When 2025 kicked off a few short weeks ago, the economics of generative AI seemed to be on a pretty clear-cut path. Businesses would need to pay big for the expensive chips and massive compute power required to train and deploy effective AI models. Access to the most powerful AI use cases and innovation would be restricted to players with the deepest pockets.
And then came DeepSeek. On January 20th, a tiny, largely unheard-of band of Chinese engineers released a large language model that comes close to matching OpenAI’s performance at a fraction of the cost. OpenAI has around 4,500 employees and has raised $6.6 billion in capital. They’ve been at it for a decade. DeepSeek was founded less than two years ago, has 200 employees, and built its model for under $5 million.
To be sure, DeepSeek achieved this milestone by standing on the shoulders of giants. Its R1 model relies on open-source technology that Meta and others have spent years developing. And OpenAI is investigating whether DeepSeek used its LLM to train its own system. But the DeepSeek team also reimagined how the technology could work and questioned basic assumptions that the Americans took as gospel. For instance, where OpenAI and others required their models to complete calculations to 32 decimal places, DeepSeek decided that eight was enough. And since the U.S. State Department bars Nvidia from selling its top-of-the-line H100 chips to Chinese companies, DeepSeek figured out how to make do with GPUs that were better suited for use in gaming consoles.
If you’re involved in the tech sector, the AI playing field suddenly looks much more level, with profound implications for consumers, entrepreneurs, and businesses.
And even if you’re not in tech, DeepSeek’s launch is a stark reminder of the tremendously uncertain and unpredictable times we’re living through. Donald Trump wants to turn Gaza into an American-owned resort. Elon Musk is trying to fire half the government. In the next year, we could begin World War III or a new era of peace. No one knows if the economic good times will roll on in the coming years or if we’re heading for a crash. Will climate change bring societies to their knees or spark a vibrant new economy founded on green innovation?
This level of uncertainty can make leading a business an unenviable task. And it can make the game feel unwinnable. How can you plan for a future five or seven years out when you can’t predict what the world will look like in just 12 months? The good news is that you don’t have to.
Shell’s Scenarios
The first step to preparing for the future is to stop trying to predict it. The world is too complex, too interconnected, and too volatile. But that doesn’t mean you can’t plan for what might happen next. Truly future-focused leaders envision multiple possible scenarios and then figure out how they might thrive no matter which future comes to pass.
That approach to scenario planning was pioneered by Pierre Wack and Peter Schwartz back in the 1960s at Royal Dutch Shell. At the time, Shell was seen as a bit of an also-ran in the oil industry compared to giants like Exxon and BP. But Shell took a different approach to how it thought about the future. Instead of using traditional methods to predict oil prices, the Shell team began mapping out different potential scenarios based on emerging industry, political, and social trends.
They looked around the world and saw that countries were moving to exert greater control over their natural resources. Libya and Venezuela were moving toward nationalizing their oil industries. In a three-hour planning session, Wack outlined six scenarios, drawn as a river forking into two streams, which in turn divided into three tributaries. These versions of the future depicted the growing potential of a radical power shift in the energy world, with OPEC nations moving to work together for their mutual interest.
The Shell team asked, “What would their world look like if Arab countries banded together to coordinate oil production and prices?” At the time, this seemed an unlikely event. Western countries and oil companies had successfully played Arab nations against each other for decades. And yet, the dire consequences for Shell of this scenario forced them to develop alternative actions, such as making investments in offshore oil drilling in the North Sea. And when OPEC came together to flex its muscles in 1973, Shell was the only player with a plan in place. Shell went on to leapfrog its competitors, becoming one of the dominant players in global energy.
The business landscape today is significantly more prone to shocks and disruptions than it was in the early 1970s. DeepSeek’s arrival is a potential OPEC moment for Nvidia, which lost nearly $600 billion in market cap the following day on fears its advanced chips may not be as essential for the AI boom as previously thought. Did Nvidia map out scenarios for this potential future and how it would continue to thrive in it? We’ll find out.
Planning For Uncertainty
The future doesn’t arrive unannounced—it’s already here if you look closely enough. Shell was able to see the future today as it surveyed what was happening in Libya and Venezuela and imagined a similar trend in the Middle East. Likewise, there are clear major trends in effect today that are shaping tomorrow’s world.
Some of these macro trends point to future certainties: The global population is aging, the planet’s climate is changing, AI is getting cheaper, and workforces are getting more diverse. These are such clear harbingers of the future that it makes sense to incorporate them into your planning no matter what happens.
But other factors are more uncertain: they look more like the question of Arab countries organizing. Will we enter an era of economic growth or stagnation? Will AI regulation stifle innovation or unleash it? The list of uncertainties facing any business is long. It’s therefore important to prioritize the most critical ones: factors that would have a major impact on your business if they came to pass. If you’re an energy company, it doesn’t matter who wins the World Series next year, but it very much matters who the president of Iran will be.
Envisioning Futures
Of course, a list of critical uncertainties often isn’t enough information to help senior leaders make decisions. Pierre Wack and his team quickly learned that it was helpful to translate these forces into vivid future “worlds.”
Nvidia execs seem to be executing a strategy that assumes massive compute centers operated by a small number of tech giants. That world may indeed come to pass. But Nvidia might also want to imagine a world in a few years where software advances make artificial intelligence cheap and easy enough to run on inexpensive local machines. Similarly, there may be a world in five or seven years where we’ve achieved artificial general intelligence: where computers are as smart or smarter than any human. But the opposite might also be true. Seven years from now, we might have discovered that LLMs are great for writing our kids’ term papers, but real human-like intelligence has proved to still be far outside our grasp.
What would each of those worlds look and feel like? Most of us struggle to imagine different futures. To help present-focused execs better understand those possibilities, the Shell team invested in rich storytelling, headlines, and posters that brought each world to life. Today, my colleagues at Jump have built on that methodology, creating different rooms for leaders to literally step into the future. The more vivid and visceral the experience, the easier it is for folks to imagine what might happen.
Taking Action
Of course, it’s not enough to imagine what the future might look like. You have to decide what you want to do about it. To improve decision-making, it can be incredibly useful to treat each scenario as a war game. Leaders are asked to divide into teams and spend a half hour in each world. They’re asked to imagine what that world looks and feels like. Then they’re asked to imagine which companies might naturally thrive in this future world. And which companies might fail. In doing so, they uncover the challenges and opportunities that each world presents. Finally, teams are asked to come up with specific strategies and investments that they wish they had made if they learned that this world had come to pass.
Some execs fight the process. They waste time arguing about why a particular world might never actually happen. The best teams take the possibility seriously. Like Shell, they take the unlikely seriously and figure out a course of action that would allow them to succeed no matter what.
That sort of war-gaming often highlights actions that a company should add to its growth portfolio. If you believe the future could go in multiple directions, invest in options that keep you competitive in any of them. Shell didn’t stop investing in the Middle East. They just added North Sea exploration.
Inevitably though, it turns out that leaders find themselves taking many of the same actions in every world. These “no regrets moves” are useful in any case. For instance, investing in renewables is a good move whether climate policy accelerates or stagnates. Even if we woke up in five years to find that global warming was a hoax (a highly unlikely outcome), it would have still made sense to invest in electric vehicles: they drive better, pollute less, and have lower maintenance costs than traditional cars. For Nvidia, it makes sense to continue driving AI efficiency no matter what world comes to pass.
The future remains uncertain. But that doesn’t mean you can’t prepare for it. As 2025 revs up, the amount of disruption we’re seeing can be bewildering. Don’t get distracted. Stay focused on the future. Because it will be here before you know it.