OpenAI Gains on Anthropic in Corporate AI Spending, Ramp Data Shows

Corporate card and spend-management platform Ramp released updated enterprise purchasing data indicating that OpenAI is growing faster than Anthropic among US businesses in the third quarter of 2026, closing the gap after losing the top spot earlier in the year. The metrics, compiled from transaction data across more than 70,000 businesses using Ramp corporate cards and invoice processing, highlight ongoing volatility in enterprise model selection. Market Share Trajectory and Model Drivers A

1 min
OpenAI Gains on Anthropic in Corporate AI Spending, Ramp Data Shows

Corporate card and spend-management platform Ramp released updated enterprise purchasing data indicating that OpenAI is growing faster than Anthropic among US businesses in the third quarter of 2026, closing the gap after losing the top spot earlier in the year.

The metrics, compiled from transaction data across more than 70,000 businesses using Ramp corporate cards and invoice processing, highlight ongoing volatility in enterprise model selection.

Market Share Trajectory and Model Drivers

Anthropic overtook OpenAI in business adoption on Ramp's platform in May 2026, capturing 41% market share compared to OpenAI's 39.5%. By July, Anthropic maintained a lead of roughly 44% against OpenAI's 40%.

Enterprise model selection dynamics and cost retention tradeoffs

According to Ramp lead economist Ara Kharazian, OpenAI's Q3 rebound is primarily driven by developer traction around its GPT-5.6 Sol model. Conversely, Anthropic's high-tier Fable 5 model faced adoption headwinds due to premium pricing structures and mandatory 30-day data retention requirements imposed by regulatory agreements.

Broadening Enterprise Penetration

While the two providers trade market share, overall corporate AI investment continues to climb across mid-market and tech-forward organizations:

  • Overall Adoption: The share of Ramp customer companies paying for AI subscriptions reached nearly 56% in July 2026, up from 50% in March.
  • Provider Volatility: Organizations continue to switch primary model providers quickly in response to checkpoint releases and price-performance shifts.
  • Scope Limitations: Ramp's dataset reflects direct card and invoice spend, excluding multi-year enterprise volume commitments through hyperscaler cloud agreements such as Microsoft Azure or AWS Bedrock.

Sources

Written by

More to read

  • TVA Board Approves Dedicated Data Center Rate Class to Shield Households from AI Compute Costs

    The Board of Directors of the Tennessee Valley Authority (TVA) voted on August 20, 2026, to establish a dedicated wholesale rate class for large data centers. The tariff restructuring is designed to insulate residential consumers and small commercial businesses from the escalating capital expenditures required to expand the power grid for artificial intelligence workloads. Approved during the board's quarterly meeting in Memphis, Tennessee, the package introduces targeted tariffs for facilities

    1 min
  • Fast Model Weight Loading in Production: Safetensors, Tensorizer, and Direct GPU Deserialization

    Fast Model Weight Loading in Production: Safetensors, Tensorizer, and Direct GPU Deserialization In modern large language model inference clusters, cold start latency is rarely bounded by GPU compute allocation. Instead, the operational bottleneck centers on storage I/O and weight deserialization. As foundation models scale from 70 billion to 405 billion parameters, raw weight footprints range from 140 GB to over 800 GB in standard 16-bit precision. On naive serving stacks, deserializing these

    1 min
  • Maximal Update Parametrization (muP): How Tensor Programs Enable Zero-Shot Hyperparameter Transfer in LLM Pre-Training

    Pre-training a frontier large language model requires hundreds of thousands of GPU hours and millions of dollars in compute. At that scale, traditional hyperparameter tuning is financially and operationally impossible: teams cannot sweep learning rates, weight initializations, or optimizer betas across multiple 70B parameter runs to find the loss minimum. Historically, practitioners relied on ad-hoc heuristic extrapolation or manual guesses from small runs, often leading to sub-optimal loss curv

    1 min