Decade after decade of widely available computing power has caused significant industry disruption and an unprecedented technology boom—and transformed the global economy in the process.
High-frequency trading firm Citadel Securities, for example, began to incorporate advanced computing capabilities into its core strategy in the early 2000s, and in turn was able to harness endless processing power and technologies such as machine learning. , to execute trades in microseconds. Citadel’s ability to rapidly deploy large amounts of computing power has made it the largest retail market-making firm in the US, executing over $400 billion in trades per day.
The innovations enabled by unlimited computing power have only expanded with the development of AI. However, for many, there seems to be a looming paradox, based on the prediction that because AI frontier models are so computationally intensive to train and use, their proliferation could lead to a shortage of computing power.
To test the likelihood of this scenario, the BCG Henderson Institute and Exponential View teamed up to quantitatively model future supply and demand for computing power. Our model has incorporated moderate assumptions about future supply, using incremental estimates of demand. What we found was that the abundance of computing power is likely to persist even in an aggressive scenario of rapid generative AI adoption.
While tech CEOs have received this message and are preparing to take advantage of the continued availability of affordable computing power, many business leaders have not fully digested the potential scope of this reality in relation to the transformative possibilities of AI – even though it can have an impact on how companies of all stripes are organized and run. Viewing change as incremental, rather than exponential, is a common mistake that Nathan Myhrvold, Microsoft’s first chief technology officer, noted back in 1993, writing at the time that most executives “act as extrapolators linemen who won’t know what hit them.”
For executives seeking to avoid being blindsided by an exponential wave of change brought about by the confluence of AI with enhanced computing power, now is the time to reflect on the implications. In what new ways can leaders create, deliver and capture value enabled by more massive computing power? Having an answer to this question as AI and abundant computing power continue to feed off each other will be key to staying at the forefront of innovation and competition.
Computing power will continue to be abundant
How likely is computing power to become scarce in the era of increasingly computationally intensive generative models? Will the world literally run out of the high-end semiconductors needed to support GenAI workloads? To examine this question, our quantitative model estimated global computing power including a scenario of incremental demand for GenAI completion, where models continue to get larger over time and enterprise adoption of the technology is very aggressive by historical standards. (We focus our model on inference because in a world of widespread GenAI adoption, the computation used to train a model will dwarf over time the overall computation spent using models.)
What we find is that even in the increasing demand scenario, the total computing power consumed to complete GenAI will still only take up a third of the total expected supply of computing power by 2028. In other words, we believe that the regime of available and affordable computing power that has fueled the digital economy will continue to hold even as GenAI adoption accelerates.
This is not to say that there aren’t real challenges to a sustainable expansion of the computing power supply. Hardware availability may be disrupted by geopolitical tensions affecting the semiconductor supply chain, while power supply may limit the expansion of data center capacity in some geographies. But advances in energy efficiency (through, e.g., new, specialized equipment), as well as efforts to make data centers self-sustaining through on-site power generation, should help cope. of this energy challenge.
Even with these risks, the main trend points to abundant computing resources in the coming years with the power to drive increasingly powerful innovations as AI capabilities continue to expand and deepen.
How businesses can harness computing power and advanced AI
There is no doubt that computing power can be a source of competitive advantage for businesses that are quick to exploit it. We have already mentioned the case of Citadel in financial services, but there are many examples in other industries as well. Amadeus, for example, originally a traditional European travel booking system, adopted open source systems and cloud computing, allowing it to process more than 100,000 transactions per second, and became a global travel technology leader that supports airlines and travel agencies worldwide.
Artificial intelligence makes computing power even more powerful, as it can be used to speed up or even automate increasingly complex problems. As we’ve argued in this column before, AI promises to make much of the current human cognitive activity in the workplace mobile to machines, in the same way that old factory floors are now often occupied by robots, in instead of human workers. Moderna, for example, has been able to disrupt the pharmaceutical industry, speeding up vaccine development from five-plus years to just months—all thanks to the automation of complex scientific processes with computing power and AI.
These cases illustrate how much can be gained by being prepared to deploy computing power, especially in the age of AI. CEOs would be wise not to discount the impact of computing power on creating or maintaining a competitive advantage—lest they risk being thwarted by forward-thinking competitors.
From Constraints to Opportunities: How Imagination Drives the Next Wave of Innovation
As computing power continues to expand alongside the relentless advancement of AI, it is imagination that will unlock new possibilities from AI and other advanced technologies. Consider other historical developments, such as the shift from dial-up/ISDN to DSL, then cable, fiber optic, and now satellite-based Internet (eg Starlink), enabling global connectivity. In many cases, the technological capabilities for such a leap already exist, but future advances may amplify them, driving innovation beyond current expectations.
In anticipation of orders of magnitude more calculations in the near future, business leaders should ask themselves what are the highest-complexity, highest-value processes and activities that have resisted automation due to complexity or their computational cost. As our colleagues at Exponential View said, “What would you do with 1000x more computing power? How would your organization use it? If you asked these questions to Sam Altman, Satya Nadella or Sundar Pichai, they would have an answer. Do you?”
The process of formulating a response to this challenge may be more systematic than many leaders realize. First, it requires an in-depth examination of a company’s business model, starting with its value proposition. Computing is the key to reaching new heights in personalization and customization of goods and services, on the way to a true “customer segment of one”.
Next, leaders must take a hard look at their value creation and delivery. What are the seemingly strange possibilities that become more plausible with sufficient calculation? Imagine an insurance company able to develop a digital twin of the Earth to better simulate and predict atmospheric and geological hazards. This is a technological feat no longer out of reach with abundant computing power; it can transform an insurer’s approach to risk management, changing the current state of competition in the insurance market based on marginal improvements over traditional actuarial methods.
Finally, when it comes to capturing value, the combination of strong models and high-frequency, high-quality data creates a tremendous opportunity to optimize pricing models.
Of course, the (now) old mantra of getting your data in order still applies. But the horizon of expanding computing power harnessed through AI is largely that of the imagination. And, again, there is a way to systematically turn complex organizations into “imagination machines.” As Martin Reeves and Jack Fuller have argued in their book, The imagination machine, Business leaders can and should actively seek out surprises in order to rethink their mental models, and then test them against the real world. If the imagination is the ability to create a mental model of something that doesn’t yet exist, then nothing will be more important than imagination itself for success in an economy increasingly driven by computing power and reality-augmenting AI.
***
Read on other Fortune columns by François Candelon.
François Candelon is a partner at private equity firm Seven2 and former global director of the BCG Henderson Institute.
Azeem Azhar is the founder of Exponential View, an executive fellow at Harvard Business School and a technology investor.
Riccarda Joas is a consultant at BCG and a former ambassador at the BCG Henderson Institute.
Nathan Warren is a senior researcher at Exponential View.
David Zuluaga Martínez is a senior director at the BCG Henderson Institute.
Some of the companies mentioned in this column are former or current clients of the authors’ employers.