The race to achieve Artificial General Intelligence (AGI) and the pace of investment that is going along with it, have brought concerns of a bubble in AI to the forefront. The race is clearly not without risk; there will be winners and losers, uncertainty and volatility along the way. But ten years from now, it’s likely we will recognize that we underestimated AGI’s impact on society rather than overestimated it. And so, it is worth taking stock of where we are in the journey towards AGI.
Where We Are Today
Today, common AI tools perform research, generate text and images and can conduct basic tasks. AGI, on the other hand, will do all of that, plus have human-like decision-making capabilities and even execution in many cases, to create improved outcomes. Big Tech hyper-scalers are investing enormous amounts to build the infrastructure for AGI. In 2025 alone, Microsoft, Amazon, Alphabet and Meta are projected to spend close to $400 billion on AI infrastructure. But this spending is not all on programming and code. Data centers, which house the computing power necessary for AGI, are comprised of chips, racks, cooling systems, transformers, pipes, sprinklers, cabling, concrete and many other “real-world” items. So AGI’s economic impact is already large, and yet the biggest impact may still lie ahead.
The potential economic upside to achieving AGI is considerable. From accelerated innovation, greater efficiency and faster workflows throughout the economy, Goldman Sachs estimates AI could boost global GDP by $7 trillion over the next decade, while McKinsey projects annual gains of up to $25 trillion if AGI becomes reality. Imagine an economy where AGI systems tirelessly design new drugs, manage global supply chains and optimize energy grids, ridding the economic system of waste and suboptimal outcomes when it comes to tasks. This shift could lower costs, accelerate innovation, harness engineering breakthroughs, all of which could create newfound puffs of wind on the global economic flywheel.
Under this scenario, for tech giants, being “the one” or “one of the ones” to achieve AGI is akin to owning the operating system of the future. Or some even suggest it could be like owning the Internet itself. Each layer of the stack required for AGI, from chips and cloud platforms to foundational models, offers annuity-like economics, something Big Tech covets. Conversely, failure could mean existential risk: losing relevance in a world where AGI redefines productivity, shatters barriers to entry and impacts competitive advantage.
Risks and Realities
A common question today is whether all this investment has driven certain valuations to the point where there are risks of overheating, fostering concerns that returns on investment will not materialize as forecast. But clearly, as evidenced by the massive capital outlays, the bigger risk in the eyes of tech companies and countries, for that matter, is being left behind. In addition, it’s important to point out that there are first-mover advantages because smarter AI models can produce the next smarter model and so forth. Tech lives on a “winner-takes-all” mentality. There is a positive feedback loop for those who are ahead, while those who fall behind will experience a negative feedback loop as their models become less relevant. This is the urgency, and these battles that will determine winners and losers are being fought literally as we speak.
What shouldn’t be forgotten, however, is that there will also be companies, large medium and small, that never enter the AI core infrastructure arms race, but rather will simply pick up the newfound tools and find ways to improve their existing businesses or start new “full AI stack” companies that disrupt the status quo based on speed, cost or customer experience.
An additional risk, or opportunity, depending on your perspective, is that AGI will impact aspects of the labor market. History is littered, however, with examples where new technologies were going to eliminate or impact jobs, yet, with all the innovation that has been unleashed on the U.S. economy through the many decades, we have the highest number of aggregate jobs in our nation’s storied history as we stand here today at roughly 163 million, which is up from roughly 61 million in the 1950s. Yet think of all the disruptive technology that was created along the way.
As more productive AI tools become available and eventually AGI itself or even humanoid robots appear, it will be important for workers to prepare their skills for jobs that emphasize AI management, creativity, relationship building and strategic thinking. All of which are skills that complement AGI deployment rather than compete with it. History suggests that technological revolutions ultimately expand economic opportunity, but the short-term disruption in some fields could be significant. There’s a saying that the definition of minor surgery is surgery somebody else is having. The same could be said for AGI and its potential disruption of aspects of the labor market.
What Won’t Change?
This is perhaps the most interesting and enduring question. What won’t change with AGI? As Morgan Housel explores in his book, “Same as Ever: A Guide to What Never Changes”, certain truths will likely hold their value through any period of change. Our take on that question is that things like diversification, relationships, discipline, patience, quality, trustworthiness, creativity, manners, common sense and good judgement are all likely to endure the AGI revolution. In closing, the buildout of the AGI-capable tech stack is more than the next release of a software update or a new phone; it’s widely considered the foundation for the next economic era. For investors, policymakers and each other, however, it is a time for awareness, diligence and yes, optimism, but perhaps with that, a renewed focus on the soft skills that aren’t likely to change in the future or any time soon.