What happened
Vercel's June 2026 AI Gateway Production Index reveals a notable shift in how developers and businesses are deploying large language models: DeepSeek has emerged as a serious contender for raw token volume, signaling rapid adoption across cost-sensitive workloads. Anthropic, however, continues to dominate on revenue spend, suggesting enterprises are willing to pay a premium for its models on high-stakes tasks. Separately, GitHub published a beginner-oriented resource addressing frequently asked questions about the platform, reinforcing its ongoing effort to lower the barrier to entry for new developers joining collaborative software workflows.
Why it matters for your business
The split between token volume and dollar spend is a meaningful signal for technology and operations leaders: DeepSeek's volume growth suggests it is being routed to high-frequency, lower-margin inference tasks where cost efficiency matters most, while Anthropic's spending dominance points to its models being selected for quality-critical applications such as customer-facing agents or compliance-sensitive automation. Organizations building multi-model AI architectures should treat these indices as a routing guide rather than a single-vendor endorsement. Meanwhile, GitHub's continued investment in onboarding documentation lowers the ramp-up cost for engineering teams expanding headcount or integrating contractors, reducing friction in collaborative development pipelines.
What to watch next
Monitor whether DeepSeek's token-volume lead translates into increased enterprise spend over the next two quarters, which would signal a maturing trust relationship with larger buyers. On the developer tooling side, watch for GitHub to expand its beginner resources into more structured learning paths as competition with platforms like GitLab and Bitbucket intensifies for first-time users who become long-term enterprise customers.