AI Index 2.0 – Evolving With the Wave
We launched the AI Index fifteen months ago to help founders and investors understand how markets value, and where they expect value to be created, in the AI Wave, just as the Bessemer index did with the Cloud. We had early conviction that AI would be the largest value creation wave in history, and it felt critical to build a living benchmark to reflect that.
But we also knew something from the beginning: this index would need to evolve.
At launch, our criteria were relatively clear. You could easily tell which companies were prioritizing AI. Some dedicated half their earnings calls to it (like ServiceNow), while others barely mentioned it (like DataDog). But by 2025, that line is blurry. Almost every software company now “does and prioritizes AI.” Our old methodology no longer separates signal from noise.
Thanks for reading Forward Feed! Subscribe for free to receive new posts and support my work.
Why change now?
The AI space is moving at unprecedented speed, AI adoption among U.S. businesses stood at 5% in early 2023, now is closer to 43% (Ramp). Just one year ago, scaling laws and model size drove the majority of the progress. Now, the focus is on reasoning, post-training and test-time computation.
Introducing AI Index 2.0
Given these advances, we also recognized the need to update our Index. Version 2.0 offers a more curated and opinionated view. It includes the top 20 public companies we believe most impact AI progress and benefit most from its adoption. We will revisit the index on an annual basis and add newly public companies to the Index after IPOs. The index selects companies based on seven categories:
Core model development: Companies that build and train their own foundation models.
Examples include Meta (Llama), Adobe (Firefly), Alibaba (Qwen), and Google (Gemini). If they were to go public, Anthropic and OpenAI would be in this category.
Ecosystem influence: Companies whose APIs, SDKs, or open source tools are broadly adopted by AI builders.
Examples: Cloudflare, MongoDB.
Critical infrastructure: Providers of compute, memory, or networking hardware essential for AI training and inference.
Examples: AMD, NVIDIA, TSMC, CoreWeave, Broadcom. If they were to go public, Cerebras would be in this category.
Data infrastructure for AI: Companies that operate platforms where enterprise data is stored, processed, and activated for AI workloads.
Example: Snowflake. If they were to go public, Databricks would be in this category.Revenue impact: Companies that are already seeing material revenue from AI products or have a credible path to AI monetization at scale.
Examples: Palantir, Duolingo, HubSpot. If they were to go public, Perplexity would be in this category.
Strategic investment: Companies allocating significant R&D budgets, CapEx, or pursuing acquisitions focused on AI expansion.
Example: Microsoft, Google, Meta.
Talent and research density: Companies with a high concentration of top researchers or open contributions to the field.
Examples: Google, Meta.
This update reflects the reality that AI is advancing quickly. As the market matures, our goal remains the same: to provide a clear lens on where value is being created, and provide insights to early stage founders and investors.
For those who want to dive deeper on the methodology and the full list companies selected, you can visit this Link
