Github copilot shifts to usage based pricing model from June

Tuesday 28 April 2026 - 10:00

By: Dakir Madiha

Github copilot shifts to usage based pricing model from June

GitHub will move all Copilot subscriptions to a usage based pricing system starting June 1, replacing fixed per seat plans with a model tied to actual AI consumption. The change introduces GitHub AI Credits, where customers pay based on the number of tokens processed during interactions. Each credit is priced at $0.01, aligning costs directly with how much the tool is used rather than the number of users.

The shift reflects a broader transformation across the software industry as artificial intelligence reshapes pricing strategies. Companies are moving away from traditional subscription models toward consumption based billing. By the end of 2025, the number of enterprise software providers adopting such models had more than doubled compared with the previous year. This transition signals a structural change in how software value is measured and monetized.

Major software firms have already introduced similar approaches. HubSpot revised its AI pricing to charge per resolved interaction instead of a fixed fee. Salesforce introduced agentic work units as a pricing metric tied to output. Workday launched a prepaid credit system for AI usage, while ServiceNow adopted a hybrid structure combining fixed commitments with flexible consumption metrics. Adobe also integrated credit based pricing for AI generated content through its Firefly tools.

The economics of artificial intelligence are driving this shift. Running large language models and AI agents requires significant computing power, particularly GPU resources, which creates variable costs depending on usage intensity. Fixed pricing models fail to capture these differences. Usage based billing allows providers to recover infrastructure costs while offering flexibility to customers whose AI needs vary widely.

Industry data shows the scale of this transition. By 2026, about 74 percent of software companies had implemented some form of usage based pricing. At the same time, software prices increased by an average of 8 to 12 percent in 2025, partly due to rising AI infrastructure costs. This trend, often described as an AI related price increase, reflects the growing financial burden of deploying advanced models at scale.

Markets have responded positively to these developments. Shares of application software companies rose as investors signaled confidence in the sector’s ability to generate revenue from AI. The Nasdaq Composite reached record levels following a recovery from earlier declines, reinforcing expectations that AI will strengthen rather than disrupt leading software firms. However, the shift may introduce less predictable spending patterns for customers, as costs will now fluctuate with actual usage rather than fixed subscriptions.