We often hear that AI will act to “democratize” access to many aspects of technology, from data analytics to coding. But what exactly does this mean?

The basic premise is that by lowering the barriers to entry, activities that previously required large budgets and teams of technical specialists are accessible to any business. Or that’s the theory, at least.

On the other hand, the most powerful algorithms and foundation models are owned and controlled by a handful of giant, multinational corporations. Meaning the advantages they give could be withdrawn at any point, should Google, Amazon or whoever decide to flick the switch, or raise the prices so they become out of reach of smaller companies.

These two possibilities paint very different pictures of how the next decade of business competition could play out. The power to use technology to build and create could increasingly be available to us all, or further concentrated in the hands of billionaires and tech giants.

So which is most likely? Let’s take a look at what we’ve learned so far…

Let’s start by digging into some relevant stats to see what we can find:

Starting with cost, it’s clear that thanks to the popularization of generative AI and the spread of tools like ChatGPT, the cost of accessing AI has fallen dramatically.

According to Stanford’s AI Index , the cost of processing a million tokens (the standard unit of AI compute power) fell from around $20 to 7 cents between 2022, when ChatGPT was launched and 2024.

On the face of it, this has made AI much more accessible, with projects that once would have needed careful scoping and budgeting now viable on a whim.

This has led to a surge in the popularity of vibe coding , the practice of using AI coding tools like Claude Code or Codex to quickly and cheaply throw together tools, experiments and prototypes.

The effect on small businesses has been tracked by analysts including those at JP Morgan Chase. By tracking payments to AI service providers, they found that firms in 2025 take an average of six months to reach a 10 percent adoption rate, whereas in 2019 it was taking around six years. This speed of adoption, it says , is unprecedented “compared to previous technologies like electricity, personal computers and the internet.”

And there’s more evidence suggesting small businesses are outpacing larger ones when it comes to leveraging AI.

The US Small Business Administration’s Office of Advocacy reported in late 2025 that companies with less than 250 employees lead the way when it comes to adoption of nearly half of the 17 tracked AI use cases. These include marketing automation, natural language processing, large language models and speech recognition.

However, larger organizations (defined as those with more than 250 employees) outpace smaller competitors when it comes to more tech-intensive deployments including image recognition, computer vision, biometrics and data analytics.

Other statistics, however, paint a vividly different picture.

A 2025 Eurostat report found a 38-point gap between larger businesses (55 percent) and smaller businesses (17 percent) using AI. And this year (2026) research by the US Federal Reserve concluded that it is “hard to say whether AI-related gains are accruing disproportionately to the largest firms or whether AI is serving as an equalizer of sorts and helping smaller enterprises compete more effectively."

Their report also goes on to say that when smaller businesses have managed to adopt AI, the impact on their business may be proportionately higher.

And here’s another unexpected finding: While larger businesses generally expect AI to cut human headcounts, smaller businesses that are using it believe it could cause them to grow. The reason? They’ll win more business and need more humans to deal with it.

So, four years into the latest leg of the AI revolution, it’s fair to say that the situation is still playing out, and both of the hypothetical scenarios we’re considering are probably too simplistic.

Evidence available today shows that giants like Google and Microsoft have reaped the benefits in terms of revenue and share price growth. However, smaller ones are adopting faster than at any other point in history, and are ahead of larger firms across a significant share of everyday use cases.

Often the barrier preventing smaller businesses from getting involved is a lack of AI literacy and understanding of what the technology offers. As the SBA research shows, many simply don’t believe that AI is applicable, or understand the benefits.

At the end of the day, I don’t believe it will be access to technology or sheer scale of compute power that will really differentiate between winners and losers in the AI race.

Regulation will be hugely important. Heavy-handed implementation could clearly benefit organizations with big enough legal teams to navigate complex new rules and legislation.

Alternatively, regulators could act to deliberately lighten the load on smaller businesses, to encourage innovation to thrive. How this will play out will vary from jurisdiction to jurisdiction but is likely to be a deciding factor.

Even more significant, though, will be human ingenuity. Fresh thinking, smart ideas, and a firm focus on building services around what people really need will, I believe, be more valuable in the long run than size and budget.

The Amazon or Netflix of the AI age will emerge thanks to humans understanding better than others what this amazing technology makes possible. Not by simply outspending competitors.

In other words, there’s still everything to play for. The rules may sometimes seem rigged in favor of the tech giants that control today’s most powerful frontier models. But the companies that truly define this era are likely to be the ones who find the most useful things to do with it.