Privateness will develop into a very powerful moat within the crypto house.
(*8*)Writer: a16z
(*8*)Compiled by: TechFlow
a16z (Andreessen Horowitz) not too long ago launched its listing of potential “massive concepts” for the know-how panorama in 2026. These concepts had been contributed by companions from its Apps, American Dynamism, Bio, Crypto, Progress, Infrastructure, and Speedrun groups.
Beneath is a collection of massive concepts from the crypto house, together with insights from some particular contributors. They cowl a variety of subjects from sensible brokers & AI, stablecoins, tokenization & finance, privateness & safety, to prediction markets and different purposes. For extra on the 2026 tech outlook, learn the total article.
Exchanges Are a Beginning Level, Not the Endgame
Immediately, nearly each profitable crypto firm in addition to stablecoins and some core infrastructure has pivoted to or is pivoting in the direction of turning into an alternate. But when “each crypto firm turns into an alternate,” what is the endgame? A proliferation of homogeneous competitors that fragments person consideration and possible leaves just a few winners. Firms that pivot to buying and selling too early might miss the possibility to construct extra aggressive, sturdy enterprise fashions.
I deeply empathize with founders struggling to make their firms financially viable, however there’s additionally a price to chasing short-term product-market match. That is particularly acute in crypto, the place the distinctive dynamics round tokens and hypothesis typically steer founders in the direction of “on the spot gratification,” like a marshmallow take a look at.
There’s nothing fallacious with buying and selling—it is an essential market operate—however it’s not essentially the tip purpose. Founders who give attention to the product itself and take a long-term view to discovering product-market match might find yourself being larger winners.
– Arianna Simpson, Normal Accomplice, a16z Crypto
New Considering on Stablecoins, RWA Tokenization, Funds & Finance
Assume More Crypto-Native for RWA Tokenization & Stablecoins
We have seen banks, fintechs, and asset managers eager to carry U.S. equities, commodities, indices, and different conventional belongings on-chain. Nevertheless, as extra conventional belongings are introduced onto blockchains, they’re typically tokenized “skeuomorphically”—primarily based on present real-world asset ideas—with out leveraging crypto-native properties.
In distinction, artificial types like perpetual futures (perps) can provide deeper liquidity and are less complicated to implement. Perps additionally present an easy-to-understand leverage mechanism, making them maybe probably the most crypto-native by-product at present. Rising market equities may be one of the attention-grabbing asset courses to “perpify.” For some shares, for occasion, the zero-days-to-expiry (0DTE) choices market is usually deeper than the spot market, making perpification an attention-grabbing experiment.
Finally, it is a query of “perpify vs. tokenize”; both approach, we must always anticipate to see extra crypto-native RWA tokenization within the yr forward.
Equally, in 2026, stablecoins will see extra “issuance innovation, not simply tokenization.” Stablecoins went mainstream in 2025, and issuance continues to develop.
Nevertheless, stablecoins with out robust credit score infrastructure are extra like “slender banks”—holding particular, extremely liquid, perceived-as-ultra-safe belongings. Whereas slender banks are a legitimate product, I do not assume they would be the long-term spine of the on-chain financial system.
We have seen many rising asset managers, curators, and protocols pushing for on-chain asset-backed loans collateralized by off-chain collateral. Typically, these loans are generated off-chain and then tokenized. Nevertheless, I feel the advantages of this tokenization are restricted, maybe solely for distribution to customers already on-chain. As a substitute, debt belongings needs to be generated on-chain, not off-chain and then tokenized. Producing debt on-chain reduces mortgage servicing prices, back-office structuring prices, and will increase accessibility. The problem is compliance and standardization, however builders are engaged on it.
– Man Wuollet, Normal Accomplice, a16z Crypto
Stablecoins Drive Core Ledger Upgrades, Unlock New Cost Use Instances
Immediately, most banks nonetheless run on legacy software program methods that might be unrecognizable to trendy builders: Banks had been early adopters of huge software program methods within the Nineteen Sixties and 70s. Within the 80s and 90s, second-generation core banking software program emerged (e.g., Temenos’s GLOBUS, InfoSys’s Finacle). Nevertheless, these have aged and are upgraded too slowly. So a lot of banking’s essential core ledgers—the databases that file deposits, collateral, and different obligations—nonetheless run on mainframe computer systems utilizing the COBOL programming language, counting on batch file interfaces slightly than trendy APIs.
Many of the world’s belongings nonetheless sit on these decades-old core ledgers. Whereas these methods are battle-tested, trusted by regulators, and deeply embedded in advanced banking workflows, in addition they stifle innovation. Including essential options like real-time funds can take months or years, wrestling with technical debt and regulatory complexity.
That is the place stablecoins are available in. Over the previous few years, stablecoins discovered product-market match and entered mainstream finance. And this yr, conventional finance (TradFi) establishments embraced stablecoins at a brand new degree. Devices like stablecoins, tokenized deposits, tokenized treasuries, and on-chain bonds allow banks, fintechs, and monetary establishments to develop new merchandise and serve extra prospects. Crucially, they’ll accomplish that with out forcing a rewrite of their legacy methods—which, whereas outdated, have run stably for a long time. Stablecoins thus provide establishments a brand new path to innovate.
– Sam Broner
On the Way forward for Sensible Brokers & AI
Utilizing AI to Do Substantive Analysis
As a mathematical economist, at the start of the yr, I discovered it troublesome to get client AI fashions to grasp my workflows; by November, I might give them summary directions as I might a PhD pupil… and they might typically return novel and accurately executed solutions. Furthermore, we began seeing AI utilized in broader analysis—particularly in reasoning, the place fashions can no longer solely instantly assist discovery but additionally autonomously remedy Putnam issues (maybe the world’s hardest undergraduate math examination).
What stays unclear is the place and how this analysis help might be most useful. However I anticipate AI’s analysis capabilities will allow and incentivize a brand new “polymath” analysis fashion: one which leans into speculating about relationships between concepts and rapidly reasoning from extra hypothetical solutions. These solutions will not be absolutely correct, however they level in the precise path inside some logical framework. Paradoxically, this strategy is a bit like harnessing the facility of mannequin “hallucinations”: when fashions develop into “sensible” sufficient, letting them roam in summary house, whereas typically producing nonsense, may yield breakthrough discoveries—very like people are most artistic when free of linear considering and clear instructions.
Considering this manner requires a brand new AI workflow—not simply “agent-to-agent” however extra advanced “agent-wrapping-agent,” the place layers of fashions assist researchers consider earlier mannequin proposals and iteratively distill worth. I’ve used this to write down papers; others use it for patent searches, inventing new artwork types, and even (sadly) discovering new sensible contract assaults.
However operating this “wrapped reasoning agent” analysis requires higher interoperability between fashions and a option to establish and correctly compensate every mannequin’s contributions—issues crypto may help remedy.
– Scott Kominers, a16z Crypto Analysis, Professor at Harvard Enterprise Faculty
The Invisible Tax AI Brokers Impose on the Open Internet
With the rise of AI brokers, an “invisible tax” is bearing down on the open net, basically disrupting its economics. This disruption stems from a rising asymmetry between the web’s context layer and its execution layer: at present, AI brokers extract information from ad-supported content material websites (the context layer) to serve customers comfort, whereas systematically bypassing the income streams (adverts, subscriptions) that fund content material creation.
To stop additional decay of the open net (and shield the various content material that fuels AI), we’d like technical and financial options deployed at scale. This might embrace next-generation sponsored content material, micro-attribution methods, or different novel funding fashions. Present AI licensing offers have confirmed to be stopgaps, typically compensating content material suppliers for solely a fraction of the income misplaced to AI site visitors diversion.
The online wants a brand new techno-economic mannequin the place worth flows routinely. Probably the most essential shift subsequent yr might be shifting from static licensing to compensation primarily based on real-time utilization. This implies testing and scaling methods—possible utilizing blockchain-enabled nanopayments and subtle attribution requirements—to routinely reward each entity whose info contributed to an AI agent efficiently finishing a job.
– Liz Harkavy, a16z Crypto Investing
Privateness as a Moat
Privateness Will Be Crypto’s Most Vital Moat
Privateness is without doubt one of the key properties that may drive international finance on-chain. But, it is a essential component lacking from nearly all blockchains at present. For many blockchains, privateness is usually an afterthought.
However at present, privateness alone is sufficient to be a key differentiating property for a blockchain. More importantly, privateness additionally allows a “chain lock-in,” or a privateness community impact. That is particularly essential in an period the place efficiency competitors is now not a ample benefit.
With bridge protocols, it is trivial for customers to maneuver between chains so long as every part is public. However with privateness, that is now not the case: it is simple to bridge tokens, however it’s extraordinarily laborious to bridge privateness. Customers danger publicity when shifting on or off a privateness chain, whether or not to a public chain or one other privateness chain, as a result of observers of chain information, mempools, or community site visitors would possibly infer their identification. Crossing the boundary between a privateness chain and a public chain, and even between two privateness chains, leaks numerous metadata, like timing and quantity correlations, which might make it simpler to trace customers.
In comparison with many homogeneous new chains whose charges may be pushed to near-zero by competitors, privacy-enabled blockchains can type stronger community results. The fact is, if a “general-purpose” blockchain would not have a longtime ecosystem, a killer app, or an unfair distribution benefit, there’s little cause for customers to make use of it, construct on it, or be loyal to it.
On public blockchains, customers can simply transact with customers on different chains—it would not matter which chain they be part of. On non-public blockchains, nevertheless, the chain a person joins issues so much, as a result of as soon as they be part of, they’re unlikely to maneuver to a different chain to keep away from privateness publicity. This creates a “winner-take-most” dynamic. And since privateness is essential for most real-world use circumstances, a handful of privateness chains might find yourself dominating crypto.
– Ali Yahya, Normal Accomplice, a16z Crypto
Different Industries & Functions
Prediction Markets Will Get Greater, Broader, Smarter
Prediction markets have been going mainstream, and within the coming yr, as they intersect with crypto and AI, they’ll get larger, broader, smarter—and current new, essential challenges for builders.
First, there might be many extra contracts listed in prediction markets. This implies we’ll have real-time odds not simply for main elections or geopolitical occasions, however for nuanced outcomes and advanced cross-events. As these new contracts unearth extra info and develop into built-in into information ecosystems (a pattern already underway), they’ll elevate essential societal questions, like the right way to worth info and the right way to higher design these markets to be extra clear, auditable, and so on.—questions crypto may help remedy.
To deal with the inflow of latest contracts, we’ll want new methods to achieve consensus on real-world occasions to settle them. Centralized platform options (e.g., confirming whether or not an occasion occurred) are essential, however disputed circumstances just like the Zelenskyy lawsuit market and Venezuela election market expose their limits. To deal with these edge circumstances and assist prediction markets scale to extra sensible purposes, new decentralized governance mechanisms and LLM oracles may help adjudicate the reality of disputed outcomes.
AI’s potential extends past LLM-powered oracles. For instance, AI brokers energetic on these platforms might collect alerts globally for short-term buying and selling benefits. This might assist us see the world in new methods and predict future developments extra precisely. (Tasks like Prophet Area are already making this house thrilling.) Past serving as subtle political analysts providing insights, these AI brokers, as we examine their emergent methods, might reveal the underlying predictors of advanced social occasions.
Will prediction markets substitute polls? No. As a substitute, they’re going to make polls higher (and ballot info may be fed into prediction markets). As a political financial system professor, I am most excited concerning the potential for prediction markets to work alongside the wealthy ecosystem of polls—however we’ll must depend on new applied sciences, like AI, which might enhance the survey expertise; and crypto, which might provide new methods to confirm that survey and ballot contributors are human, not bots.
– Andy Corridor, a16z Crypto Analysis Advisor, Professor of Political Financial system at Stanford
Crypto Will Broaden to New Functions Past Blockchains
For years, SNARKs (succinct non-interactive arguments of information, cryptographic proofs that confirm a computation’s correctness with out re-running it) have been used primarily for blockchains. It is because their computational overhead was too giant: proving a computation could possibly be one million instances extra work than operating it. That overhead is worth it when amortized throughout 1000’s of verifiers, however impractical elsewhere.
That is about to vary. By 2026, zkVM (zero-knowledge digital machine) prover overhead will drop to roughly 10,000x, with reminiscence footprints of only a few hundred megabytes—quick sufficient to run on telephones and low cost sufficient for many use circumstances. This is why “10,000x” may be a key threshold: high-end GPUs have roughly 10,000x the parallel throughput of a laptop computer CPU. By the tip of 2026, a single GPU will be capable to generate proofs of CPU executions in real-time.
This unlocks a imaginative and prescient from early analysis papers: verifiable cloud computing. When you’re already operating CPU workloads within the cloud (as a result of your compute is not amenable to GPU acceleration, otherwise you lack experience, or for legacy causes), you can get cryptographic proofs of computational correctness at affordable price overhead. And since provers are optimized for GPUs, your code wants no adjustments.
– Justin Thaler, a16z Crypto Analysis, Affiliate Professor of Laptop Science at Georgetown
– a16z Crypto Editorial
