An asymmetric bet is one where the possible loss is small and capped while the possible gain is large, so a business can afford many of them and needs only a few to pay off. The idea is close to what Nassim Nicholas Taleb calls convexity, and to Jeff Bezos’s split between reversible “two-way door” decisions and irreversible ones.
Decide what you can afford to lose on a bet before you look at the upside.
02
Check reversibility
Bezos describes reversible Type 2 decisions as two-way doors that can be made quickly by small groups.
03
Look for open-ended upside
The best bets have gains that can grow far beyond the stake, not a fixed return.
04
Expect to be wrong often
A portfolio of asymmetric bets will have many failures; judge it by the total, not each bet.
Where you see it today
Cheap AI prototypes
Coding assistants and model APIs let a small team build a working prototype in days. That lowers the cost of each experiment, so a company can test more ideas with the same budget.
Amazon's playbook
Jeff Bezos’s letter to shareholders for 2015 argued that a bet with a ten percent chance of a 100 times payoff is worth taking every time, while accepting being wrong nine times out of ten (Bezos, 2016). The same letter set out the two-way door test.
When bets are not asymmetric
Large AI infrastructure commitments have big, fixed downsides. The IEA reports that global investment in data centres nearly doubled since 2022 and reached half a trillion dollars in 2024 (IEA, 2025). Those are one-way doors and need the slower, Type 1 process.
Convexity as a test
Taleb and Douady define fragility and antifragility through how a position responds to volatility (Taleb and Douady, 2012). A bet is convex when uncertainty helps it more than it hurts, a useful check before calling a project asymmetric.
Asymmetric bets vs related ideas
Asymmetric bet
Core questionIs the loss capped and the gain open?
ExampleA low-cost pilot of an AI agent in one team
Write down the maximum loss of each bet in money and time before approving it.
Label each decision as a one-way or two-way door and route it to the right process.
Run several small AI pilots in parallel instead of one large programme.
Set clear kill criteria so failed bets end quickly.
Move budget to the few bets that show results, and do it fast.
The Business Engineer lens
AI cuts the cost of trying things, so the number of affordable asymmetric bets inside a company rises sharply, while the infrastructure layer demands the opposite discipline. The Business Engineer separates cheap, reversible experiments at the application layer from heavy, irreversible commitments below it.
A bet where the possible loss is small and limited and the possible gain is much larger, so a few wins can pay for many losses.
What is a two-way door decision?
Jeff Bezos's term for a reversible decision; if it turns out badly, you can go back, so it can be made quickly.
How is an asymmetric bet related to convexity?
Both describe positions that gain more from good outcomes than they lose from bad ones; Taleb calls such exposure convex or antifragile.
How does AI change asymmetric betting?
It lowers the cost of prototypes and experiments, so companies can afford more small bets, while large infrastructure spending remains a high-stakes commitment.
The complete explainer, cleaned up and reformatted for mobile.
Another dimension of asymmetric betting is given by how impactful the idea can be to the business. When we have asymmetric bets that can have a high impact and are easy to reverse, we get to the “Jackpot” and go into an “All-In-Mode” of action! And how easy to reverse.
Introducing Asymmetric betting
There was a time when the motto for Facebook was “move fast and break things.” This was the foundation of the “Hacker’s mindset.”
Over the years, as Facebook scaled up, its motto also changed. And it went from “move fast and break things” to “move fast with stable infrastructure — as explored in the economics of AI compute infrastructure — .”
In short, when a company has scaled at a certain size, it loses the ability to experiment fast because almost all the experiments it undertakes are hardly reversible.
For a company like Facebook, even a small experiment targeting a few thousand users will be under journalists’ radar, amplifying the experiment if it goes wrong.
That is why – I argue – scale, to a certain extent, reduces the ability of companies to experiment.
Yet, if you’re just building a company, and your scale is limited, I think you have a valuable opportunity. As you can undertake, I like to call “asymmetric bets.”
Asymmetric bets are experiments, which are quick to implement, while they carry a very low downside, a potentially high/or unlimited upside, and they are reversible!
These, to me, are the goldmine for business people, and I’ve been obsessed with this concept for years.
Indeed, my main argument for why small startups can undertake large players it’s that small startups have in their favor these asymmetric bets.
As we saw, the case of large, scaled-up companies (like Facebook) is very hard to take advantage of because those large companies are under the radar, and any experiment they will undertake will be overseen by millions of people.
But how do you find these asymmetric bets? Let’s start by defining them!
Speed vs. Reversibility
In the speed vs. reversibility matrix, I defined four main decision-making models of action.
By looking at two main variables.
On the one hand, assessing an experiment’s risk is simply based on the worst-case-scenario test. In short, the question there is, what happens if the experiment goes wrong?
In a scenario where you know that if the experiment goes wrong, it will have a limited downside on the business, then the risk is low.
In the opposite scenario, where the worst-case scenario test can have a large negative impact on the business, the risk is high.
On the other hand, we want to ask whether the experiment is reversible.
In short, can you go back to how things were before the experiment? If the answer is “yes,” then you have a reversible experiment. If the answer is “no,” you have a non-reversible experiment.
Now we can define where asymmetric bets sit on top of this matrix!
Asymmetric bets: the sweet spot between speed, reversibility, and potential impact.
Understanding Asymmetric Betting
Asymmetric bets live at the intersection of low-risk, highly reversible experiments with a potentially unlimited upside.
But if these asymmetric bets are so powerful, why doesn’t everyone uses them?
1. Smart business people use asymmetric bets in combination with bolder bets to lower entrepreneurial risk. In short, asymmetric bets serve as a way to balance out other risks you might have to undertake as an entrepreneur. 2. Asymmetric bets require a specific mindset. You need to be wired to understand problems in the real-world (rather than looking for problems for your solutions). You need to understand that the world of perception can be more powerful than the engineering side. In short, things like UX can impact your bottom line as hard code changes the core functionalities of the product. 3. When you find these asymmetric bets, others might start to copy and paste them, thus making them less effective over time. In short, asymmetric bets follow the law of diminishing returns. 4. Those asymmetric bets might be highly contextual. They might work in one domain, not in another, so testing is critical. 5. Asymmetric bets work depending on the scale of the business. Certain asymmetric bets work until a certain size of the company. They kind of stop being effective.
With that in mind, let’s see some of these asymmetric bets in action!
Unlocking a fast-decision-making mode of action!
With that in mind, let’s see some of these asymmetric bets in action!
But before, let’s give some priorities.
Within the asymmetric betting model, we want to look for the “Jackpot” or those asymmetric bets that will unlock the “All-In-Mode!” To scout them, we’ll use two more dimensions:
1. How impactful can it be? We have either low or high-impact ideas.
2. How easy to reverse? In this case, we have ideas and experiments which are hard to reverse vs. the ones that are much easier to reverse.
Based on that, we try to give maximum priority to the “pure asymmetric bets.”
All-in-mode!
Key Highlihgts
Historical Context: The concept of asymmetric betting is discussed in the context of the evolution of companies. As companies grow and scale, their ability to experiment rapidly decreases because the consequences of experiments become more significant and less reversible.
Asymmetric Bets: Asymmetric bets are defined as experiments or actions that meet specific criteria:
They are quick to implement.
They carry a low downside risk.
They have the potential for a high or unlimited upside.
They are easily reversible.
Advantage for Startups: Small startups have an advantage in pursuing asymmetric bets because they are not under the same level of scrutiny as large, scaled-up companies like Facebook. Startups can undertake experiments with relatively low risk and potentially high rewards.
Speed vs. Reversibility Matrix: The decision to pursue an experiment is evaluated based on two key factors:
Risk assessment: What is the worst-case scenario if the experiment fails?
Reversibility: Can you go back to the previous state if the experiment doesn’t work?
Sweet Spot of Asymmetric Bets: Asymmetric bets are positioned at the intersection of low-risk, highly reversible experiments with the potential for significant impact. They are considered the “sweet spot” for business opportunities.
Why Not Everyone Uses Them: While asymmetric bets can be powerful, not everyone uses them for several reasons:
They require a specific mindset and problem-solving approach.
Asymmetric bets can lose effectiveness as others start to copy them.
Their effectiveness may be highly contextual.
They may work for a company up to a certain size and then become less effective.
Unlocking the “Jackpot”: To identify the most valuable asymmetric bets, two additional dimensions are considered:
Impactfulness: How much of an impact can the experiment have?
Ease of Reversal: How easily can you reverse the experiment if needed?
All-In Mode: When a high-impact, easily reversible asymmetric bet is identified, it can unlock the “All-In Mode” of action, where the company dedicates significant resources and efforts to pursue that opportunity.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
