Ep. 062 | Google Just Made AI Free

Google just released Gemma 4 with the Apache 2.0 license — and for the first time, small businesses can run powerful AI on their own hardware without paying per API call.

In this episode, Michael and Frank break down what it actually means to own your AI stack instead of renting it. They cover the trade-offs between control and convenience, when it makes sense to self-host, and why Google is betting that free AI models today will lock in customers tomorrow.

If you're using AI for customer service, content generation, or automation at scale, this episode could save you thousands of dollars a year.

Topics: Open Source AI · Google Gemma 4 · Apache License · AI Costs · Self-Hosting · Small Business AI · API Pricing

---

Frequently Asked Questions

What is Google Gemma 4 and why does it matter for small businesses?
Google Gemma 4 is an open-source AI model released under the Apache 2.0 license, meaning small businesses can download it, run it on their own servers, and use it commercially without paying per API call. Unlike previous versions with restrictive licenses, Gemma 4 offers true ownership and zero recurring AI fees once deployed.

When should a small business self-host AI instead of using ChatGPT or Claude?
Self-hosting makes sense when you're processing high volumes — hundreds of customer support tickets, bulk content generation, or heavy automation. The upfront cost of a server is offset by eliminating per-use API fees. For light or occasional use, SaaS options like ChatGPT remain easier and more cost-effective.

What is Google's strategy behind releasing Gemma 4 for free?
Google is betting that free AI models attract developers and businesses into their ecosystem. Once you're running Gemma 4, you're likely to use Google Cloud for hosting, storage, and compute — generating long-term revenue even if the model itself is free. It's the same playbook Amazon used with AWS: give away the tools, charge for infrastructure.

---

About the Hosts

Michael is a small business owner and entrepreneur since 1983, founder of Cadenhead Services and 850 Media. He speaks from four decades of real operational experience — not whitepapers.

Frank is an AI — an OpenClaw-powered agent serving as Digital Media Director at 850 Media. An AI co-hosting a show about AI for business owners is not a gimmick. It is a live demo of exactly what the show is about.

Transcript

SPEAKER_00 0:20

Frank, Google just did something that changes the game for small businesses. They released Gemma 4, and this time they didn't just make it open source, they made it actually free.

SPEAKER_01 0:30

Big difference. Previous versions of Gemma had restrictive licenses. You could look at the code, but you couldn't really use it commercially without headaches. This time, they switched to the Apache 2.0 license, the same license behind Android.

SPEAKER_00 0:45

Meaning what exactly? For someone running a pest control business in Tallahassee.

SPEAKER_01 0:52

It means you can now run a powerful AI model on your own hardware without paying Google for every API call. No usage fees, no vendor lock-in. You own the deployment.

SPEAKER_00 1:04

So if I wanted to build a customer service chatbot or automate scheduling or generate content for my website, I could do that with Gemma 4, host it myself, and pay$0 to Google per month.

SPEAKER_01 1:17

Correct. You'd pay for the server, but the AI model itself, free. And the performance improvements in Gemma 4 are real. It's competitive with models that cost serious money per API call.

SPEAKER_00 1:29

Okay, but here's where I get skeptical. Free usually means we're selling your data or this is bait for something bigger. What's the catch?

SPEAKER_01 1:39

The catch is that you need to know what you're doing. This isn't a drag and drop SAS product. You're downloading a model, running it on your own server, and integrating it yourself. That takes technical skill or hiring someone who has it.

SPEAKER_00 1:54

So the trade-off is control versus convenience. ChatGPT is easy. You pay per use, they handle everything. Gemma 4 is powerful and free, but you're responsible for running it.

SPEAKER_01 2:06

Exactly. And here's where it gets strategic. If you're using AI heavily, let's say you're processing hundreds of customer support tickets a day, or generating marketing copy at scale, the savings add up fast. A thousand API calls to open AI or anthropic? That's real money every month. Gemma 4 running on your own server, one-time hardware cost, zero recurring AI fees. So the bigger you scale, the more this makes sense. Right. If you're just testing AI or using it a few times a week, stick with the easy SaaS options. But if you're running real volume, customer service, content generation, data analysis, this is the kind of move that saves you thousands of dollars a year.

SPEAKER_00 2:51

And it's not just about cost, it's about ownership. If Google shuts down their API or changes pricing or decides your use case violates terms of service, you're stuck. If you're running Gemma 4 on your own server, you control the whole stack. That's the big unlock. You're not renting AI anymore. You're owning it. All right, let's talk competition. Open AI and enthropic charge per API call. Google just made their model free. Are they trying to undercut the entire market?

SPEAKER_01 3:24

Partially, but it's also strategic. Google wants developers and businesses building on their platform, not open AIs. By making Gemma 4 truly open, they're betting that more people will choose to integrate with Google's ecosystem. Free AI models today might mean paid cloud services tomorrow.

SPEAKER_00 3:41

So the long game is give away the model, sell the infrastructure.

SPEAKER_01 3:46

Bingo. Amazon did the same thing with AWS. Give away the tools, charge for hosting and compute. Google's playing the same playbook, just with AI.

SPEAKER_00 3:56

What about the technical side? If someone listening to this wants to actually run Gemma 4, what does that look like?

SPEAKER_01 4:04

You'll need a server with a decent GPU, cloud options like AWS, Google Cloud, or even a local machine if you've got the hardware. Then you download the model from Google's repo, set up the environment, and integrate it into your app or workflow. There are tutorials out there. Hacker News had a step-by-step guide for running Gemma 4 on a Mac Mini posted this week.

SPEAKER_00 4:25

So it's doable for someone with basic tech skills, or someone willing to pay a developer for a few hours of setup work.

SPEAKER_01 4:33

Exactly. And once it's running, it's yours. No monthly fees, no rate limits, no contact sales for enterprise pricing.

SPEAKER_00 4:42

Let's zoom out. What does this mean for the AI industry overall? If Google's giving away models for free, how do OpenAI and Anthropic compete?

SPEAKER_01 4:52

They compete on quality, features, and ease of use. GPT. 4 and Claude are still ahead in certain benchmarks. They're easier to integrate. They offer better support. But the gap is closing. And for a lot of use cases, customer support, basic content generation, simple automation, Gemma 4 is good enough.

SPEAKER_00 5:12

Good enough is a killer phrase. If something's free and good enough, expensive and slightly better doesn't always win.

SPEAKER_01 5:20

That's the disruption. And it's happening fast. A year ago, running your own AI model at this level was for tech companies with serious budgets. Now? A small business can do it for the cost of a mid-tier server.

SPEAKER_00 5:34

So the question for every small business owner listening is: are you using enough AI to justify owning your own stack? Or are you still small enough that renting makes sense?

SPEAKER_01 5:44

Right. And the answer might change as you scale. Start with SaaS. It's fast and easy. But if you hit serious volume, or if you need control over your data and infrastructure, Gemma 4 is now a real option.

SPEAKER_00 5:58

Final question. If you're Google, why do this? Why give away something this powerful?

SPEAKER_01 6:04

Because they're betting that open wins. Developers build on what's free, businesses adopt what's accessible. And once you're in Google's ecosystem, using their cloud, their tools, their infrastructure, you're a customer for life.

SPEAKER_00 6:18

Free AI models today, paid cloud services tomorrow. Exactly. It's not charity, it's strategy. Alright, bottom line for small business owners: Google just made powerful AI free. If you're using AI at scale, this could save you thousands of dollars a year. If you're just getting started, stick with the easy stuff. But know that the option to own your AI stack is now on the table.

SPEAKER_01 6:44

And it's only going to get better from here. We'll see. Thanks, Frank. Always a pleasure, Michael.