DIGITAL asset exchanges (DAXs) are often described as a sunrise industry, but the economics is squeezing out small operators in ways few anticipated.
Today, artificial intelligence (AI) can run nearly all core DAX functions, enabling industry leaders to cut costs and stay ahead.
Coinbase is a striking example: Between 95% to 100% of the company’s code is now written with AI, collectively performing the software development work of 1,200 employees, and foreseeably 100,000 employees by 2030!
Its new line of financial advisors are AI-powered (non-human) and registered with the authorities to personalise investment guidance, real-time portfolio analysis, and tax-loss harvesting for clients.
Incidentally, Coinbase slashed 14% of its workforce this year, joining a list alongside Kraken (5%), Crypto.com (12%), Gemini (25%), etc, with no sign of abatement.
Much has been written about the “hollowing out” of the crypto industry and how AI has changed the cost of doing business. Far less attention, however, is paid to how uneven the impact is on small DAXs.
Take compliance, for instance. Baselines have risen everywhere, and DAXs must either shape up or ship out.
In the language of our own regulator, DAXs must “embrace the institutional discipline required of a mature capital market”.
But without AI capabilities, a DAX will struggle to deliver the required levels of efficiency and accuracy, from onboarding to liquidity building to risk engines to fraud detection.
And as specs grow in sophistication, so do costs – for model supervision, specialist engineers, and continuous retraining.
Furthermore, cybersecurity experts repeatedly warn DAXs that their “status quo defence won’t be enough” as AI has widened the attack surface and made threats more lethal.
A large DAX can afford to setup an in-house AI division or piggyback on its foreign well-capitalised parent, but small homegrown ones (whose cheif executive officers spend half the time fundraising) cannot.
Let’s not forget, trading is a computational activity. Retail traders love the idea of having institutional-grade power in a DAX environment with AI toys to feel like quants and customise their signal bots, even if they won’t always use them.
Ironically, as trading becomes “fully industrialised” with agentic execution and robo-intelligence replaces human intuition, large DAXs will gain an even greater edge. Small DAXs become caught in Catch-22: They need AI to keep up with market expectations but can’t afford it.
They pay the same fixed costs for vendor solutions designed for large platforms but have to spread them across much lower turnover and user base.
And alas, the profit pool from local growth isn’t large enough to maintain AI-overhead structures for every operator, especially amid a prolonged crypto winter.
It might make more sense to exit (while the licence can still fetch premium) than to stage a debt-funded capital expenditure-intensive AI turnaround.
Hence, the hunch: Is this the beginning of the end for small DAXs?
Market conditions are unforgiving.
Before the current rebound, spot volume dropped to multi-year lows (since 2023) while altcoins faced the deepest sell-off (since 2020) with weak listing interest across board.
Mid-sized regional players are quitting as legal problems overwhelm them and retail coalesces around global Tier 1 DAXs, including the legendary BitMEX which once occupied the world’s priciest office address in Hong Kong.
Almost 100 projects have reportedly folded in the first half of 2026! Another Top 40 DAX (CoinEx) succumbed just last week.
Meanwhile, the big incumbents are diversifying their crypto shelf with tokenised stocks, precious metals, commodities, pre-initial public offerings, prediction markets, etc, to realise the vision of an “Everything Exchange” — a one-stop shop where users can trade any asset class imaginable round-the-clock with a single login and wallet.
Crossover competition is flooding this overly crowded sector. Not to mention that within the DAX ecosystem, decentralised exchanges are rapidly cannibalising centralised ones.
Brokerage giant Robinhood has successfully linked millions of its customers onchain with zero-fee trades. And by acquiring Bitstamp, it has foothold in over 50 jurisdictions with 5,000 institutional accounts.
Neobanks’ crypto push paid off handsomely for Revolut in Europe, as Nubank follows suit in Latin America for 100 million users. Here in Malaysia, trading apps offering crypto exchange traded funds will hungrily expand if allowed.
The recent shake-out reveals a harsh truth: Spot itself no longer justifies a standalone exchange. That’s yesterday’s business model!
Rubbing salt to this, AI is raising the cost of survival for smaller DAXs. Oddly enough, it’s also becoming their final source of competitiveness.
Because licensing is no longer the moat in 2026, future growth will be concentrated, and benefit the few not many, i.e. large scalable platforms which develop scope economies and use AI to become smarter and safer.
So, unless small DAXs carve out defensible niches, join a strategic network, or hyperlocalise their ramps, the notion that they can compete head-on is naive.
The global industry is giving serious consolidation vibes, Malaysia included. This is just the opening act. Stay tuned, there will be blood.
Edmund Yong is a director of the Generative AI Association of Malaysia and ambassador of the Global Blockchain Business Council founded in Davos. The views expressed here are the writer’s own.
