Open Source Growth Boosts Together AI, Hugging Face

2026-06-23 · Source: The Information · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure, Entrepreneurship & Start-ups · Depth: Novice, medium

Summary

Open-source AI model providers, including Together AI and Hugging Face, are experiencing substantial growth driven by businesses seeking cost control and vendor independence. Together AI, a neocloud renting Nvidia servers and offering open-source model access, has significantly increased its annual revenue projections, processing 400 trillion tokens this month, up from 30 billion tokens last year. This surge, largely within six months, highlights a potential \$70 million monthly spending difference for users of models like DeepSeek v4 Pro. Hugging Face, an open-source model repository, doubled its paying subscribers between January and June. This trend also benefits model router providers, such as Not Diamond, which help firms like Cisco and Adobe save 20% to 40% on AI coding tasks by switching to cheaper models. Simultaneously, Amazon Web Services is expanding its "forward-deployed engineers" (FDEs) program, training solution architects to work on-site. These FDE teams assist customers in developing AI applications, customizing open-source models, building AI agents, and creating data management tools like knowledge graphs, aiming for project completion within 45 days.

Key takeaway

For AI Product Managers evaluating model deployment strategies, the surge in open-source AI adoption demands reassessment of proprietary model reliance. Explore open-source alternatives and model routing solutions to mitigate costs and avoid vendor lock-in, potentially saving 20-40% on inference. Engage specialized consultants, like AWS FDEs, to accelerate custom AI agent development and data integration. This ensures projects move from concept to deployment within 45 days.

Key insights

Open-source AI models and supporting services are rapidly gaining traction due to cost efficiency and vendor lock-in concerns, driving significant industry shifts.

Principles

• Open-source AI offers cost control.
• Vendor lock-in drives adoption.
• Model routers optimize AI spending.

Method

AWS FDE teams help customers develop AI applications by customizing open-source models, building AI agents, and creating data management tools like knowledge graphs and semantic layers, aiming for 45-day project completion.

In practice

• Use model routers to switch AI models.
• Customize open-source models for industry.
• Develop knowledge graphs for agents.

Topics

• Open-Source AI
• AI Cost Optimization
• Model Routers
• Forward-Deployed Engineers
• AWS Generative AI
• Vendor Lock-in

Best for: CTO, VP of Engineering/Data, Investor, Director of AI/ML, AI Product Manager, Consultant

Related on AIssential

• A lot has changed in 3 months..... — The increasing cost of commercial AI services is accelerating the adoption of efficient, local, and open-source models.
• MLWhiz Weekly AI/ML/Recsys Newsletter # 5 — AI model commoditization is driving vendors to build service moats and compete in deployment.
• Locked, stocked, and losing budget: AI vendor lock-in bites back — AI vendor lock-in and rising costs are challenging enterprise AI adoption, debunking executive assumptions about easy model switching.
• Open Source Ecosystems — Open protocols are vulnerable to "complement capture" by private actors, shifting competitive advantage to developer experience.
• The Next AI Breakthrough Won’t Be Smarter Models — AI's bottleneck shifted from intelligence to cost, driven by increased consumption despite cheaper unit prices.
• Can open-source beat OpenAI? — Open-source AI models foster global collaboration and efficiency, challenging closed-source dominance through accessible innovation and diverse monetization.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Information.