August 2024
Generative AI is following the classic hype cycle. After almost two years of unrealistic expectations and euphoria, a rising tide of disappointment is mounting because GenAI hasn’t transformed the world overnight.
Over the past few weeks, reports have said corporate spending hasn't delivered ROI, GenAI startups aren’t building sustainable businesses, and investors have driven early-stage valuations in this sector beyond reason.
While these reports are factually correct, this analysis frames the issue incorrectly. To GenAI skeptics, I want to send a clear message:
GenAI is not disruption. It’s discontinuity.
The distinction is more than academic. GenAI marks the start of a transformation far more profound and sweeping than even many sophisticated technology leaders have truly grasped. That means it will play out over a much longer time frame than investors and entrepreneurs are accustomed to tolerating.
This shift requires everyone across the innovation economy to fundamentally reset how they think about building new companies, making investments, and defining market shape. Faced with the unknown, too many people rely on gut instinct when making decisions.
Instead, this moment demands a long-term view to map out which investments and technologies make sense, when, and why. Investors and founders must develop a rigorous, multi-scale time frame for their analysis to know which questions to ask about emerging GenAI businesses, how to identify the most resilient startups, and how this technology will impact existing businesses across all markets.
Breaking from the past
Almost 15 years ago, I worked in the traditional financial analysis industry and grew increasingly frustrated. When evaluating innovative companies, everyone seemed stuck in silos, trying to understand innovation through the lens of previous generations of technologies and industries. I believed we needed a more holistic approach—one that started with a clear understanding of the business model (starting with the revenue model)—with subsequent financial analysis built on that foundation. Numbers always tell the truth, but they only matter when you surface the numbers that actually matter. New technologies make this complex, and it is easy to skip rigorous financial analysis on the pretext that no precedent exists.
I founded my firm, D’Ornano + Co, based on that conviction. Since then, we’ve built our practice around the philosophy that traditional tools and thinking are insufficient for analyzing disruptive technologies.
Over the past 18 months, a growing share of the deals we analyzed have involved GenAI, from infrastructure to horizontal and vertical applications. I've found this space fascinating.
I have noticed a pattern that's been hard to break. Expectations about growth and returns still follow historic patterns, and so does the timing: GenAI should already be transforming the earnings of those who adopt it; it is just overhyped. Really?
Sure, bets are being placed in areas of GenAI that are not mature enough to deliver impactful products. FOMO proves to be overwhelming. Many bad investments will be made in companies unable to build sustainable business models in the long run. Part of that is perfectly normal and linked to the nature of Venture Capital: funding innovation. Of course, adequate diligence will help avoid — we hope — some of those.
To show why this thinking is problematic, I want to pull back and focus on discontinuity versus disruption.
When disruption comes along, it disrupts the linear trend and changes the curve, increasing or decreasing it by, say, +10%, +20%, +50%, and so on. The curve becomes steeper or less steep, but it still rests on the same fundamentals that created the curve in the first place. That’s not to say navigating disruption is easy, but leaders and investors can use existing tools to analyze risks and opportunities and understand how the game has changed.
Discontinuity stops the curve and creates a new curve that has an entirely different profile. The curve can continue in the same direction and at the same speed, but it can also collapse or accelerate in unexpected ways. In geology, discontinuity designates structural breaks, usually unhealed. But when they heal, they exhibit high tensile strength. The same applies in business.
All the previous strategies, tools, and historical lessons are ripped away. In this dynamic, discontinuity breaks with the past and can lead to something very positive or to failure. But it’s philosophically and fundamentally different from disruption, which doesn't introduce such clear breakpoints.
Such moments pose a fundamental challenge beyond tech: people’s brains aren't wired for discontinuity.
Several years ago, John Kao, the Turing Fellow at Yale’s Center for Collaborative Arts and Media and a noted innovation thinker, wrote an essay discussing the dilemmas leaders face when confronted with the disruption versus discontinuity paradigm: “Optimizing for disruption and discontinuity requires different skill sets and mindsets. Addressing discontinuity requires creativity and the ability to bridge today to an uncertain future. It requires the ability to generate meaningful ideas that can serve as beacons for a journey into the unknown.”
Understanding and adapting to discontinuity requires understanding the underlying components driving this phenomenon, a clear vision of how it impacts an existing business model, and a sophisticated view of the structures and assets that realign or emerge to enable something new.
From what I have seen in the deals I’ve worked on and in conversations with top private equity investors, they still don't fully recognize the discontinuity. This is understandable, given that the asset class has largely been built on predictable, stable cash flows in its investable targets.
But today, they are missing the larger story of GenAI.
The GenAI Discontinuity
Though critics like to mock Silicon Valley when bubbles emerge, tech leaders have proven remarkably prescient about what will happen. They tend to get the timing wrong.
During the dot-com bubble, founders and investors rushed to create and fund companies based on the assumption that businesses and consumers would change their habits in a blink and the old economy would be washed away in a few months. US telecom companies raised billions in debt to build networks that lay unused in the ground, creating a glut of “dark fiber” across the country. When the bubble burst, the “New Economy” became a punchline.
Amid the rubble of the dot-com bust, many Silicon Valley VCs began discussing a book Science Historian David Nye wrote a decade earlier, “Electrifying America: Social Meanings of a New Technology.” The book traces the many ways in which electricity transformed the nation.
Cities installed electric streetlights, which expanded nightlife and led to a wave of new restaurants opening. Companies that furnished electricity needed a reason for people to use it, so they built streetcars. Then they needed more riders, so they built amusement parks at the ends of the lines. To increase demand for electricity in homes, they designed appliances such as refrigerators and washing machines. As residential buildings became electrified, people stayed up later to read and do other things.
No one could have envisioned most of these innovations when Thomas Edison made his first lightbulb in 1879. It took decades to build the infrastructure, raise investment capital, develop use cases, and refine business models.
VCs saw a parallel with the dot-com bust: the real transformation would unfold over a much longer timescale as all the pieces came together. Though it may have taken longer than initially predicted, we now live in the digital world many envisioned in 2000, thanks to smartphones, 4G wireless, broadband, and a host of business-model innovations enabled by these infrastructure waves.
This is where we are with GenAI. While we are not seeing a GenAI “bust,” the clouds forming around it are obscuring its discontinuity.
GenAI represents a fundamental break because its impact could extend far beyond tech disruptions like the cloud, smartphones, or the social Web. GenAI's scale will likely rival, and most likely exceed, the internet because its ability to leverage both structured and unstructured data removes previous limitations on what data can be used.
However, this technology is still in its infancy. The LLMs that enable GenAI are evolving rapidly, and many elements remain unstable. Open or Closed LLMs? Big or small models? Nobody can say for sure. In July 2024, Meta announced it had started working on Llama 4, which it expects will require almost 10 times the compute used to train Llama 3. Mistral has unprecedented traction. The race seems to be on. The capex surge from these developments must be weighed against the asymmetric risk of missing the next platform shift.
At the application level, software companies are still trying to understand what data they may have that could be tapped for GenAI, a process that could require huge investments to create the right data flows.
Meanwhile, the infrastructure needed for GenAI is only just being put into place. That includes a scramble to deliver the processing power and a boom in data center construction. This explosion in data centers has a secondary effect: surging energy consumption. Goldman Sachs projects that AI will increase data center power usage by 160% by 2030.
Given the uncertainty across this GenAI discontinuity, applications built on existing foundational models are shaky. Many adopting companies are still experimenting. So much still needs to be figured out, and making sweeping judgments about GenAI based on short-term ROI is premature.
Any meaningful impact will not become apparent for another 2 to 4 years. The real transformation will take even longer.
This is not a reason to stand on the sidelines and passively wait to see how things play out. Execs and investors need to ensure their business does not lose ground before reaping the benefits. GenAI can transform many non-tech companies into tech companies, an evolution many industries should start recognizing now. This is already happening at full speed in the B2B services space, a space Private Equity investors prefer.
Now is the time to take the first steps. Investors should start by acknowledging that their view of returns, timing, and approach structure no longer fits this new era.
Instead, investors must begin to grapple with the ways these different silos within the GenAI tech stack — data centers, real estate, chips, energy, software — are blending to create a new framework. The same applies to asset classes, with boundaries blurring among Venture Capitalists, Private Equity Investors, Asset Managers, and, of course, Hyperscalers.
Investors must learn to conceive of what has not yet happened and to imagine what will happen. Navigating discontinuity requires the ability to abstract, see how these pieces might eventually align, and imagine what consequences could be unleashed.
Discontinuity leaves investors in an uncomfortable position, with everything they’ve known lost. Therefore, very rigorous measurement is needed. When an investor values a startup at 100X versus 10X, they need to be 10 times more rigorous in their due diligence.
GenAI is not a mysterious black box. You can measure every element in granular detail. No investment is risk-free, but you can make smart, disciplined decisions by identifying GenAI's key business levers and measuring them to build confidence. A more strategic, informed approach will direct capital to the right businesses that are driving this discontinuity and transforming themselves in preparation for this huge opportunity. For more mature businesses, GenAI will enable unprecedented productivity gains, helping expand margins in a challenging macroeconomic context.
GenAI requires a fundamental change in investing and in how leaders should think about their businesses going forward. Only by decoding this discontinuity can we achieve higher returns and create long-term value.
How the discontinuity unfolded
The argument developed here is extended in the AGNT Manifesto:
From Disruption to Discontinuity — where the distinction is restated with the SaaSpocalypse as evidence
Six Tremors, One Fault Line — the six shifts that made the break
When Exponential Growth Becomes Visible — the winter the models crossed
Intelligence Is Sufficient for Production — why AGI is not the precondition