What are the forces behind AI’s exponential growth?

Everywhere you turn, AI is in the headlines. The pace feels breathless , new models, new breakthroughs, new valuations, new use cases. Many executives are asking: Why is this happening so quickly? What’s really powering this exponential growth?

The short answer is that AI’s rise is not random. It is systemic, driven by eight powerful forces that reinforce each other in a compounding loop, a flywheel effect that is now spinning on its own.

The eight forces accelerating AI are:

1. Data:  AI is incredibly data hungry and getting hungrier. Training state of the art models once required millions of data points, now it takes trillions. We are approaching the limits of high quality human generated data and are turning to synthetic data. Synthetic data is artificially generated datasets produced by algorithms to mimic real world information.

   
   

   

   
   

2. Algorithms & Architectures:  A breakthrough that accelerated large language models (LLMs) came with the transformer model in 2017 from Google. This unlocked the leap to LLM’s by enabling parallel processing of data and far better handling of context. That architecture gave us ChatGPT, Claude, Gemini, and a wave of other models. Progress doesn’t stop there, innovation is relentless. Providers are now pushing to higher scale and experimenting with architectures that improve reasoning, efficiency and memory.

3. Adoption:  Adoption has been explosive. ChatGPT became the fastest growing consumer app in history, reaching 100 million users within the first two months and today it has over 700 million weekly active users and a billion daily queries. Other AI services have followed the same trajectory were image generators like MidJourney and Stable Diffusion. This wave of demand shows that AI isn’t just a niche tool, it’s becoming embedded in everyday life and business workflows at unprecedented speed.

4. Capital:  Capital appears to be unlimited. Corporate AI investment surged to $252.3B in 2024 (private funding plus M&A). And 2025 is set to be even bigger. Amazon, Alphabet, Meta, and Microsoft combined are forecast to pour $320–$365B into capex, largely on GPUs, data centres, and cloud infrastructure to power AI at scale.

5. Energy:  AI is incredibly power hungry. Training and running large models consume vast amounts of electricity, not just for GPUs, but also for cooling, storage, and networking. By 2030, global AI data centre demand could reach over 320 GW of power, up from less than 90 GW in 2022. To put that in perspective, that’s roughly the equivalent to the entire annual electricity consumption of Japan. A staggering footprint.

6. Regulatory:  Today’s relatively light regulatory environment has fuelled AI’s rapid growth. With few hard rules in place, companies have been free to move fast, experiment, and scale adoption without the drag of compliance overheads. That flexibility has been critical to the momentum we see today.

   But this landscape will not last forever. As models grow more powerful and increasingly embedded in critical systems, so too will the risks of misuse and unintended consequences. Governments will inevitably step in with clearer guardrails to protect safety and trust. The challenge lies in striking the right balance, introducing enough regulation to safeguard society, without stifling innovation, and doing so in a way that achieves some level of global consensus.

7. Talent:  The world’s top AI researchers and engineers are in extremely short supply, concentrated in a handful of universities, labs, and big tech firms. This scarcity drives fierce competition, salaries can reach into the millions, and companies routinely acquire start ups just to secure their teams.  Over time, we can expect the gap to narrow as more talent is trained, open source tools improve, and specialised AI skills become mainstream.

8. Compute power:  Access to compute has become the defining factor in AI progress. Training frontier models requires tens of thousands of GPUs running for weeks at a time, at costs that can reach hundreds of millions of dollars per model.  Supply is constrained, NVIDIA’s high end GPUs remain scarce, lead times stretch into months, and the cost of access is soaring.  This is a clear barrier to entry for smaller technology companies.

   
   

   

Three of the eight are significant risks to further exponential growth, big tech has realised this and is working hard to mitigate:

1. Energy:  The scale of consumption, now and forecast, has triggered concerns about sustainability and grid resilience. In response, Big Tech is not waiting for governments to act. Companies like Microsoft, Google, and Meta are investing directly in their own nuclear fusion energy sources.

2. Talent:  Scarcity of talent has prompted a fierce battle with some AI talent able to write their own cheques.  Meta, for example, has recently offered multimillion dollar compensation packages and long term equity incentives.  Open AI has awarded large retention bonuses to stop top talent from being poached by rivals.  This is driving costs up further and pricing smaller tech companies out of the market, which may in the long run stifle innovation.

3. Compute power : Nvidia’s trillion dollar company valuation demonstrates how important and scarce compute power is to exponential growth.  To reduce this dependence, several of the largest players are taking matters into their own hands.  Google has developed its own TPUs (Tensor Processing Units).  Amazon has invested in Trainium and Inferentia chips for AWS.  Microsoft, in partnership with AMD, has unveiled its own Athena AI accelerator to supplement Nvidia GPUs. Meta, too, is doubling down on in-house silicon with its MTIA (Meta Training and Inference Accelerator) chip program.

Understanding these forces isn’t just an academic exercise, it’s vital for leaders who need to separate hype from reality. When you see the systemic drivers behind AI’s rise, it becomes easier to believe in its staying power and to justify bold investment decisions. Companies that grasp these fundamentals will be better placed to commit early, build conviction, and position themselves to capture the value that AI will inevitably create.

Every revolution in technology has a tipping point where momentum becomes self-sustaining.  AI has already passed that threshold. Each cycle of the flywheel feeds the next, more adoption generates more data, which improves models; better models attract more capital, more capital unlocks more compute, talent, and innovation. The result is a compounding loop where progress accelerates, barriers to entry rise, and late movers find it increasingly difficult to catch up. What once looked like optional experiments will quickly become table stakes, defining who leads and who lags in every industry.

The AI boom is not a passing trend, it’s the next chapter of evolution after the digital transformation.  At Tyde Consulting, we’ve helped organisations through major technology shifts, from digital and cloud to today’s AI revolution. Please get in touch to discuss how we can help, www.tydeconsult.com