The Hidden Strategy Behind Open Source AI That Malaysian SME Owners Need to Know
If you own a small or medium business in Malaysia, you might feel a constant pressure to keep up with the latest AI models. But recent data suggests this pursuit isn’t necessary—and could even misdirect your efforts. Understanding how open source and premium AI really work together might save you time and reduce the fear of falling behind.
TL;DR: Open source models are not replacing premium AI. They serve different stages: premium models for discovery, open source for mature tasks. Overall spend on premium tools stays high as new use cases emerge. For SME owners, this means you can start with accessible models and only upgrade when testing new ideas.
How Open Source Models Fit Into Real AI Usage
Decagon CEO Jesse Zhang published a theory suggesting that open source and frontier models aren’t competitors but phases of the same lifecycle. As use cases mature, they shift to lighter models, but new applications keep driving demand for premium tools. Data from Vercel’s AI gateway dashboard shows that while models like DeepSeek now process more tokens in volume, Anthropic still accounts for over half of total AI spend on the platform. Similarly, OpenRouter figures indicate that premium models maintain dominant spending despite lower token volumes.
“The frontier labs will keep owning discovery. Open source will increasingly own production.” – Jesse Zhang, Decagon CEO (via TechCrunch)
What This Two-Tier System Means for Your Business
For Malaysian SMEs, this dynamic offers flexibility. You can rely on free or low-resource open source models for established processes like customer support or data entry, while reserving premium models for innovative projects that need robust performance. This approach helps you stay competitive without constantly chasing the latest AI release.
Consider your own operations. Routine tasks that AI handles well can likely be managed by open source models. For new product ideas or complex problem-solving, you might want the extra capability of frontier tools. The key is matching the model to the task’s maturity.
Comparing AI Model Types for SME Use
| Aspect | Frontier Models (e.g., Anthropic) | Open Source Models (e.g., DeepSeek) |
|---|---|---|
| Best for | Testing new ideas, complex problems | Running proven processes, daily operations |
| Resource intensity | Higher demand, more specialized | Lower demand, easier to deploy |
| Adoption stage | Early-stage discovery | Mature production |
The Bigger Picture: A Stable AI Economy for Business Owners
This division between frontier and open source models is likely not a temporary phase. According to the TechCrunch report, it may become a persistent feature of the AI industry. For SME owners, this stability means you don’t need to predict the next big AI winner. Instead, focus on what works for your current needs and scale as your business evolves.
Even with new entrants like Nvidia’s Nemotron on the horizon, the overall pattern seems set to continue. This should reduce any urgency to switch tools frequently. You can adopt AI at your own pace, using different models as your processes mature. The path from discovery to production is becoming clearer, and your business can benefit from both tiers without being tied to one.
Book a free 15-min call to see how open source AI can fit into your business → https://autorunbiz.com