The tech giant, Google, launched “implicit caching” for its Gemini 2.5 models to cut costs without the need for manual setup. The feature, which quietly began rolling out this week, promises up to 75% savings on repetitive inputs.
This development will provide a big boost to crypto startups that planning to use artificial intelligence (AI) infrastructure.
How Google’s New Update Makes Gemini AI More Affordable
According to the official blog post, implicit caching allows AI models to recognise and reuse repeated parts of a user’s request.
One can think of it like a smart memory. If you often send similar instructions to the model, it will not charge you the full price every time. Instead, it records the repeated part and gives you a hefty discount.
The new development will bring major savings for those crypto and blockchain developers who often run large batches of similar queries for wallet analysis, fraud detection, or identifying market trends.
Unlike previous systems that required developers to manually mark which parts of a prompt to save (known as “explicit caching”), this new system works automatically in the background.
This update supports Gemini 2.5 Pro and Flash, with token limits adjusted to 2,048 and 1,024 tokens, respectively. It is roughly equivalent to 1,500 and 750 words.
It will encourage developers to keep the repetitive part of their prompt at the beginning and add changing details at the end to improve their chances of triggering a “cache hit.”
The change comes after a wave of developer frustration with Google’s earlier caching model, which some claimed led to unexpectedly high API costs. In response, Google not only apologized but also promised smoother, more transparent billing, and implicit caching is the answer.
Still, some questions remain. There is no independent verification yet that these savings will hold up at scale. But if early feedback is positive, this could reshape how AI is deployed across the crypto ecosystem, particularly for lean teams that depend on cost-efficient automation.
For now, Google’s implicit caching is enabled by default, meaning developers already using Gemini 2.5 may start seeing savings automatically. For an industry where every satoshi matters, that is a welcome development.
Also Read: Meltwater Launches Mira to Speed Up AI Media Monitoring
