm of artificial intelligence is evolving into a new arena where competition is defined not only by capability but also by accessibility and cost-effectiveness.

On February 6, OpenAI made a significant move by announcing that its ChatGPT search feature would be available to anyone without requiring an account

Advertisements

This decision signals OpenAI’s aggressive strategy to challenge Google’s dominance in the search engine domainThe previous year had already seen the introduction of ChatGPT’s free search functionality, and this latest development indicates a bold step towards further encroaching upon Google’s long-held market position.

The news quickly rose to trending status on Weibo, showcasing significant public interestIn response, Google unveiled its Gemini 2.0 model series on the same day, which includes various versions such as Flash, Pro, and Flash-Lite, aimed at enhancing AI capabilities while providing budget-friendly solutions for users.

A year ago, ChatGPT stood alone as a powerhouse in generative AI, but the landscape has since been transformedThe rise of DeepSeek presents challenges not only for OpenAI and Anthropic, representatives of closed-source AI, but also raises the stakes for Google and Meta, which both operate within the open-source ecosystem

Advertisements

As competition heats up, a new chapter in AI development is underway.

The question arises: Has OpenAI entered the search arena too late? Historically, Google's search engine revolutionized entire industriesNow, generative AI holds the potential for a fresh wave of transformation, with leaders like OpenAI beginning to carve their niche within the search engine realmThis latest update simplifies the user experience considerably: prospective users can access ChatGPT’s search feature without the cumbersome requirement of an account, effectively lowering barriers to entry and seeking to amplify user growth.

Since its rise to prominence, insiders have consistently highlighted the significance of search engines as a pivotal application for AINo longer are users confined to sifting through countless web pages; instead, they receive highly consolidated answersTraditional searching hinged on keyword matching

Advertisements

In contrast, AI-powered systems utilize semantic understanding to directly generate summaries, enhancing efficiency drastically.

According to a recent report by Semrush, Google currently boasts an impressive 6.5 billion unique visitors as of December 2024, while ChatGPT sits at 566 millionAlthough the existing user base for AI products falls short of that of traditional search engines like Google and Baidu, AI is reshaping users' search interactionsAs the barriers to accessing AI applications diminish, the user base for ChatGPT and similar offerings could very well experience a steady increase.

Some commentators posit that as AI products continue their ascent, search engines led by Google are likely to encounter a significant impactNot only ChatGPT but also others like Doubao, DeepSeek, and Kimi are rapidly incorporating free search capabilities, establishing themselves as crucial channels for information aggregation

• India Cuts Rates by 25 Basis Points
• How to Navigate the AI Investment Boom?
• Amazon Doubles Revenue and Profit in Q4
• Global Public Debt on the Rise
• The Illusion of Ever-Rising UK Stock Market

Unlike the previous era of superficial content integration, modern applications typically cite source links, facilitating users’ ability to verify and confirm the accuracy of information.

The search arena is undeniably crowded, yet Google’s stronghold is not easily overcomeAn experienced AI professional commented, "ChatGPT has opened up too late; its strategy should have focused on quicker solutions to API requests to capitalize on incoming trafficGoogle's solutions have already undergone adjustments; their advantage lies not only in their search algorithms but equally in the vast reserves of network data snapshots they have accumulatedThey are well aware of users' search intentions." As numerous search products in the Chinese market transition to fully free models, it seems OpenAI may have missed a critical window of opportunity.

The expert added, "OpenAI is too constrained in its approach; ChatGPT's positioning is limited to chat functionalities—it should diversify into more application fields

As competitors leverage open-source platforms to claw into these markets, OpenAI’s narrow moat is insufficient for maintaining its lead."

On the flip side, Google too is hastening its AI search developmentsIts Search Generative Experience (SGE) has been revamped into AI Overviews, integrating Gemini into traditional search functionalities to provide AI-generated answers directly on search pages, establishing direct competition with ChatGPT search.

The recent Gemini series upgrades also significantly lowered pricing, emphasizing a commitment to attracting usersYet, ChatGPT’s search function faces obstacles regarding the transparency of information sourcesSome AI-generated content remains difficult to trace back to its original data, leaving potential issues of inaccuracies—often referred to as 'hallucinations'—to be contended withMeanwhile, Google maintains a broader coverage when it comes to search results, holding a distinct advantage in real-time and news content delivery.

This highlights an intriguing distinction: Google is building upon its existing search capabilities by integrating large models to consolidate answers, whereas ChatGPT and similar platforms are reconstructing search capability around AI technology

The rivalry marks a fresh transformation in how users engage with search platforms.

The rapid rise of DeepSeek has prompted almost simultaneous updates from OpenAI and Google, signaling the impact of this emerging competitorLaunched globally on January 11, DeepSeek quickly distinguished itself amid other AI products, boasting a steep growth trajectorySensor Tower's data reveals that in its first 18 days, DeepSeek garnered 16 million downloads—almost double that of ChatGPT’s comparable launch period.

Statistics from AI product rankings indicate that by January 31, DeepSeek achieved a remarkable 22.15 million daily active users in just 21 days post-launch, representing 41.6% of ChatGPT’s daily user count and exceeding Doubao’s 16.95 millionThis positions DeepSeek as the fastest-growing AI application worldwide, even if its total user count still trails that of ChatGPT.

The urgency created by DeepSeek’s swift ascent is palpable, compelling major AI manufacturers to take note

Counterpoint’s lead analyst Wei Sun remarked, "A noticeable trend is emerging—the gap between open-source AI and closed-source AI is narrowingDeepSeek and Meta's open-source AI create substantial competition for proprietary models like OpenAI’s." The performance prowess of open-source large language models marks a transformative shift within the AI sector.

In recent weeks, OpenAI has been extremely active, not only opening up search capabilities for all users but also prioritizing cost reductionLast week, OpenAI swiftly debuted its new reasoning model, o3-mini, granting access across ChatGPT and its API.

Simultaneously, the new model has seen significant price drops, with o3-mini costing $1.10 per million tokens for input and $4.40 for output, representing a 63% decrease compared to o1-mini, yet still appearing relatively high compared to DeepSeek's more competitive pricing strategy

DeepSeek’s R1 offers API pricing at $0.14 per million tokens (cache hit) and $0.55 per million (cache miss), with output pricing at $2.19 per million tokens.

The competitive landscape has shifted dramatically as the price war intensifies, with Google also joining the frayIn its latest model releases, the Gemini 2.0 Flash-Lite stands at $0.30 per million tokens, making it the most economically viable option from GoogleThe pricing for Gemini 2.0 Flash has also been adjusted to $0.40 per million tokens for output, halving costs compared to its predecessor, Gemini 1.5 FlashThe Gemini 2.0 Flash Thinking experimental version is now available free of charge and offers functionalities to access, summarize, and distill content from YouTube videos.

DeepSeek's rapid rise signifies not merely a technological achievement but a pivotal moment reshaping competitive dynamics within the AI landscape