
Direct Competition: The Only Moat for Startups to Stand Up to Giants
By: Not Boring
Translated by: TechFlow
TechFlow Summary: This article breaks down the only moat a startup can use to directly take on giants: counterpositioning. It is not about doing things better; it is about making it impossible for competitors to copy you. For any entrepreneur trying to carve out space under the shadow of a giant, this is the most practical strategic weapon.
Hello everyone 👋,
Happy Thursday! We launched the new version of the notboring.com website this week, marking our second piece for the week. We are in a deep-work sprint season now.
This is a continuation of the strategy matters, moats are critical series. There is still a common belief that startups do not need these things—they just iterate faster on their products and do them better. It is overly harsh to demand network effects or economies of scale from such young companies.
But startups should not lie flat strategically just because they are young. Fortunately, there is one moat that buys them time. This article is a hymn to it.
Let’s dive in.

There are seven powers, but if I had to pick my favorite, it would be counterpositioning.
In Mind the Moat, an introductory guide to Hamilton Helmer’s 7 Powers, Lindy CEO Flo Crivello defines counterpositioning as: "build your business model so that the incumbent cannot effectively compete due to conflicting incentives."
This definition captures why I love it so much: before a startup grows large enough to attain the other six powers, it uses this as a moat to attack the incumbent. The incumbent has a larger business and more resources, and the startup leverages those exact advantages against them. The incumbent does too well to move recklessly.
The first time I wrote about Ramp in December 2020, I compared counterpositioning to the five-strike piercing palm from Kill Bill. When a startup lands a counterpositioning blow on an incumbent, the latter might not even realize it is already dead; but take one step toward the startup, and it collapses.

If the incumbent’s corporate card encourages customers to spend more via points, then Ramp’s business model helps customers spend less. Incumbents cannot follow suit for many reasons, including that their software capabilities lag far behind Ramp’s, but primarily because if customers spend less, they earn less—and shareholders will not stand for that. Other startups like Divvy and Brex play the same game as the incumbents, using rewards and points to stimulate spending. They ended up fine: Bill.com acquired Divvy for $2.5 billion, and Capital One acquired Brex for $5.15 billion. Today, Ramp is valued at $44 billion and rolling out a suite of products designed to save customers money and time, including a model router.
If you want to challenge a king, do not try to beat him at his own game with better technology.
I thought of counterpositioning today because of CEO Zach Dell of Base Power Company’s conversation with host David Senra on David Senra’s podcast. At the 16:18 mark, when discussing how he and Justin landed the business model, he delivered a sermon on counterpositioning.
"If you are going to challenge an incumbent," he said, "my thesis is that the best way to do it is to have a counterpositioned business model."
If I show up saying, "I make the best home battery on the market, and I sell it 10% cheaper than the big guys," that margin is my opportunity, right? Those big companies will fight price wars. They will copy my product, slash prices, and eventually push me out of the market. But if I show up saying, "I do not sell batteries, I sell electricity"—we will install batteries at your house, and because of our different business model, you only pay one-twentieth or even one-fortieth of what it costs to own one outright—then I win. If you are one of the incumbents wanting to compete with me, you would have to completely overhaul your business model, and you know how hard that is for a public company, right?
Base cannot rely on counterpositioning forever. The moat protecting its margins will come from other powers: economies of scale, control of a key resource, switching costs, brand, and possibly network effects and process capability. But as Zach pointed out, there is no better way to start challenging an incumbent than with counterpositioning.
This is a theme Ben and David repeatedly hammer on the Acquired podcast: "As Hamilton Helmer says, counterpositioning is usually the power of takeoff." Helmer himself appeared on the show to explain why: "It is the only incomplete source of power. If you want true durability, you need another source of power."
Counterpositioning is here to keep you company for a while, not to stay with you forever. And that is exactly what makes it interesting. It is the power of the trickster god. Counterpositioning buys you time to fight the incumbent (those incumbents who happen to be positioned in a way that leaves them vulnerable to a counterpositioning strike), but it does not necessarily protect you from being attacked by another startup. You will need other moats for that.
Happily, Helmer uses Dell to illustrate this point. Dell counterpositioned Compaq because Compaq’s dealer channel made direct sales painful. But eventually, anyone could adopt a direct-sales model, and Dell’s counterpositioning power evaporated. By then, Dell had already used that window to build economies of scale around its direct-to-consumer, JIT model.
The classic case studies in Acquired are full of examples like this.
When Google launched Android, it could give it away for free because its parent company made money on search. It could have subsidized all the losses, but practically speaking, it did not need to—the more mobile searches happened on Google, the more revenue it generated. Meanwhile, Nokia had to make money directly from the mobile OS; giving it away would have killed its business.
Here is a good way to think about counterpositioning: Can Challenger A do X precisely because it makes money in a fundamentally different way than Incumbent B?
To be clear, having Google’s money-printing machine behind you certainly makes it easier, but it is always worth thinking about how to disrupt the way your incumbent competitors make money.
Amazon was born on the internet and built around a direct-to-consumer fulfillment system. Barnes & Noble, on the other hand, was optimized around physical stores. For Barnes & Noble, fully committing to Amazon’s model would mean reallocating capital, executive attention, and distribution infrastructure, all while enduring short-term profit dips at existing stores. Amazon’s model "is lower margin for them… compared to those high-margin brick-and-mortar stores," Ben said. Amazon "counterpositioned every competitor whose cost structure was built for physical retail," David added.
Another angle to consider: the more an incumbent has invested in its existing infrastructure, the harder it is to adapt.
Startups without Google’s funding can absolutely deploy this strategy too. My favorite example is Somos Internet. As I explained in Cable Caballero, Somos designed a completely new network architecture and built its own hardware to run it. Incumbent telecoms could not respond because doing so would require writing off billions in capex spent buying and upgrading third-party hardware under legacy configurations. Their dilemma was obvious. Add in vendor dependence, and most firms simply lacked the technical capability or flexibility to rearchitect the entire system. Worse, they had leveraged themselves heavily to finance network rollouts, leaving them unable to cut prices to crush Somos, let alone fund an uncertain new network build-out.
Those two points have another side. Anyone who has seen a movie where villains kidnap the hero’s family to hold them hostage knows the trope well: having nothing to protect is, in itself, a form of freedom.
When Microsoft challenged the company known for "no one gets fired for buying IBM," Ben explained, "Microsoft basically had nothing to lose." IBM sold integrated computer systems through the strongest enterprise sales team in the industry (software, hardware, everything). To rush a PC to market, the Big Blue broke its usual mold, using off-the-shelf components, Intel processors, and even adopting the OS supplied by Microsoft. But Microsoft did not grant an exclusive license to IBM; it licensed DOS non-exclusively and had a strong incentive to make PCs as cheap and widespread as possible to put its software on as many desktops as possible. As Ben and David put it, they could say: "We do not need to make money on hardware. We do not even need to make hardware."
Microsoft did not win because its software was better written than IBM’s. IBM’s software was fine in the end. But IBM needed software to sell high-margin machines, whereas Microsoft profited when the hardware layer became commoditized. Once hardware was commoditized, all hardware providers had to support the software layer. Or, as Ben put it, Microsoft could "just ship bits and freely become the integration core of the entire ecosystem."
This is a vicious form of counterpositioning: the entrant can counterposition not only because it is willing to cannibalize a profit pool, but because it actively benefits from destroying the incumbent’s profit pool.
Then there is Facebook. It launched in 2004 facing off against MySpace and its one million users. MySpace is a joke now, but back then it was skyrocketing. It went live in 2003, hit one million users in February 2004, surpassed Friendster the following month, and grew fivefold to five million users by November.
Facebook’s counterpositioning was to argue that stuffing a massive user base into a network from day one is not actually ideal. A better approach is to start small, seed the network with Harvard’s elite, and then expand slowly within tight circles.

What would you do if you were MySpace? Growth is good, and these nerds are just building a little thing for their nerd friends. Would you stop growing?
But it turned out starting small was better: seeding a network with Harvard’s elite and expanding slowly within cliques. Obviously not forever, just initially. Facebook leveraged that counterpositioning window to build a highly passionate network where people chatted with new college friends and poked each other. The product attracted outsiders—regardless of age—so intensely that when they finally opened the floodgates, the floodwaters rushed in, propelling them into the realm of network effects.
People have not quite learned that lesson yet.
Earlier this week, a16z partner Josh Elman tweeted, "New moats are just like old moats. Every few years, we fall in love with shiny new technologies and forget the basic laws of physics for consumer software." I would extend that: what we forget are the basic laws of physics for business strategy. But since we are on Facebook, let us stick to consumer.
AI agents are the hottest product category right now, and Instinct is the most hyped among them. The experience is fantastic, surpassing previous assistant products.
But as Ben said: "Better is not counterpositioning." Strategy master Michael Porter wrote, "Operational effectiveness means performing similar activities better than rivals. Competitive strategy is about being different."
"Better" is an advantage; moats require barriers. Counterpositioning only exists when an incumbent sees you are better but cannot replicate whatever makes you better without harming its existing business.
(For clarity: you cannot choose a model so terrible that pursuing it harms everyone, including yourself.)
Who gets hurt by launching a useful agent? Labs have already done it in desktop apps, Grok Bot did it, and other startups are doing it (and by definition, they cannot be counterpositioned by startups), but Instinct is still better. Because it is better, it raised a $350 million Series B at a $2.5 billion valuation in late August. One theory is they are raising that much cash to keep offering the service for free, drowning competitors on the path to a new business model like transaction fees.
But better is not a moat. Good luck trying to drown Meta’s money printer.
Meta (formerly Facebook) also released its own assistant this week. It is called Muse.

Some early reviews claim it is better than Instinct. I have not tried it yet; better is not the point.
The point is Meta replicating Instinct’s model and product is not just painless—it is brilliant! Meta knows you better than anyone else can, recommending things you did not even know you needed, solving the cold start problem. As Ben Thompson writes, they can give every user "a virtual machine with 8GB RAM and 8GB storage," which is "a real computer, and Meta is providing one to everyone in America, and eventually to everyone in the world. That is pretty incredible!" They can integrate directly into WhatsApp (that massive but under-monetized messaging app), plug into FB Marketplace so you can buy what you want cheaply, embed it in Meta Ray-Bans so you can literally tell your face what to do, and train it in their massive data centers. They can offer it for free until the heat death of the universe (or until that >10% existential risk probability from Anthropic materializes). They can marshal network effects, economies of scale, brand, and any necessary moat because they counterpositioned early and dug all these moats first.
Of course, Zuckerberg has tried copying or creating new products before, failed, or has not managed to kill the target. Threads did not cure my Twitter addiction. But Twitter has network effects. It has a moat.
As Josh noted, too many startups believe new technology is so wondrous that they no longer need moats. They think they just need speed and taste, focus and agility, plus some indefinable startup magic.
A lot of startups do just fine using this approach. Large companies with moats are happy to pay big bucks to bring that indefinable element into their fortress rather than letting it land in a competitor’s hands.
I cannot point to a specific moat for Cursor, but it is an outstanding product and worth $60 billion to SpaceX. Nvidia acquired Hugging Face for $12.9303 billion, which I conveniently used as an example in When to Dig a Moat to show how an AI company leverages a window of complexity and uncertainty to develop early network effects. I recently chatted with a founder who, when asked about moats, said something like, "We do not care about moats, we just want to build great products." So I dropped it. Now that company is valued roughly 50 times higher, so what do I know?
Rumor has it Meta (and others) even made ten-figure offers for Instinct!
But after Instinct declined, Meta said, "Fine, we will build it ourselves." Because for them, it is just a matter of spending money, and they have plenty. Of course, now that AI can write code, copying a product with no moat might be easier than ever, but that misses the point. Without a moat, any good company gets copied, even if humans have to do it manually.
We are in a period where AI is so shiny and fresh that giants are willing to pay premium prices to buy better products from others. But hoping to be acquired is not a strategy, and "better" is not a moat.
I hope Instinct wins, just as I hope many startups can take down Goliath in their respective industries. They are fighting. Just yesterday, they launched a trusted contacts network, enabling people’s assistants to chat with the assistants of their partners, family, and friends.
This is an attempt to create network effects, and they are going all-in. Respectable. The question is whether they have enough time to build genuine network effects before people start using Muse, which will inevitably interoperate with other Muses.
Time is precisely why counterpositioning is such a powerful weapon for companies without barriers.
It makes copying your superior product painful for others, bordering on suicidal. You need to sprint as hard as you can during that window until you have built your long-term moat.
That is all for today. Tomorrow, we will be back in your inbox with a fresh Weekly Optimism Dose.
Thanks for reading.















