The Myth About AI — AI Is Expensive
AI costs are declining fast. Open source models and local hardware are making AI accessible to everyone.
In my previous posts, I showed how AI is an advantage for everyone who dreams and how individuals should not worry about job replacement but focus on how AI creates new opportunities. In this post, I address those who aren't excited about AI and believe the normal way of doing business should continue without it. The biggest concern I hear is cost — and even major organizations have been lamenting about this.
There is no sugarcoating it: AI usage is costing billions worldwide. Google just paid SpaceX almost one billion for AI compute storage, and to stay profitable, that expense must be passed to end-users. So why should non-AI users consider it? Because it would be FREE soon. Knowledge can neither be created nor destroyed, and AI is built on knowledge — LLMs get their intelligence from mass data input from already existing systems and content.
It is actually immoral to put a price on access to that knowledge. What people pay for is the compute power that processes it. Study the trends: LLM profitability is a dying dream. Companies depending solely on that model have been pivoting. Everyone believes there's an AI bubble about to burst. But one company showed us how the future of AI looks, and the world missed it.
In June 2026, Nvidia announced RTX Spark PCs with up to 128GB unified memory and up to 1 Petaflop FP4 AI performance — computers that can run AI locally with ease. What if you didn't need software to interact with servers? What if the server was your own computer? What if you didn't need an internet connection to communicate with AI because it's sitting right in your machine? That is the power of local models.
Nvidia just showed the world that Jarvis is no longer just for one rich billionaire named Iron Man. In a few years, AI will be for everyone with a computer. Start your journey with local models and open source tools like Ollama. Within the next five years, software AI agents will be out of vogue — everyone with a modern computer will have an AI agent working locally. That is the future I envision.
