Be a part of our each day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Study Extra
New information reveals dramatic AI market share shifts in 2025, with fast modifications in how companies and customers make the most of synthetic intelligence instruments. Poe, a platform that hosts greater than 100 AI fashions, has launched a complete report that gives an unprecedented look into real-world utilization patterns throughout textual content, picture and video technology applied sciences.
Poe’s evaluation, primarily based on interactions from tens of millions of customers over the previous yr, presents technical decision-makers essential insights right into a aggressive discipline the place utilization information is usually intently guarded. “As AI models continue to progress, we believe they will become central to how people acquire knowledge, tackle complex tasks and manage everyday work,” the corporate writes.
The findings spotlight important market fragmentation throughout all AI modalities. Whereas established gamers like OpenAI and Anthropic preserve dominant positions in textual content technology, newer entrants equivalent to DeepSeek (in textual content) and Black Forest Labs (in picture technology) have rapidly captured significant market share, suggesting a dynamic ecosystem regardless of large investments flowing towards {industry} leaders.
Listed here are the 5 most shocking takeaways from Poe’s evaluation of the early 2025 AI ecosystem.
1. Google exhibits uneven efficiency throughout AI modalities
Google’s assorted efficiency throughout totally different AI modalities reveals the challenges of attaining cross-modal management. Its Gemini household of textual content fashions “saw growing message share through October 2024,” however has been “declining since” regardless of substantial funding and technical capabilities.
This contrasts sharply with Google’s efficiency in different classes. In picture technology, Google’s Imagen3 household has secured a powerful 30% market share, whereas in video technology, its Veo-2 mannequin has quickly captured 40% of messages.
This blended efficiency means that technical excellence alone doesn’t assure market management. For enterprise decision-makers, this underscores the significance of evaluating AI capabilities on a modality-by-modality foundation moderately than assuming management in a single space interprets to excellence throughout all AI capabilities.
2. Video technology experiences high-velocity competitors
Video technology, the latest frontier in generative AI, has already witnessed intense competitors and quickly shifting management positions. In accordance with the report, “The video generation category, while only existing starting in late 2024, has rapidly expanded to more than eight providers now offering diverse options to subscribers.”

Runway, an early pioneer, “has maintained a strong position with 30 to 50% of video gen messages” regardless of having solely a single API mannequin. Nonetheless, Google’s entrance has instantly disrupted the established order: “Google’s Veo-2, since its recent launch on Poe, rapidly captured nearly 40% of total video gen messages in just a few weeks.”
Chinese language-developed fashions collectively account for roughly 15% of video technology messages. Fashions like “Kling-Pro-v1.5, Hailuo-AI, HunyuanVideo and Wan-2.1 continue to push the frontier on capabilities, inference time and cost,” demonstrating that worldwide competitors stays a major think about driving innovation regardless of geopolitical tensions.
3. Picture technology undergoes radical transformation
The picture technology discipline demonstrates maybe probably the most dramatic market shift in gen AI, with established gamers quickly dropping floor to newcomers. “First-mover image gen models like Dall-E-3 and various Stable Diffusion versions were pioneers in the space, but have seen their relative usage share drop nearly 80% as the number of official image gen models has grown from 3 to ~25,” the report states.
Black Forest Labs emerged because the shock chief: “Black Forest Labs’s Flux family of image generation models burst onto the scene in mid 2024 and has maintained its dominant position as the clear frontrunner since, capturing close to 40% of messages.” This represents a outstanding achievement for a relative newcomer in opposition to established rivals with huge sources.

Google’s strategic funding in picture technology can also be bearing fruit, with “Google’s Imagen3 family…on a steady growth since its late 2024 launch, carving out almost 30% usage share.” This positions Google as a robust second-place contender regardless of its later market entry.
Poe’s information reveals a regarding pattern for AI corporations investing closely in sustaining older fashions: “As frontier labs release more capable models, usage of the new flagship model in a provider’s offering quickly cannibalizes the older versions.”
This sample manifests throughout corporations, with customers quickly abandoning GPT-4 for GPT-4o and Claude-3 for Claude 3.5. The implication is obvious: Sustaining backward compatibility and assist for legacy fashions might have diminishing returns as customers persistently migrate to the latest choices.
Corporations might must rethink their product lifecycle methods, doubtlessly focusing sources on fewer fashions with extra frequent updates moderately than sustaining in depth households of choices with various capabilities and worth factors.
5. Textual content AI duopoly faces new challengers
OpenAI and Anthropic preserve dominance in textual content technology, however face rising stress from newer entrants. In accordance with Poe’s information, “text usage across OpenAI and Anthropic models has been nearly equal, showcasing growing competition in the highly expressive text modality” since Claude 3.5 Sonnet’s launch in June 2024. Collectively, these two corporations command roughly 85% of textual content interactions on the platform.
Anthropic’s fast ascension to parity with OpenAI means that high quality and functionality enhancements can rapidly translate to market share shifts, even in a discipline with robust community results and first-mover benefits.
Extra intriguing is DeepSeek’s emergence as a professional third contender. The report notes that “DeepSeek-R1 and V3 went from no usage in December 2024 to gain 7% of messages at their peak, a significantly higher level than any previous open-source model family, such as Llama and Mistral.” This dramatic rise signifies that obstacles to entry for brand new textual content AI suppliers could also be decrease than beforehand anticipated.
The highway forward: A fluid panorama for AI decision-makers
The Poe report illuminates a gen AI market characterised by fast evolution, the place technical excellence alone doesn’t assure sustained market management. Enterprise decision-makers face an more and more advanced vendor panorama, the place right this moment’s dominant participant is perhaps tomorrow’s also-ran.
What stays clear is that consumer preferences can shift dramatically with new mannequin releases, suggesting organizations ought to construct versatile AI stacks that may adapt to altering capabilities moderately than locking into single-vendor options. The multimodal nature of AI adoption — with totally different leaders rising throughout textual content, picture and video — additional complicates enterprise technique.
As Poe notes in its conclusion: “We hope these findings offer a glimpse into the shifting dynamics of the AI model landscape.” For companies navigating this fluid ecosystem, the message is unmistakable: The AI revolution continues at breakneck velocity, rewarding those that preserve flexibility whereas punishing those that wager too closely on yesterday’s know-how leaders.