71.1% of companies active on Databricks’ Unity Gateway in the week of 2 November had used AI models from two or more providers, up from 36% in the week of 20 July, according to the Databricks AI Index, which Databricks published on 16 November. 47.9% had used three or more.
Databricks CEO Ali Ghodsi put it at 75% with two or more and 50% with three or more in an X post the same day. Our number: the two-or-more share roughly doubled in 15 weeks, and the share using four or more rose from 4.4% to 31.4%, by our arithmetic.
Business Pill · The model and the harness
An evergreen Business Engineer explainer on the model and the harness around it. It teaches the general idea only and does not discuss the Databricks AI Index.
The key insight: As we read the Databricks AI Index, using several model providers is now the majority pattern on Unity Gateway: 71.1% of active companies had used two or more providers by the week of 2 November and 31.4% four or more, against 36% and 4.4% in the week of 20 July.
What the Index Counts
The index page counts, “Among companies active each week, counts every provider a company has used since July 20, 2026. Gemini and Gemma count as one provider, as do GPT and GPT-OSS.” Databricks co-founder Patrick Wendell wrote that the data “covers >10,000 companies with broad representation (Digital Natives, Fortune 500, Global 2000, etc…)”.
Because providers accumulate from 20 July, a company that tried a second provider once in August still counts in November. The chart rose in most weeks; its biggest one-week gains came in the week of 27 July, from 36% to 44.5%, and in the week of 10 August, from 42.7% to 49.8%.

What Was Posted on X
Ali Ghodsi, Databricks’ CEO, posted at 17:26 UTC on 16 November, quoting Patrick Wendell’s thread, that six months ago most enterprises had one model provider and that today 75% have two or more and 50% three or more. This piece reports his figures as posted next to the index chart’s values for the week of 2 November, 71.1% and 47.9%.
In the second post of his thread Wendell, a Databricks co-founder, gives the same 75% for two or more providers, says almost half use three or more, and names OpenAI‘s GPT models and open-source models as the most common new providers. The headline’s ‘OpenAI Among Adds’ rests on this post.
What the Executives Said
Ghodsi wrote: “6 months ago most enterprises had 1 model provider, today 75% have 2+ and 50% have 3+ model providers.” He described the data as “a cross-section of over 10,000 large enterprises using our gateway”.
Wendell wrote: “At mid year most companies had one model provider. Today, 75% have two or more, and almost half use three or more.” The page’s chart, marked updated 2 November, shows 71.1% and 47.9% for the latest week, 2 November; both executives posted on 16 November.

Who Gets Added
Wendell wrote that “The most common” new providers “are OpenAI (GPT models) and OSS models”. The index’s share charts point the same way: OpenAI’s share of proprietary-model tokens rose from 16.4% to 37.0% by our sums, and open-weight models went from 7.3% to 30.9% of tokens to Databricks-hosted models.
Our set on the index also covers Anthropic’s 70% of closed-model spend, the open-weight rise and median user spend. Earlier, we covered Menlo Ventures’ enterprise AI spend figures and Claude models in Microsoft 365 Copilot.
The Structural Read
As we read the method, the measure is cumulative: the page counts every provider a company has used since 20 July, so it shows how many companies have tried a second provider, not how evenly they split their work.
As we read the executives’ posts, their round numbers run a little ahead of the chart: Ghodsi and Wendell wrote 75% for two or more providers, while the chart’s latest weekly value is 71.1%.
Ali Ghodsi, Databricks CEO, on X
“6 months ago most enterprises had 1 model provider, today 75% have 2+ and 50% have 3+ model providers.”
Three Implications
FOR MODEL MAKERS Wendell names OpenAI’s GPT models and open-weight models as the most common new providers; OpenAI’s proprietary token share rose from 16.4% to 37.0% by our sums.
FOR BUYERS The page counts Gemini and Gemma as one provider and GPT and GPT-OSS as one, so mixing a lab’s open and closed models does not add a provider.
RELATED Our first piece on the index covers which labs get the spend.
The Business Engineer Lens
This story maps onto the Business Engineer framework Enterprise AI: From Software to Substrate.
The framework argues that enterprise vendors sell irreversibility, and that “The product is the interconnection.”
As we read it, the index points at where that interconnection sits for model access: 71.1% of active Unity Gateway companies have used two or more model providers through one gateway. The framework predates the index and does not discuss it.
What Is Not Established
The index counts providers a company has used since 20 July, not how much work each provider carries. The figures cover companies that manage AI through Unity Gateway, and the 75% and 50% are the executives’ figures as posted, a little above the chart’s latest weekly values.
We read the index page in full, including the values behind each chart, Wendell’s six-post thread and Ghodsi’s post. We did not contact Databricks for this piece.
The Bottom Line
71.1% of active Unity Gateway companies had used two or more model providers in the week of 2 November, up from 36% in the week of 20 July, the Databricks AI Index shows; Ghodsi and Wendell put the figure at 75%.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. The figures come from the Databricks AI Index page at developers.databricks.com/ai-index, which we read in full on 16 November 2026, including the exact values its charts plot, from Databricks co-founder Patrick Wendell’s six-post thread on X and from Databricks CEO Ali Ghodsi’s post on X, both of the same day. The page marks the multi-provider chart as updated on 2 November 2026; the 75% and 50% figures are the executives’ as posted. Growth multiples and provider totals are our arithmetic from the page’s values. Databricks was not contacted for this piece.
Sources: Databricks AI Index (page updated 2 Nov 2026, published 16 Nov 2026) · Ali Ghodsi on X, 16 Nov 2026 · Patrick Wendell on X, 16 Nov 2026 (post 2 of 6)









