The AI Super Cycle – where is the wave headed?
ChatGPT’s blockbuster release with over 100 million users in its first two months, triggered a frenzied capital play in generative AI across the industry and governments, akin to the dot com of yore, triggering an AI super cycle. Déjà vu all over again!
Looking back we have had super cycles led by steam engine, electricity, computers, smartphones, economic super cycles like the industrialization of US, Europe and Japan followed by China, or the commercial use of oil, geopolitical super cycles of monarchy, capitalism, communism and the recent seeming revert into the ravages of cycles of war or even the environmental ones like greenhouse effect or climate change and the ramblings for saving our planet.
Is the AI super cycle beginning to wane?
The hype of buying anything with AI continued relentlessly with companies transforming into .ai avatars, the stock market bursting through the skies for chip makers led by NVidia with their AI powered GPUs, who leapfrogged to $26b revenue in a quarter from an annual $26b just a year ago. Almost all industries have their pie in AI with the big technology players, seemingly far ahead and others at various stages of their journey.
However, there are musings about all this being overhyped and whether investments are leading to predicted productivity and efficiency gains or whether acceleration in growth seen by some(NVidia) can be sustained?
To the pundits who set up expectations of quick growth and quicker results, the AI wave has thrown up challenges on multiple fronts and while it is on the path to delivering results, the timescale will be much, much longer. The associated cultural and social change involved has its own paradoxes as for the first time the white-collar knowledge worker is also feeling the need to reskill! Responsible AI for a future which avoids any conflicts between a digital and biological human is repeatedly being emphasized or imagined, though a far cry from where we are today!
Then how should we look at this super wave?
Like earlier, this hype too shall pass and industry or society should look at AI as a tool for achieving economies of innovation which can scale up and free up humans from the drudgery of cycles of repetitive work which the digital assistants can do better while humans can, not only control these AI avatars but with the democratization of innovation, also indulge in their passions with paybacks, hopefully enjoying more of life as well! Areas like creation of new LLMs currently need a massive compute power affordable by only a few which will limit the speed of further growth in this area. In fact, the 1.6 trillion parameters expected in GPT- 4o have not necessarily made it better than the earlier models. However, we now also have DeepSeek with similar performance but many orders of magnitude cheaper, that has triggered a fresh race to supremacy and suggests we might see costs coming down drastically in the years to come!
Nevertheless, things will evolve, quantum computers may yet become affordable in future leading to AI super clouds reaching all, like the electricity in our homes. Research will cease to be a prerogative of research labs and instead embrace people labs! AI along with generative AI will become the biggest accelerator to the convergence of technology areas like robotics, genome sequencing, energy storage, quantum computing, wearables, neuralink and public blockchains while also reducing complexity of use and deployment.
With people thus truly empowered, and other converging technology areas reaching the common man in future, a new age of super industrialization that guarantees equality of opportunities and ensures progress in all spheres from education to healthcare, to jobs, to access to capital like never before would soon be upon us. Global economic growth of ~7% annually, could possibly become a reality in less than a decade!!!