Why Your Blockchain Is Going to Fail

Apps that don’t need the chain

Of the 54 apps on MegaETH, 31 run on other chains as well, Aave and Uniswap among them, so MegaETH is a fraction of their business. 23 run only on MegaETH. Those are the apps the chain gave a reason to exist.

The remaining 23 hold under $3M between them, and three hold 97% of that. The DEX Kumbaya has ~$2M. World Markets, another DEX, has ~$660K and Euphoria, a tap-to-trade app, has ~$200K. The other 20 are dead, with zero balances or forgotten funds. There is no trading app, perps exchange or anything else built that could use the speed. The one thing MegaETH said it did better than anyone, almost nothing on it needed. Hyperliquid did the reverse, and the second half of this piece shows how, stage by stage.

23 apps built only for MegaETH, 3 hold 97% of the $2.9M

1. Kumbaya$2M
2. World Markets$659K
3. Euphoria$199K
4. Canonic$50K
5. Avon MegaVault$10K
6. Chisino$8K
7. Velkonix$156
8. WarpX V3$60
9. CurrentX V2$24
10. Offshore Protocol$11
11. evently$7
12. CurrentX V3$5
13. Megaswap$2
14. WarpX V2$1
15. Syscall-SDK$0
16. Prism DEX$0
17. Tulpea$0
18. Premarket$0
19. Backed AAO$0
20. Blackhaven$0
21. MegaChad$0
22. MegaPump DEX$0
23. MegaPump Launchpad$0

A blockchain is an app store

A chain like MegaETH works a lot like Apple’s App Store. It supplies what apps run on and earns when users use them, and it has the same problem at launch, the cold start. No apps means no users, and no users means nobody builds apps. Apple solved it in 2008 because it already owned the users. Over 6 million people had bought an iPhone before the store opened and they downloaded 10 million apps in three days, so Apple could approve every app and still have 500 at launch. Four months earlier, with Apple’s backing, Kleiner Perkins had put up $100M to fund iPhone developers. Apple set the quality bar but left the picking to people who back apps for a living, and gave them its market insight. Deciding who can build also decides who can’t, and that decision, network curation, is the one a new chain gives away when it lets anyone build anything.

Curation is the bouncer on the door

Think of curation as the bouncer at a nightclub. You decide who comes in, why, and how they behave once inside. The job on the door is a full dance floor at midnight, and you get there by making sure the people inside are having the best night they can because of who you let in, and who you didn’t. The better you do that, the longer the queue, and the longer the queue, the more anyone you let in will act the way you need them to.

The bouncer’s job changes through the night. The people who turn up at nine are not the people who make the club at midnight. They are there to make the room worth walking into for the ones who do. Curation is how you move from the users you can get to the users you need, and a chain that lets anyone build on it is not doing that job.

The job at launch

Blockchains are not Apple. Most new chains do not have 6 million users waiting. Every user a new chain wins is pulled off another they already use, so what it offers has to be worth the move. That hands the power to developers. A new chain needs the builder more than the builder needs the chain, and an app that knows it can hold the chain to ransom or leave. Apple’s answer was to buy the app or build its own, many times over. Chains have learned it the other way round. dYdX launched its own chain in 2023 and in 2025 Uniswap did too.

Three ways to get the apps you need

Three ways. Two of them are how any strategy wins, be different in a way that is worth a lot, or be the lowest cost option. The third is the one every chain would take if it could.

Be different. Be different in a way that is valuable enough that developers will build what you ask in order to get access. Apple could pull it off, but for any new chain it is a very hard ask, closer to impossible than realistic. The surest way to control what gets built is to pick a niche and build a killer app yourself.

Subsidise. The alternative is to make the developer the lowest cost producer. Cut their cost of building on you so far that it covers the risk of a new chain and buys you a say in what they build. That means subsidy, and the first question is how much. How is the second. A chain is not an expert in which apps will work. Apple’s answer was a fund run by professional investors, with Apple’s insight behind it. MegaETH’s was an incubator it ran itself, which took no equity in 20 teams and watched the best of them leave. The mandate for MegaETH was obvious. Fund only apps that need the speed.

Betting on a single outside app makes no sense, so the chain spreads the bet across a set of launch apps. Those apps have to create enough value that users come back, so that more and better developers follow. MegaETH’s 54 show the count is not the point. Fifty-four apps, almost none needing its speed, so users still had no reason to stay.

Own the users. Most chains can’t. Apple did. Coinbase launched Base in August 2023 into its own customer base, with more than 100 apps integrated on day one. Few chains have a Coinbase behind them.

MegaETH subsidised, but without a mandate. Its incubator backed 20 teams, took no equity and never asked whether what they built needed the speed. Outside the incubator anyone could build anything. Most of it didn’t need the speed either. A swap on a fast chain is still a swap, and it’s the same swap you can do anywhere else. In July MegaETH said its goal is now apps “only possible on MegaETH”, built by MegaETH itself. That’s a lesson others had learned already.

Speed is not enough

Solana is fast. Ethereum’s L2s are fast. A long list of chains have sold themselves on speed, and by November 2026 most held a fraction of what they once did.

Every chain sold on speed kept under 5% of its peak

1. 89%HyperliquidPeak $2.8B, now $2.5B
2. 4.9%MegaETHPeak $246M, now $12M
3. 3.5%SeiPeak $628M, now $22M
4. 3.1%AptosPeak $1.3B, now $40M
5. 1.3%SonicPeak $1.1B, now $15M
6. 0.4%MovementPeak $296M, now $1.3M

Sei peaked at $628M and was down to $22M by November 2026. Aptos went from $1.3B to $40M. Movement from $296M to $1.3M. Sonic, the Fantom team’s second chain, from $1.1B to $15M. The company behind Kadena gave up and shut down in October 2025. Monad sold itself on speed and held $951M in November. It only peaked in October, so we’ll give it some time before we say.

Solana and Hyperliquid became market leaders, and neither did it by being the fastest. Both were fast enough. Solana built a large app ecosystem early, when being different from Ethereum still worked. Hyperliquid owned one high value app, an on-chain perps exchange, and grew it into the leader of its niche.

Arbitrum is the same story with a different kind of different. It opened on 31 August 2021, when a transaction on Ethereum cost about $37 and Optimism, the only other rollup, was still open to a handful of invited projects. Cheap, open and EVM was enough. GMX, a perps exchange, launched on it the next day. Arbitrum held $2B before it had a token, 18 months later, and only then came the airdrop. Same order as Hyperliquid. The differentiator was timing, and timing runs out, which is why cheap and fast was worth a fortune in 2021 and almost nothing by 2026. Choosing speed in 2026, when speed no longer set a chain apart, was the strategy failure. Blast proved the other half: points before mainnet, $2.3B at the airdrop in June 2024, $5M by October 2026.

How Hyperliquid did it and won

Hyperliquid did the opposite to MegaETH. Its founder, Jeff Yan, came from high-frequency trading. He built one product, a perps exchange, and a chain fast enough to run it.

Hyperliquid’s median order takes 0.2 seconds, fast for an on-chain order book in 2023, too slow for the fastest traders, and fast enough for the users who came. The speed was there to serve the one app that needed it. The product came first and the infrastructure was built to deliver it with the resources at hand.

Its growth followed the adoption curve, which explains who adopts new markets, most often in technology, and when they do. The curve tells you how to launch something users have not used before. You don’t need a new market, a new segment of an existing one will do. Perps were already a big market, on Binance and dYdX. Hyperliquid’s segment was traders who wanted perps near as fast as Binance without handing it their money.

Hyperliquid ran the curve by the book. Innovators first. They are drawn to anything new, take risks others won’t, and can usually afford to lose. Then early adopters, who are more selective and the most influential group on the curve, because the early majority watch what they do. Trust built over months. Then the incentives, which carried it across the chasm to the early majority. The chasm is the gap between the early adopters, who use it because something is new, and the early majority, who use it because other people have proved it works. It is where most products die, loved by a small niche and never wanted by the mainstream. Then a regulated exchange, which means the late majority had come.

Hyperliquid’s adoption curve, and who came on at each stage

1. Innovators NFT collectors on invites, early 2023
2. Early adopters HLP depositors, first outside market maker, 2023
3. The chasm Points programme, Nov 2023 - Nov 2024
4. Early majority Traders after the airdrop, then builders, 2024 - 2025
5. Late majority ICE and a regulated US market, 2026
6. Laggards Not here yet, maybe never

Innovators

Mainnet opened in February 2023 as an invite-only closed alpha for 20 to 30 testers. From March anyone with a referral code could join. The first users were mostly NFT collectors who had never traded a perp, placing $10 trades. Nobody targeted them. They just turned up. In early 2023 NFT trading was still a $2B a month market, full of people who would try anything new. Hyperliquid did not need them for long. It needed them to prove the product worked, with almost no money at risk. The invite was the curation.

The house makes the market. Liquidity came from the founders’ own trading firm, which quoted every market, 25 pairs by June 2023, with no trading fees at first. In May 2023 those strategies became HLP, a vault anyone could deposit into, with no fees and no cut for the team. That solved one side of the market on day one, and it answered the obvious question about a founder trading against his own users.

Early adopters

Hyperliquid did not pay market makers to turn up. Yan wrote the documentation, explained market making on X and helped firms integrate. In September 2023, seven months after launch and two months before any incentives, one of Binance’s largest market makers met the team at a conference in Singapore and was live two weeks later. Trust had formed the slow way, from a product that kept working.

Then a few professional traders arrived, and it only takes a small number to make the market. Retail traders came with them, because the deep market proved further that the product worked and could be trusted.

On 16 March 2026 the top 100 of 1.7M addresses accounted for 81.3% of everything Hyperliquid had ever traded. That is $3.34T of $4.11T. One firm can run many addresses, so trading is likely more concentrated still. On Trade[XYZ]’s markets on Hyperliquid, 8 wallets did half of all market making. Retail was 28% of wallets and 13% of volume.

Crossing the chasm

The incentives. Only once the professionals were there did Hyperliquid pay for growth. The points programme opened on 1 November 2023. The month before, Hyperliquid averaged just under 500 daily traders. By April 2024 it averaged more than 16,000. The points were the push that got growth across the chasm. What came was the early majority, thousands of traders, with the professionals among them the ones who matter.

The points brought users on, which is what they are meant to do. When they paused in May 2024, daily traders fell by more than half, to 6,951. That, however, was still 14 times where they started. They came back when the second phase opened at the end of May.

What matters is what happened when the points ended for good, in November 2024, when 31% of the token supply went to users. Daily traders averaged 12,000 that month. In December, ~28,000. By March 2026, more than 51,000, over 4x November, 16 months on. Monthly volume went from $62B in October 2024 to $436B in March 2026. The early majority had arrived, and it came and stayed without being paid.

Hyperliquid’s traders came for the points and stayed after them

1. Sep 2023 First outside market maker, about 400 traders a day
2. May 2024 Points paused, daily traders fall by more than half
3. Nov 2024 Airdrop, points end for good
4. Mar 2026 51K traders a day, 16 months after points ended

The early majority turns up

The professionals kept coming and got even bigger. One set of wallets that Hyperdash links to Jump Trading was first funded in December 2025, a year after the points ended. By July 2026 it did about 18% of the exchange’s volume. The product, and the market the professionals made, had them come and kept them.

The totals kept growing for 16 months with no points to farm. MegaETH lost 95% of its TVL in just six months.

After the points ended, Hyperliquid grew and MegaETH fell away

Once the traders were there the developers followed. Hyperliquid opened to outside front ends in October 2024 and to outside market builders a year later. One builder, Trade[XYZ], went beyond crypto perps and runs stock, index and commodity markets.

The model is worth a look, because it is where Hyperliquid started to become infrastructure, not just an app. A front end can route orders in for nothing. Listing your own markets costs a stake of 500,000 HYPE, about $40M at November 2026 prices, locked for at least six months and slashable by the validators if you misbehave. You run as many markets as you like, set your own leverage and fees, and your traders pay Hyperliquid’s chain fees as well. Hyperliquid’s fees buy back its own token, which means every builder pays Hyperliquid on every trade and has to buy $40M of its token first, then lock it away.

The late majority whales in

On 29 October 2026 Intercontinental Exchange (ICE), the owner of the NYSE, took a stake in Trade[XYZ], intending to run a regulated US market for Trade[XYZ]’s Hyperliquid perps. The late majority is on its way, through a regulated exchange rather than a referral code. With it Hyperliquid transitions to being an infrastructure play with an app, rather than the other way round. A chain other people’s markets run on, with enormous trading volume underneath. It’s worth noting ICE didn’t go to the fastest chain. It went where the traders are.

Hyperliquid did it in the order that works. Build something different that is valuable enough that users change what they do. Build a community of those users. Then attract the high value users and apps that benefit from those users being there. MegaETH did it in reverse.

The questions to ask when launching a chain

Four questions are worth answering before launch, not after.

1. Which app shows off what only you do, so well that a user would come for it alone?
2. Who are the few hundred users of that app who will bring most of your volume, and what would make them come without being paid?
3. If nobody else will build that app, will you?
4. Which apps will you turn away, because they make you look like every other chain?

If you can’t answer the first one, the points programme won’t save you. It’ll just tell you how many people came by for the rewards, and then how many left.