Google Cloud Indonesia has warned companies that the biggest challenge in deploying AI at scale is no longer the technology itself, but the cost behind it. As AI moves further beyond the experimentation stage, token consumption can quickly push business implementations far above initial budgets.
Country Director Google Cloud Indonesia Karim Siregar said AI has entered a more advanced phase and should now be treated as a practical tool for solving everyday business problems. That shift, however, makes cost control far more important than before.
Token-based billing can raise expenses quickly
Karim explained at the Google Cloud Media Briefing in Jakarta on Wednesday, 15/7/2026, that AI billing is now often based on tokens. He compared the system to prepaid electricity or mobile credit, where payment depends on usage.
“We must not forget the impact of AI usage costs. AI is charged using tokens now, friends. If you know tokens, it is like PLN, where payment is based on usage, like prepaid credit. With AI, it is based on tokens,” Karim said.
That model means companies can face a sharp rise in costs if usage grows faster than expected. The issue is not only how much is spent, but also whether the output creates enough business value to justify the bill.
Why small-scale testing matters
Karim stressed that companies should begin with small-scale trials before expanding AI across a full organization. A limited rollout makes it easier to understand usage patterns and estimate future expenses with more accuracy.
This approach also helps businesses avoid surprises when consumption rises during broader deployment. By testing first, companies can measure whether token spending remains reasonable before committing to a larger rollout.
| Focus | Explanation | Impact on Companies |
|---|---|---|
| Token costs | AI is billed based on token usage | Spending can rise quickly when usage is heavy |
| Small trials | Implementation starts on a limited scale | Helps estimate costs before a wider rollout |
| Usage control | Tokens must be monitored carefully in enterprise use | Makes cost and business value easier to assess |
The message from Google Cloud Indonesia reflects a broader reality for companies adopting AI: capability alone is not enough. Businesses also need to manage token consumption carefully so the technology remains financially sustainable as it moves deeper into daily operations.