How AI‑Native Companies Shape the Global Large‑Model Ecosystem

With large‑model‑centric AI driving a new wave worldwide, a cohort of AI‑native companies has emerged, defined as firms that embed artificial intelligence as the core product, service, or capability from inception, differentiating them from businesses that merely add AI as a tool.

Tencent Research Institute surveyed over 100 global AI‑native firms, focusing on three key questions: 1) What new technologies, product forms, and business models are they spawning? 2) How does AI deeply integrate into industry processes to enable upgrades? 3) How is the AI ecosystem evolving structurally?

1. Global AI‑Native Enterprise Ecosystem Overview

(1) Three Foundational Model Ecosystems

The generative AI field now clusters around three major ecosystems led by OpenAI, Anthropic, and Google, each offering distinct innovation soils for AI‑native startups.

OpenAI ecosystem : Largest scale and strongest attraction, with 81 startups valued at ~US$63.5 billion, covering AI search, content generation, legal services, embodied intelligence, and more.

Anthropic ecosystem : Mid‑size, targeting enterprise‑level markets and safety, hosting 32 firms (~US$50.1 billion) focused on high‑trust domains such as knowledge management, automated decision‑making, finance, law, and healthcare.

Google ecosystem : Smallest but fastest‑growing, with 18 firms (~US$12.8 billion) leveraging Gemini models for creative generation, marketing optimization, data analysis, and other niche tracks.

(2) Multi‑Model Access Strategies

To mitigate competition and reduce reliance on a single ecosystem, many AI‑native firms adopt multi‑model integration, supporting both OpenAI and Anthropic APIs (e.g., Anysphere, Clay, Glean, Hebbia, Jasper). This approach often aligns with a B2B2B model, delivering AI capabilities to service‑oriented clients who further embed them downstream.

(3) Self‑Developed Model Focus

Beyond the three ecosystems, a growing segment builds proprietary models. Some unicorns (e.g., xAI, Cohere, Mistral, Zhihu) aim for breakthrough general‑purpose models, while others target vertical niches such as content generation (Midjourney, Stability, Suno) or embodied intelligence (Physical Intelligence, World Labs).

2. Ecosystem Layout Differentiation: The Three Players' Tactics

(1) Core Positioning

OpenAI pursues a “general AI tool platform” with extensive plugins and API openness, positioning itself as a “super entry point.”

Anthropic emphasizes safety and enterprise‑grade reliability, branding itself as a “secure, enterprise‑focused AI service.”

Google embeds Gemini across its product suite, creating a full‑stack, tightly integrated AI stack.

(2) Developer Strategy Comparison

OpenAI : Offers APIs, SDKs, function calling, fine‑tuning, and a revenue‑share model, plus ChatGPT Plus and Enterprise tiers.

Anthropic : Provides the MCP protocol for Claude integration, long context windows, and B2B‑oriented API partnerships, with Pro and Team subscriptions.

Google : Delivers Gemini via Google Cloud, Agents service, and A2A protocol, integrating with IDEs (e.g., Replit) and other developer tools.

(3) Channel Strategy Comparison

OpenAI : Leverages Microsoft Azure for enterprise distribution while maintaining direct consumer channels through ChatGPT web and apps.

Anthropic : Partners with cloud providers (AWS Bedrock, Google Cloud Vertex AI) and embeds Claude in third‑party tools like Slack.

Google : Pre‑installs Gemini in Gmail, Docs, Android, Chrome, and bundles it with Google Workspace and Cloud services.

(4) Vertical Industry Penetration

OpenAI : Broad coverage across medical, legal, finance, creative, and education via partner startups and Microsoft integrations.

Anthropic : Concentrates on high‑sensitivity sectors such as finance, law, security, and coding, leveraging Claude’s safety.

Google : Deploys specialized models (e.g., Med‑PaLM, Gemini for coding) across healthcare, manufacturing, media, and education.

(5) Pricing Strategy Comparison

OpenAI : API usage‑based billing with subscription tiers (ChatGPT Plus, Enterprise), gradually lowering prices.

Anthropic : Token‑based pricing with longer context windows, offering Pro/Team plans and occasional cloud‑partner discounts.

Google : Low‑price bundled offerings, leveraging TPU cost advantages and cross‑product subsidies.

Conclusion

While OpenAI, Anthropic, and Google each pursue distinct ecosystem strategies, the overall AI‑native landscape remains in early formation, with rapid technological shifts and evolving product experiences. The firms that sustain open collaboration, continuous innovation, and multi‑dimensional value delivery are likely to secure a leading position in the competitive large‑model ecosystem.