---
title: "Strategic Decision or Structural Advantage?"
description: "Databricks Unity Gateway data shows Anthropic capturing 94% of SOTA tokens while OpenAI captures 81% of low-end tokens \u0026 stalls open-weight growth with GPT-6 Luna."
categories: ["AI","Strategy","Pricing"]
keywords: ["databricks ai index","unity gateway","anthropic","openai","gpt-6 luna","claude opus 5.5","claude sonnet 5.5","open-weight models"]
ai_summary: "Sixteen weeks of Databricks Unity Gateway telemetry across 10,000+ enterprises show Anthropic \u0026 OpenAI splitting the market by tier. Anthropic captures 94% of frontier SOTA tokens \u0026 70% of proprietary spend, while OpenAI captures 81% of low-end tokens \u0026 coincides with open-weight share stalling at 31% after cutting Luna prices by 90%."
date: 2026-11-16
lastmod: 2026-11-16
wordCount: 704
readingTime: 4 minutes
canonical_url: https://tomtunguz.com/two-pricing-strategies-in-enterprise-ai/
author: "Tomasz Tunguz"
---


OpenAI & Anthropic each hold roughly 90% market share of an AI segment, but those segments are different : Anthropic holds dominant share at the premium end (94%), & OpenAI at the low end (81%).[^1]

Across 10,000+ enterprises, that split raises a business question : is this a deliberate pricing strategy to capture different ends of the market, or a structural cost difference in how the two labs serve tokens?

{{< email_image src="https://res.cloudinary.com/dzawgnnlr/image/upload/v1794857256/analysis/databricks-ai-index/chart1-sota-share-d13283b0c7b0.png" alt="Anthropic captures 94.1% of frontier SOTA token volume on Databricks Unity Gateway" width="760" height="429" >}}

Anthropic sells a premium product & dominates the SOTA (state-of-the-art) tier. Anthropic's share of SOTA token volume rose from 81% in July to 94% on November 2, propelled by the September 21 rollout of Claude Opus 5.5.[^1]<sup>,</sup>[^2]<sup>,</sup>[^3]

{{< email_image src="https://res.cloudinary.com/dzawgnnlr/image/upload/v1794864333/analysis/databricks-ai-index/chart5-mid-tier-09d4c2bed824.png" alt="OpenAI climbed from 0% to 47% of mid-tier tokens" width="760" height="429" >}}

The mid tier is where the two labs now collide directly. Across all Sonnet & Sol versions (Sonnet 5 & 5.5 vs GPT-6 Sol & 6.1 Sol), OpenAI climbed from 0% in July to 47% of mid-tier volume by November 2, & took 57% of the newest Sonnet 5.5 vs Sol 6/6.1 cohort within six weeks of launch.[^1]<sup>,</sup>[^4]

{{< email_image src="https://res.cloudinary.com/dzawgnnlr/image/upload/v1794857258/analysis/databricks-ai-index/chart2-low-end-share-2a842af16b9a.png" alt="OpenAI captures 81.1% of low-end token volume after Luna price cuts" width="760" height="429" >}}

OpenAI pursued the opposite strategy, using aggressive price cuts on Luna to capture the high-volume floor. On July 30, OpenAI cut Luna pricing by 80% & then another 50% just a few months later, for a 90% total reduction.[^5]<sup>,</sup>[^6]

On September 22, OpenAI launched GPT-6 Luna at $0.10 per 1m tokens.[^6] Its low-end share surged from 30% to 66% after the first cut, peaked at 84%, & finished at 81% on November 2.[^1]<sup>,</sup>[^7]

{{< email_image src="https://res.cloudinary.com/dzawgnnlr/image/upload/v1794857260/analysis/databricks-ai-index/chart3-open-weight-3bc699a560aa.png" alt="Open-weight token share quadrupled to 33.2%, then stalled after GPT-6 price cuts" width="760" height="429" >}}

Open-weight token share climbed every week from 7% in July to 33% in September, then stalled at 31% in the same week GPT-6 Luna & Sol cut prices by 50%.[^1]<sup>,</sup>[^8]

That raises the strategic question in foundation models today : is Luna's pricing a short-term market-share subsidy, or a structural cost advantage in OpenAI's model architecture?

[^1]: Databricks, [Databricks AI Index](https://developers.databricks.com/ai-index) & [Introducing the Databricks AI Index : The state of enterprise AI tokenomics](https://www.databricks.com/blog/introducing-databricks-ai-index-state-enterprise-ai-tokenomics), published November 16, 2026 (weekly share series through November 2, 2026 ; coding benchmark updated November 10, 2026). Unity Gateway reflects Databricks' enterprise customer base (data-platform & agentic-coding workloads) rather than the entire enterprise AI market.
[^2]: SOTA-tier share sums Anthropic's frontier models (Opus 4.8, Opus 5, Opus 5.5 & Fable 5.1, reaching 39.8% of total proprietary tokens on November 2) against OpenAI's frontier models (GPT 6 Astra & GPT 5.6 Sol, 2.5% of total proprietary tokens on November 2).
[^3]: Blended token prices throughout this post assume a 200:1 input-to-output token ratio : `(200 x input + output) / 201` per 1m tokens.
[^4]: Mid tier compares Anthropic Sonnet 5.5 against OpenAI GPT-6 Sol & GPT-6.1 Sol. Databricks does not report Claude Haiku separately in its token-share series (it sits inside Older Anthropic Models), so budget-tier comparisons rest on list prices & benchmark cost per task rather than token share. Standard API list prices per 1m tokens as of November 2026 : OpenAI [API Pricing](https://developers.openai.com/api/docs/pricing), Anthropic Claude Sonnet 5.5 ($2.00 / $10.00 / $0.10 cached) & Claude Haiku 5.5 ($0.10 / $0.50 / $0.01 cached up to 100k prompt tokens ; $0.50 / $2.50 above 100k), Zhipu GLM 5.3 Flash ($0.15 / $0.50 / $0.03 cached), & DeepSeek V4.1 Flash ($0.15 / $0.60 off-peak ; $0.30 / $1.20 standard ; $0.003–$0.006 cached).
[^5]: OpenAI, [Advancing the price-performance frontier with GPT-5.6](https://openai.com/index/advancing-the-price-performance-frontier-with-gpt-5-6/), published July 30, 2026 : GPT-5.6 Luna cut 80% from $1.00 / $6.00 to $0.20 / $1.20 per 1m input & output tokens ; serving cost reduced 20% via GPU kernel improvements & token generation efficiency increased more than 15% via speculative decoding.
[^6]: OpenAI, [Introducing GPT-6 Sol & Luna](https://openai.com/index/introducing-gpt-6-sol-and-luna/), originally published September 22, 2026 (updated September 29, 2026) : GPT-6 Luna priced permanently at $0.10 / $0.50 per 1m input & output tokens ($0.01 cached input), 50% below GPT-5.6 promotional pricing.
[^7]: Low-end tier share compares OpenAI's economy models (GPT 6 Luna, GPT 5 Nano & Older OpenAI Models, 19.3% of total proprietary tokens on November 2) against Anthropic's economy bucket (Older Anthropic Models including Haiku, 4.5% of total proprietary tokens on November 2). Including mid-tier workhorses (GPT 6.1 Sol & GPT 6 Sol vs Sonnet 5.5 & Sonnet 5), OpenAI's non-SOTA share rose from 13.6% on July 20 to 61.4% on November 2.
[^8]: Databricks, [The Enterprise Case for Open-Weight Models on Databricks](https://www.databricks.com/blog/enterprise-case-open-weight-models-databricks), published November 4, 2026.
