The shifting landscape of internet information retrieval has reached a new milestone as OpenAI’s ChatGPT begins to exert significant pressure on the traditional search engine dominance held by Google. Adam Fry, the consumer product lead at OpenAI, recently disclosed that ChatGPT now handles approximately one-fifth of the search-like query volume managed by Google. This revelation, shared during a private industry dinner hosted by OpenAI in New York, provides a rare quantitative glimpse into the internal metrics of the artificial intelligence giant. The disclosure underscores a rapid evolution in how users interact with the web, transitioning from keyword-based indexing to conversational, intent-driven AI agents.
The data, later summarized in a report by Big Technology, highlights a pivotal shift in consumer behavior. OpenAI’s internal metrics define a "search" specifically as any instance where a user’s prompt necessitates that ChatGPT reach out to the live internet to retrieve information. This methodology differs significantly from traditional search engine metrics, which often count every individual character string typed into a search bar, regardless of whether it triggers a web crawl or a direct database response.
Chronology of the AI Search Evolution
The rise of AI-driven search has been meteoric. Historically, ChatGPT was utilized primarily as a tool for creative writing, software coding, and summarization—tasks that did not require real-time web access. However, over the past twelve months, the utility of the platform has expanded. OpenAI has observed a distinct shift toward questions regarding current events, financial markets, healthcare, and biographical information. These topics inherently require up-to-date verification, prompting the integration of search tools into the core ChatGPT experience.
This trajectory is supported by broader usage statistics. Reports indicate that weekly active users for ChatGPT surged by 20% in the last two months alone, climbing from 1 billion to 1.2 billion. As the user base grows, the frequency of "search-like" prompts has increased in lockstep. This growth trajectory poses a fundamental question for market analysts: to what extent is AI cannibalizing traditional search traffic, and to what extent is it generating entirely new categories of information retrieval?
The Methodological Divide: Defining a Search
The primary challenge in comparing OpenAI to Google lies in the definition of a search query. Fry’s assertion—that "every time you ask a question and go search the web, that’s a search"—is a functional definition focused on the backend trigger of a web crawl. Conversely, Google’s search engine operates on a vastly different scale, managing trillions of queries annually.
In February 2025, search analytics firm Ahrefs offered an independent perspective. Utilizing a study conducted by Harvard University and OpenAI, Ahrefs noted that ChatGPT processed approximately 2.5 billion daily prompts as of July 2025. By analyzing the linguistic intent behind these prompts, researchers determined that roughly 65% of those interactions were queries that users would have historically funneled into a traditional search engine like Google. When reconciled with Google’s reported annual volume of over 5 trillion searches, Ahrefs estimated that ChatGPT’s search footprint was closer to 12% of Google’s total capacity.
The discrepancy between the 12% figure and Fry’s 20% claim can be attributed to the divergence in classification. While Ahrefs attempts to categorize user intent, OpenAI is tracking technical web-access events. As AI models become more sophisticated, the line between a "prompt" and a "search" continues to blur, making the comparison increasingly complex for industry analysts.
Referral Traffic: The Real-World Impact
While OpenAI’s search volume metrics are compelling, a secondary metric—referral traffic—tells a different story. Referral traffic tracks the actual flow of users from a search platform to third-party websites. In this arena, the dominance of Google remains nearly absolute.
According to data tracked in October across 118,671 websites by Ahrefs’ free Web Analytics tool, Google accounted for 24.74% of tracked referral traffic. In stark contrast, ChatGPT contributed only 0.36%. This wide gap suggests that while users are increasingly turning to ChatGPT to answer direct questions, they are not yet using the platform as a primary gateway to navigate to other websites.
This behavior is consistent with the "answer engine" model of AI, where the objective is to provide a complete response within the interface, thereby reducing the need for the user to navigate away to external sources. For publishers and content creators, this represents a significant structural change. If the platform successfully answers the query without the user ever clicking a link, the traditional model of web traffic monetization faces a potential existential threat.
Google’s Defensive Posture and Growth
Despite the rise of ChatGPT, Alphabet’s Google has not seen a decline in query volume. In fact, Google’s third-quarter 2025 earnings report provided a counter-narrative to the idea that AI is eroding the search giant’s market share. CEO Sundar Pichai stated that investments in AI—specifically those integrated into the search experience, such as "AI Overviews"—have actually accelerated query growth.
Pichai noted that Google’s year-over-year query growth rate is currently at its highest level in the last five years. This suggests that the total "pie" of information retrieval is expanding rather than shrinking. The convenience of AI is encouraging users to perform more searches, for more complex tasks, at a higher frequency than previously possible. Google’s strategy appears to be one of "co-option," where it integrates AI features into its existing ecosystem to maintain user loyalty while simultaneously growing the total volume of traffic.
Implications for the Digital Ecosystem
The implications of these developments for the broader digital ecosystem are twofold. First, the definition of a "search" is undergoing a permanent transformation. The industry must move away from counting clicks or keyword inputs and toward measuring the successful resolution of a user’s intent. Whether that resolution happens via a list of blue links or an AI-generated summary is becoming secondary to the quality of the answer provided.
Second, the reliance on two distinct metrics—search requests and referral traffic—is essential for any realistic market analysis. If search requests represent the "demand" for information, then referral traffic represents the "supply chain" that supports the web. As OpenAI continues to refine its search capabilities, the disparity between these two numbers will serve as a key indicator of whether ChatGPT is becoming a destination in its own right or an intermediary that eventually directs traffic to the broader web.
Future Outlook and Analytical Limitations
Looking ahead, the industry remains in a period of intense volatility. OpenAI has yet to provide full transparency regarding the regional scope or the specific Google benchmarks used to calculate its 20% figure. Without access to the specific datasets, independent researchers find it difficult to reconcile these claims with third-party observations.
Furthermore, the lack of a standardized reporting metric for AI search volume makes it difficult for advertisers and publishers to plan their strategies. In the near term, those in the digital sector should treat these metrics as complementary rather than competitive. As Ahrefs continues to provide monthly reports on referral traffic, the data will offer a clearer view of how AI adoption is affecting the distribution of web traffic.
Ultimately, the rise of ChatGPT signifies that the "search engine" is no longer a monolithic concept. We are moving toward a dual-mode information economy: one where users go to traditional engines for navigation and deep-link discovery, and another where they go to AI models for synthesis, analysis, and rapid decision-making. As these two modes continue to interact, the only certainty is that the volume of human-computer interaction will continue to grow, forcing both incumbents and challengers to rethink the fundamental architecture of the internet.
While OpenAI’s executive disclosures provide a snapshot of the current state of play, the true measure of success will be found in the coming fiscal years. If ChatGPT’s referral traffic begins to climb alongside its query volume, the disruption to the current search paradigm will be undeniable. Until then, the industry will continue to navigate the friction between traditional search engines that thrive on traffic volume and AI models that thrive on information synthesis. The evolving relationship between these two systems will define the next decade of digital interaction, fundamentally changing how humanity accesses and processes the sum of human knowledge.
