State of AI: 2025 Mid-Year LLM Market Update - Foundation Model Landscape + Economics episode artwork

EPISODE · Aug 4, 2025 · 30 MIN

State of AI: 2025 Mid-Year LLM Market Update - Foundation Model Landscape + Economics

from State of AI · host Ali Mehedi

Dive deep into the rapidly evolving world of Large Language Models (LLMs) with our discussion on Menlo Ventures' '2025 Mid-Year LLM Market Update: Foundation Model Landscape + Economics' report. Released July 31, 2025, this essential update from Tim Tully, Joff Redfern, Deedy Das, and Derek Xiao reveals critical shifts and emerging trends shaping the future of computing.In this episode, we unpack the latest data from a survey of over 150 technical leaders across startups and enterprises, exploring:Anthropic's ascendancy as the new enterprise LLM leader, now commanding 32% of enterprise usage and surpassing OpenAI, which holds 25%. We'll discuss how innovations like Claude Sonnet 3.5, 3.7, and 4 propelled this growth.The significant shift in AI spend from model training to production inference, with LLM API spending more than doubling from $3.5 billion to $8.4 billion in just six months, marking a clear move towards models running in production.The emergence of code generation as AI’s first breakout use case, with Claude quickly becoming the developer’s top choice, capturing 42% market share—more than double OpenAI's.The breakthrough importance of Reinforcement Learning with Verifiers (RLVR) as the new path to scaling intelligence, pushing the envelope beyond traditional pre-training limits.Why 2025 is known as the 'year of agents,' as LLMs are increasingly trained to think step-by-step, reason through problems, and integrate external tools, dramatically boosting their real-world utility.The stagnation of open-source model adoption in the enterprise, which has seen market share slightly decline. We'll examine the persistent performance gap with closed-source frontier models, technical complexities, and enterprise reluctance to use certain foreign APIs as key contributing factors.The surprising market dynamic where enterprises prioritize and pay for performance over price, consistently upgrading to the best-performing models even as older ones become cheaper. Builders rapidly switch to newer, higher-performing models, with 66% upgrading within their existing provider.Join us to understand these pivotal market dynamics and gain insights into where long-term value might accrue in the LLM ecosystem."

Dive deep into the rapidly evolving world of Large Language Models (LLMs) with our discussion on Menlo Ventures' '2025 Mid-Year LLM Market Update: Foundation Model Landscape + Economics' report. Released July 31, 2025, this essential update from Tim Tully, Joff Redfern, Deedy Das, and Derek Xiao reveals critical shifts and emerging trends shaping the future of computing.In this episode, we unpack the latest data from a survey of over 150 technical leaders across startups and enterprises, exploring:Anthropic's ascendancy as the new enterprise LLM leader, now commanding 32% of enterprise usage and surpassing OpenAI, which holds 25%. We'll discuss how innovations like Claude Sonnet 3.5, 3.7, and 4 propelled this growth.The significant shift in AI spend from model training to production inference, with LLM API spending more than doubling from $3.5 billion to $8.4 billion in just six months, marking a clear move towards models running in production.The emergence of code generation as AI’s first breakout use case, with Claude quickly becoming the developer’s top choice, capturing 42% market share—more than double OpenAI's.The breakthrough importance of Reinforcement Learning with Verifiers (RLVR) as the new path to scaling intelligence, pushing the envelope beyond traditional pre-training limits.Why 2025 is known as the 'year of agents,' as LLMs are increasingly trained to think step-by-step, reason through problems, and integrate external tools, dramatically boosting their real-world utility.The stagnation of open-source model adoption in the enterprise, which has seen market share slightly decline. We'll examine the persistent performance gap with closed-source frontier models, technical complexities, and enterprise reluctance to use certain foreign APIs as key contributing factors.The surprising market dynamic where enterprises prioritize and pay for performance over price, consistently upgrading to the best-performing models even as older ones become cheaper. Builders rapidly switch to newer, higher-performing models, with 66% upgrading within their existing provider.Join us to understand these pivotal market dynamics and gain insights into where long-term value might accrue in the LLM ecosystem."

NOW PLAYING

State of AI: 2025 Mid-Year LLM Market Update - Foundation Model Landscape + Economics

0:00 30:12

No transcript for this episode yet

We transcribe on demand. Request one and we'll notify you when it's ready — usually under 10 minutes.

No similar episodes found.

No similar podcasts found.

Frequently Asked Questions

How long is this episode of State of AI?

This episode is 30 minutes long.

When was this State of AI episode published?

This episode was published on August 4, 2025.

What is this episode about?

Dive deep into the rapidly evolving world of Large Language Models (LLMs) with our discussion on Menlo Ventures' '2025 Mid-Year LLM Market Update: Foundation Model Landscape + Economics' report. Released July 31, 2025, this essential update from Tim...

Can I download this State of AI episode?

Yes, you can download this episode by clicking the download button on the episode player, or subscribe to the podcast in your preferred podcast app for automatic downloads.
URL copied to clipboard!