Tag: blockchain

  • Forecasting Cryptocurrency Prices with AI and Machine Learning: Data, Methods, and Limitations

    Forecasting Cryptocurrency Prices with AI and Machine Learning: Data, Methods, and Limitations

    Cryptocurrency markets generate vast amounts of data every second. Artificial intelligence (AI) and machine learning systems can process price movements, trading volume, derivatives positioning, blockchain activity, and text-based sentiment to help researchers estimate probabilities for future market behavior. While these tools can improve analysis, they cannot reliably predict exact future prices.

    How Machine Learning Identifies Patterns in Market Data

    Traditional technical analysis focuses on price, volume, and indicators. Machine learning models can process a wider range of variables, including historical returns, volatility, order book imbalances, funding rates, open interest, liquidations, and correlations with conventional markets. Models learn relationships between inputs and outcomes. Classification models can estimate whether Bitcoin is more likely to rise or fall, while regression models can estimate returns or price ranges. The CFTC’s Technology Advisory Committee has identified predictive analytics, asset-price forecasting, and analysis of large amounts of unstructured data as potential AI applications in financial markets.

    Blockchain Data Adds Another Information Layer

    Cryptocurrency markets provide data unavailable in many conventional markets because activity occurs on public blockchains. Models can incorporate exchange inflows, wallet balances, realized profits, transaction activity, and long-term holder behavior. Glassnode’s September 2026 research showed how on-chain data can provide context for Bitcoin markets. It reported that Bitcoin gained 23% over 21 sessions while remaining 10% lower for the year. The firm also identified USD 83,000–USD 86,000 as a resistance zone based on long-term holder cost basis and institutional break-even levels.

    AI Can Analyze News and Sentiment

    Natural-language processing can convert news, regulatory announcements, corporate disclosures, and social media discussions into sentiment scores or event classifications. Combining these signals with quantitative data can help models react to information that price charts alone may not capture. However, misleading information, changing narratives, and poor-quality data can reduce forecast reliability.

    Backtests Can Overstate Accuracy

    Historical performance is not proof that a model will work in future markets. Overfitting occurs when an algorithm learns historical noise rather than durable relationships. Performance must also account for spreads, commissions, slippage, latency, and market impact. The CFTC warns that hypothetical trading systems may overlook bid-ask spreads, execution differences, and fees. The regulator also cautions that AI cannot predict the future or sudden market changes.

    Why This Matters

    Crypto trades continuously across fragmented venues and generates vast market and blockchain data. AI can process these signals faster than human analysis, but reliable forecasting still depends on data quality, model validation, and disciplined risk controls across different market cycles and conditions.

    Final Thoughts

    AI can strengthen crypto forecasting by combining market, blockchain, and sentiment information. Its value lies in identifying patterns and probabilities rather than guaranteeing future prices. Successful models still require testing, monitoring, and realistic assumptions about trading costs.

    Frequently Asked Questions

    Can AI accurately predict cryptocurrency prices?

    AI can identify patterns and estimate probabilities using historical and real-time data. However, it cannot reliably predict exact future prices or unexpected events that rapidly change market conditions.

    What data does AI use to forecast crypto prices?

    Models can analyze prices, trading volume, volatility, order books, funding rates, open interest, and liquidations. They can also incorporate blockchain activity, wallet movements, news, and market sentiment.

    How does machine learning help with Bitcoin price forecasting?

    Machine learning models can identify relationships across large datasets that may be difficult to analyze manually. Classification models can estimate price direction, while regression models can estimate potential returns or price ranges.

    How does on-chain data improve crypto forecasting?

    Public blockchains provide information such as exchange flows, wallet balances, realized profits, and transaction activity. Combining these metrics with market data can provide additional context about investor and network behavior.

    What are the main risks of AI crypto predictions?

    Models can suffer from overfitting, poor-quality data, and sudden changes in market conditions. Real-world results can also be affected by trading fees, spreads, slippage, latency, and unexpected market events.

    Disclaimer: This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency investments carry significant risk. Always conduct your own research (DYOR) before making any investment decisions.

  • Ethereum Wallet Growth: What It Really Reveals About Network Adoption and Price

    Ethereum Wallet Growth: What It Really Reveals About Network Adoption and Price

    Ethereum wallet growth is frequently cited as evidence that blockchain adoption is accelerating. More addresses holding ETH or interacting with the network can indicate expanding participation, but wallet statistics are easy to misinterpret. An Ethereum address does not necessarily represent one person, and one person can control hundreds or thousands of addresses. Understanding that distinction is essential when interpreting on-chain adoption.

    Wallet Addresses Measure Accounts, Not People

    Ethereum is pseudonymous. Blockchain data records addresses and transactions rather than verified individual identities. Consequently, counting addresses cannot produce an accurate count of Ethereum users. A trader might operate separate wallets for long-term holdings, decentralized finance, NFTs, and testing. Institutions and exchanges can control numerous addresses, while a single exchange address can represent assets belonging to thousands of customers. Automated bots can create additional addresses without representing additional human users.

    Active Addresses Provide Different Information

    Analysts therefore frequently examine active addresses rather than total addresses. An active-address metric generally counts addresses that send or receive transactions during a defined period. Growth can indicate increasing blockchain activity, but it still measures participating addresses rather than unique people. Activity can also change based on market conditions. During periods of intense trading, token launches, or airdrop campaigns, existing users may create more wallets or interact more frequently. Address growth can therefore increase even without equivalent growth in the number of people using Ethereum.

    Layer 2 Networks Complicate Ethereum Metrics

    Ethereum’s scaling strategy creates another measurement challenge. Increasing amounts of user activity occur on Layer 2 networks rather than directly on Ethereum Mainnet. Rollups execute transactions outside Layer 1 before settling data or proofs back to Ethereum. A Mainnet-only wallet metric can miss activity occurring across Ethereum’s broader scaling ecosystem. Wallet applications themselves increasingly support multiple Layer 2 networks. Ethereum.org’s wallet directory evaluates products partly on Ethereum-native and Layer 2 support, illustrating how wallets are evolving into multi-network interfaces.

    Wallet Growth Does Not Equal Investment Demand

    More wallets also do not automatically mean ETH demand or price must rise. An address can hold negligible ETH, interact primarily with stablecoins, or exist only temporarily. Conversely, institutional custodians may hold substantial ETH through relatively few addresses. Useful adoption analysis therefore combines wallet statistics with transaction activity, fees, stablecoin transfers, decentralized application usage, Layer 2 activity, and capital flows. No individual metric captures the entire network.

    Final Thoughts

    Ethereum wallet growth measures changes in blockchain addresses, not a verified population of investors or users. It can reveal important activity trends, particularly when combined with active-address and transaction data, but interpreting it as direct evidence of user growth or future ETH price performance can produce misleading conclusions.

    Also Read: How Ethereum Scales Without Sacrificing Decentralization

    FAQs

    1. What does Ethereum wallet growth measure?
      Ethereum wallet growth tracks changes in blockchain addresses holding assets or interacting with the network. It can indicate expanding activity but does not directly measure the number of individual Ethereum users.
    2. Does one Ethereum address represent one user?
      No. One person can control multiple Ethereum addresses, while exchanges and custodians can use individual addresses to hold assets belonging to thousands of customers.
    3. What are active Ethereum addresses?
      Active addresses generally refer to addresses that send or receive transactions during a specified period. They provide insight into network activity but still cannot identify the number of unique human users.
    4. Why do Layer 2 networks complicate Ethereum wallet statistics?
      Much of Ethereum’s activity occurs on Layer 2 networks before settling back to Ethereum. Looking only at Mainnet addresses may miss activity taking place across the broader Ethereum ecosystem.
    5. Does increasing Ethereum wallet growth mean ETH prices will rise?
      Not necessarily. Wallet growth does not directly measure investment demand, as addresses can hold small balances or mainly use stablecoins and applications. ETH prices are influenced by many additional market factors.

    Disclaimer: Analytics Insight does not provide financial advice or guidance on cryptocurrencies and stocks. Also note that the cryptocurrencies mentioned/listed on the website could potentially be risky, i.e., designed to induce you to invest financial resources that may be lost forever and not be recoverable once investments are made. This article is provided for informational purposes and does not constitute investment advice. You are responsible for conducting your own research (DYOR) before making any investments. Read more about the financial risks involved here.

  • XRP Market Cap Crosses $100B as XRPL Permission Upgrade Enters Countdown

    XRP Market Cap Crosses $100B as XRPL Permission Upgrade Enters Countdown

    XRP has regained the $100 billion market cap milestone after a sharp 24-hour rally, while the XRP Ledger (XRPL) moves one step closer to activating a major new feature. Data from CoinGecko shows XRP rose 8% in the latest session, with the token testing resistance near $1.65 amid expanded trading activity.

    Market-Wide Crypto Strength

    The move placed XRP among the top performers in the broader cryptocurrency market. Other assets including Arbitrum, Flare, and Zcash also recorded meaningful gains during the same trading period.

    Even with the latest surge, XRP remains more than 55% below its all-time high from nearly a year ago. The gap underscores how much ground the token still needs to cover to regain its previous peak.

    Trader Peter Brandt has suggested XRP could eventually reach $5.40. Meanwhile, Coinbase data indicates that expected volatility around XRP has risen in recent sessions.

    ETF Inflows and Whale Activity Pick Up

    US spot XRP ETFs recorded $13.03 million in net inflows on September 22, according to CoinGlass. That followed a day with zero net inflows. Total cumulative inflows across those funds now stand at roughly $936.66 million.

    One fund accounted for about $12.47 million of the September 22 total, or around 95.7% of the day’s net buying, signaling that a single product drove most of the demand.

    On-chain data also points to rising activity. Santiment logged 1,917 XRP transactions worth at least $100,000 on Tuesday, the highest one-month count. The XRP Ledger added 3,647 new wallets during the same stretch.

    Notably, the burst in high-value transfers came while XRP was rallying, not during a selloff. That timing sets the current increase apart from the large transfers that often accompany market declines.

    XRPL Permission Upgrade Moves Toward October Activation

    The PermissionDelegationV1_1 amendment entered a 14-day activation countdown on September 21 after 29 of the XRP Ledger’s 35 trusted validators voiced support. If that support is maintained, the amendment could activate on October 5 at 11:18 UTC.

    The feature must keep the backing of at least 80% of validators, meaning 28 or more validators must stay supportive during the countdown. If support falls below that level, the activation clock resets.

    When active, the protocol will allow one XRPL account to grant limited authority to another account without giving up control of the primary keys. For example, a stablecoin issuer could authorize a compliance system to approve customers, while keeping its main keys offline.

    Another operations account could receive permission to process payments, but it would not be able to change keys or transfer permissions to a third party. Each delegated account can receive up to 10 permissions, and the main account can later modify or revoke them.

    This design mirrors traditional banking controls, where payment, compliance, and other functions run under separate authority structures.

    Revised Upgrade Follows Earlier Security Flaw

    PermissionDelegationV1_1 is the second attempt to introduce delegated permissions on the XRP Ledger. The original version contained a vulnerability that could have allowed an attacker to force another account to pay transaction fees without a valid signature.

    Repeated transactions with deliberately high fees could have drained a victim’s XRP balance. The issue stemmed from the order in which the software validated transactions: account permissions were checked before the system verified whether a signature was valid.

    As a result, some failed transactions could still charge fees before the invalid signature was detected. A community tester found the flaw on September 15, 2025, while testing outside the main network.

    Validators were told to reject the original amendment, and it never reached activation. The revised PermissionDelegationV1_1 proposal has now gathered enough initial support to start the countdown, but it still needs sustained validator backing through October 5 to become active.

  • Solana’s Alpenglow Consensus Upgrade Reaches Public Testnet With 150ms Finality Target

    Solana’s Alpenglow Consensus Upgrade Reaches Public Testnet With 150ms Finality Target

    Solana has advanced its Alpenglow consensus upgrade to the threshold of public testnet deployment, giving developers and validators a broader environment to test a major shift in transaction finality. The upgrade is designed to cut finality time from roughly 13 seconds to about 150 milliseconds.

    What Alpenglow Changes

    Solana currently relies on TowerBFT, a consensus mechanism that requires validators to record votes on-chain. Under TowerBFT, validators must accumulate enough votes across 32 slots before a block is considered final. Alpenglow replaces this with a new protocol called Votor, which lets validators exchange votes directly rather than depending on the same on-chain voting sequence.

    With Votor, agreement can be reached in one or two voting rounds. One round can finalize a block when sufficient validator stake participates; a second round offers an alternative path when participation is lower. Anza estimates median finality near 150 milliseconds, with earlier simulations reaching about 100 milliseconds under favorable conditions.

    From an application or wallet perspective, the upgrade is designed to be mostly invisible. Transactions continue through the same interfaces; the notable changes take place in validator communication and consensus.

    Expanding Testing to Public Testnet

    Alpenglow has been running for more than four months on a smaller community cluster built specifically for consensus testing. Moving to Solana’s public testnet brings more validators and infrastructure providers into the evaluation process. Because testnet tokens hold no real-world value, developers can restart the network, inspect failures, and test migration procedures without risking mainnet assets.

    Anza first moved Alpenglow into community validator testing in May, describing it as the largest consensus change in Solana’s history. Validators participating in the test are expected to run Agave 4.3, the latest branch of Anza’s validator software.

    Agave 4.3 and Mainnet Timeline

    Anza recommended Agave 4.3 for broader mainnet validator adoption on September 21. The rollout initially targeted operators representing 10% of mainnet stake, and the recommendation later expanded to 25%. By August, developers expected the 150-millisecond finality improvement to arrive through Agave 4.3.

    Anza’s schedule lists September 28 as a tentative date for resuming mainnet feature activation, though the date is subject to change. That date should not be read as confirmation that Alpenglow will activate on Solana mainnet. As of early Wednesday, Anza’s feature gate tracker still listed Alpenglow testnet activation as pending.

    Finality is a separate measure from other performance upgrades. Solana recently reduced its target slot time from 300 milliseconds to 250 milliseconds via SIMD-0525, bringing the network to four slots per second. That change also shortened each validator’s four-slot leader window from 1.2 seconds to one second. SIMD-0525 ultimately targets 200-millisecond slots, or five slots per second, but no confirmed mainnet activation date has been announced for that stage.

    Alpenglow follows a separate process through SIMD-0326. Rather than shortening slots, it replaces TowerBFT with Votor and changes how validators agree that blocks have reached finality.

  • MoonPay Bolsters Tokenized Asset Strategy With $60M All-Stock North Capital Acquisition

    MoonPay Bolsters Tokenized Asset Strategy With $60M All-Stock North Capital Acquisition

    Cryptocurrency payments company MoonPay is expanding beyond digital asset payments with an all-stock acquisition of North Capital, a private-markets investment platform, in a deal valued at more than $60 million. The transaction, announced Wednesday, is pending regulatory approval.

    North Capital to Operate as MoonPay Subsidiary

    Once the deal closes, Salt Lake City-based North Capital will become a wholly owned subsidiary of MoonPay. The platform has processed about $9 billion in combined primary and secondary transaction volume.

    North Capital supplies technology infrastructure for private securities issuers and fund managers, covering capital raising, asset management, clearing, custody, and secondary trading. Its affiliates also hold broker-dealer, trading, transfer, and investment advisory registrations with the U.S. Securities and Exchange Commission (SEC).

    MoonPay said the acquisition will add North Capital’s securities tokenization infrastructure and regulatory capabilities to its existing payments and tokenization stack.

    Why MoonPay Is Buying North Capital

    The deal supports MoonPay’s focus on tokenized real-world assets (RWAs). MoonPay CEO Ivan Soto-Wright said the acquisition will help create the regulatory foundation needed for wider adoption of tokenized RWAs.

    “We believe bringing those capabilities into the MoonPay ecosystem can help connect different parts of the financial system through modern, programmable infrastructure,” Soto-Wright said.

    Tokenization converts real-world assets such as equities, bonds, and commodities into blockchain-based tokens that can be bought, sold, and traded on distributed ledgers.

    MoonPay’s Push Beyond Crypto Payments

    The purchase is the latest step in MoonPay’s move beyond its core cryptocurrency payments business. The company’s Trade platform connects banks and fintech companies with tokenized assets, decentralized finance (DeFi) protocols, and stablecoin liquidity.

    Earlier this year, MoonPay also acquired Solana-based trading infrastructure provider DFlow and security startup Sodot. The North Capital deal adds another layer to its push into tokenized assets, a growing blockchain use case among traditional financial institutions.

  • How XRP Ledger’s Native AMM and Order Books Could Power Institutional Tokenized Markets

    How XRP Ledger’s Native AMM and Order Books Could Power Institutional Tokenized Markets

    Why XRP Ledger Is Building for Tokenized Finance

    The XRP Ledger (XRPL) is increasingly positioning itself as a foundation for tokenized financial assets. Unlike networks that depend heavily on external smart contracts to run decentralized exchanges, XRPL embeds both an order-book DEX and an Automated Market Maker (AMM) directly into the protocol. This native integration of trading, token issuance, and permissioned access gives institutions a distinct set of tools for building compliant tokenized markets.

    Order Books and AMMs: Two Liquidity Engines on One Ledger

    XRPL’s DEX lets participants create offers that specify the exact assets and prices they are willing to trade. In parallel, the protocol’s AMMs supply liquidity through pooled reserves, with exchange rates set algorithmically. When a trade is routed through the open DEX, it can tap order-book offers, AMM liquidity, or both to find the most efficient execution path.

    Liquidity providers who deposit into an AMM receive LP tokens representing their share of the pool and earn income from trading fees. It is also worth noting that on XRPL, only one AMM is permitted for any given asset pair, a design choice that simplifies the pool landscape.

    Permissioned DEX: Adding an Institutional Gate

    Institutional markets often require strict control over who can trade with whom. XRPL’s Permissioned DEX approach uses permissioned domains and credentials to create controlled trading environments. A domain acts like an allow-list, while credentials prove that a participant meets the requirements needed to access restricted trading.

    Trades in this permissioned setting are matched only with offers in the same domain. Hybrid offers offer more flexibility by interacting with both a designated permissioned DEX and the open DEX. However, one important limitation is that permissioned trades cannot directly consume AMM liquidity. That creates a potential trade-off between regulatory compliance and unfettered liquidity access.

    Multi-Purpose Tokens: Built-In Compliance Controls

    XRPL’s Multi-Purpose Tokens (MPTs) are designed for fungible assets that need institutional-grade controls. MPTs can support supply caps, transfer fees, holder authorization, freezing, and clawbacks—capabilities that map naturally to tokenized funds, debt, or other regulated assets where issuers need meaningful oversight of ownership and transfers.

    Native MPT trading on the DEX is still being developed under the MPTokensV2 proposal. If implemented, it would extend order-book offers, cross-currency payments, and AMM support to MPT-based assets.

    Institutional Pilots Are Already Underway

    In February 2026, Aviva Investors and Ripple announced plans to explore tokenizing traditional fund structures on XRPL. The initiative marks Aviva Investors’ first foray into blockchain-based fund tokenization and signals growing interest in XRPL as a settlement layer for regulated assets.

    Technology Alone Won’t Build Institutional Markets

    Protocol features are only part of the equation. Tokenized assets still require regulated issuers, qualified custody providers, professional market makers, reliable pricing sources, clear legal frameworks, and enough buyers and sellers to sustain deep secondary-market liquidity.

    XRPL advertises settlement in roughly three to five seconds and transaction costs of fractions of a cent. But institutional adoption will ultimately hinge on whether compliant markets using those rails can achieve real depth.

    Why This Matters

    Institutional tokenization is not just about putting traditional assets on-chain. XRPL’s combination of order books, AMMs, permissioned trading, and token controls could provide the connective tissue between compliant issuance and secondary-market liquidity—making tokenized financial products more practical for institutional participants.

    The Bottom Line

    XRP Ledger brings together native order books, AMMs, credentials, and tokenization controls in one settlement environment. The tools are compelling, but their success depends on liquidity, issuer adoption, regulatory clarity, custody standards, and how smoothly permissioned and open-market infrastructure integrate.

  • Crypto Matures: Arc and Prosper Take Opposite Routes to Institutional Trust

    Crypto Matures: Arc and Prosper Take Opposite Routes to Institutional Trust

    Crypto has spent years chasing the same goal from two directions: more real economic activity, more serious capital, without becoming either fully permissioned finance or a pure speculative casino. This week produced two answers to that problem, launched a day apart, and neither one looks like the other.

    Circle’s Arc: Permissioned Trust via Institutional Validators

    Circle’s answer is Arc. The Layer 1 network went live with a founding validator cohort that reads like a partial list of global finance itself: BlackRock, the Depository Trust & Clearing Corporation, ICE, Mastercard, MoneyGram, SBI Group, Standard Chartered, Sumitomo, Visa, Worldpay and Galaxy. The chain is explicitly permissioned. Circle presents that as a selling point, pairing it with a defined governance perimeter it says lets banks use a public chain for treasury, trading and confidential payments. USDC, with roughly $74 billion in circulation, is wired in as the network’s gas token. The company also completed the genesis mint of ARC this week, all 10 billion tokens created at once, while being careful to say the mint is not a commitment to a public launch. It’s a technical step toward a possible move to proof of stake, not a trading debut. Arc’s bet is that institutions will trust a public chain more readily if the validator set looks like the institutions themselves.

    Prosper’s Performance Markets: Permissionless Transparency

    Prosper’s answer is close to the opposite bet. Performance Markets, launched on Pharos the same week, keeps everything public and permissionless. Any curator with an onchain track record can deploy a vault. There’s no validator gate, no cohort of institutional backers, no allocation set aside for insiders on the token side. Instead of importing trust through recognizable names, Prosper tries to manufacture it through transparency: every vault’s NAV, fees, and buybacks are verifiable onchain, and a high-water mark — the standard hedge-fund mechanism that only credits gains above a fund’s previous peak — governs when a share of performance fees buys back and burns the vault’s token.

    What that token is, and isn’t, is the part worth sitting with. Laura Shi, Chief Business Officer at Pharos, where she leads business architecture, RealFi innovation and ecosystem expansion, frames it this way:

    “Think of p{VAULT} as a token built around a public track record, not a share in a fund. Anyone can see how the Curator and the strategy are performing onchain, and use that information to decide what the token is worth. Strong results may build confidence; weak results may reduce it. But there is no automatic one to one link to the Vault’s NAV. Buyback and burn affects supply, it does not promise price support.”

    That’s the core difference between the two experiments. Arc is trying to earn trust by putting familiar institutions in the room. Prosper is trying to earn it by putting the data in the open and letting the market decide what a track record is worth, without wrapping that decision in an ownership claim.

    Which Model Will Prevail?

    Neither model has proven itself yet. Arc’s real test is whether the institutions that signed on as validators actually route meaningful volume through the chain, rather than treating the launch as a pilot. Prosper’s real test is narrower and more mechanical: whether curators show up in numbers, and whether the first vaults clear their high-water marks publicly — the moment the buyback mechanism actually has fees to work with.

    Watching which of these gains traction first won’t settle which model is “right.” But it will say something concrete about what kind of trust crypto’s next phase actually runs on: borrowed from institutions, or built from data anyone can check.

  • How AI-Powered Apps Are Revolutionizing Cryptocurrency Trading

    How AI-Powered Apps Are Revolutionizing Cryptocurrency Trading

    Artificial intelligence is rapidly becoming a cornerstone of cryptocurrency trading infrastructure. AI-powered applications can process price action, derivatives data, blockchain activity, news, and sentiment far faster than any individual trader. While some systems only generate signals, newer agentic tools can connect directly to exchanges or wallets and execute transactions autonomously. However, these systems produce probabilistic decisions, not guaranteed forecasts.

    AI Combines Multiple Data Sources

    Conventional technical analysis typically relies on price, volume, and indicators. AI, on the other hand, analyzes vast datasets including Bitcoin momentum, order-book depth, funding rates, open interest, liquidations, exchange flows, and volatility. Machine learning identifies historical relationships associated with specific market events, while natural language processing (NLP) incorporates financial news, regulatory announcements, and social media discussions. The result: systems can classify market conditions as bullish, bearish, or neutral, assign probabilities to potential movements, or rank possible trading sequences.

    From Signals to Autonomous Execution

    AI analysis is increasingly extending beyond recommendations. Applications connected to exchange APIs can convert trading signals into orders, adjust position sizes, and monitor positions continuously. This trend became more visible in 2026: Kraken launched its CLI in March with 134 commands supporting spot trading, futures, staking, transfers, and WebSocket streaming. Kraken said the system was designed for developers and AI agents to access crypto markets directly. Coinbase introduced Agentic Wallets in February 2026, providing infrastructure that lets AI agents autonomously spend, earn, and trade while using built-in security guardrails.

    Historical Performance Can Be Misleading

    Machine learning models depend heavily on training data. A strategy optimized during a Bitcoin bull market may identify relationships that weaken during sideways or declining markets. Overfitting is another risk: a model can perform exceptionally well against historical observations but deteriorate once deployed with live capital. Trading costs—exchange fees, bid-ask spreads, slippage, and execution latency—can reduce or eliminate profits suggested by backtests. The CFTC warns that AI cannot predict sudden market changes and urges skepticism toward bots or signal services promising guaranteed or unusually high returns.

    Risk Controls Remain Essential

    Automated execution makes risk management particularly important, as errors can be repeated quickly. Position limits, stop-loss rules, exposure controls, model monitoring, and independent validation help constrain losses. The CFTC’s Innovation Task Force, announced in 2026, specifically includes artificial intelligence and autonomous systems among the technologies for which it is developing regulatory frameworks.

    Why This Matters

    The crypto market operates 24/7 and generates vast real-time data. While AI can process and respond rapidly, the need for reliable models, transparent strategies, and robust risk management increases proportionally.

    Final Thoughts

    AI can play a significant role in monitoring the cryptocurrency market, identifying trends, and executing trades, but it cannot eliminate trading uncertainty. Performance still depends on data quality, model design, changing market conditions, transaction costs, and disciplined risk controls.

    Frequently Asked Questions

    1. How does AI analyze cryptocurrency markets?
    AI systems process price movements, volume, order books, funding rates, open interest, liquidations, and blockchain activity simultaneously. NLP also incorporates financial news, regulatory developments, and social media sentiment.

    2. How do AI-powered apps generate crypto trading signals?
    Machine learning models search historical and real-time datasets for patterns associated with particular market conditions. They classify conditions as bullish, bearish, or neutral, calculate probabilities, or rank potential trading opportunities.

    3. Can AI automatically execute cryptocurrency trades?
    Yes. AI applications connected to exchange APIs or compatible wallets can convert signals into orders, determine position sizes, and monitor positions. Newer agentic infrastructure makes autonomous execution increasingly accessible.

    4. Can AI accurately predict Bitcoin and cryptocurrency prices?
    AI identifies patterns and estimates probabilities but cannot reliably predict future prices. Unexpected events, changing variable relationships, poor training data, and overfitting can cause previously successful models to fail.

    5. What are the major risks of using AI for crypto trading?
    Key risks include overfitting, inaccurate signals, sudden market changes, automation errors, slippage, and transaction costs. Position limits, stop-loss rules, exposure controls, and continuous model monitoring remain critical when automated systems trade.

  • European Central Bank Launches Pontes Bridge for Tokenized Asset Settlement via TARGET Services

    European Central Bank Launches Pontes Bridge for Tokenized Asset Settlement via TARGET Services

    The European Central Bank (ECB) has officially launched Pontes, a new service enabling euro-area financial institutions to settle tokenized asset transactions using central bank money. The platform bridges distributed ledger technology (DLT) platforms with the Eurosystem’s TARGET Services, offering a secure settlement method for wholesale markets.

    Starting Monday, September 21, 2026, eligible credit institutions can leverage Pontes for wholesale transactions, with initial availability limited to standard European business hours. The ECB plans to expand the service to round-the-clock settlement in later phases.

    ECB President Christine Lagarde is scheduled to open a roundtable event in Frankfurt at 17:00 CET, followed by a panel discussion chaired by Executive Board member Piero Cipollone. The launch represents a key milestone in the Eurosystem’s strategy to integrate DLT while preserving the role of central bank money as the settlement anchor.

    Pontes, Latin for “bridges,” connects tokenized markets—where bonds, deposits, and other claims are represented as digital tokens on distributed ledgers—with existing Eurosystem infrastructure. This interoperability allows participants to settle the cash leg of tokenized trades in central bank money, addressing a fundamental need as tokenization gains traction.

    The ECB developed Pontes following a comprehensive testing program from May to November 2024, involving 64 market participants in over 50 trials. The program tested connections between DLT platforms and central bank settlement systems, processing more than €1.5 billion in transactions. Insights from these trials shaped the final service design.

    Pontes is part of a two-track strategy alongside Appia, a longer-term initiative exploring a future European financial ecosystem built on tokenization and DLT. While Pontes focuses on immediate interoperability, Appia aims to produce a blueprint by 2028 for common infrastructure and standards. Together, they reflect the Eurosystem’s commitment to innovation without disrupting current systems.

    The service is distinct from the proposed digital euro, which targets retail users for everyday payments. Pontes operates exclusively in wholesale financial markets, providing institutions a pathway to use central bank money in tokenized transactions without altering consumer payment access.

  • Absa Rolls Out Institutional Digital Asset Custody in South Africa Using Ripple

    Absa Rolls Out Institutional Digital Asset Custody in South Africa Using Ripple

    Absa Corporate and Investment Banking (CIB) has launched an institutional digital asset custody service in South Africa, built on Ripple’s custody technology. The new offering gives corporate and institutional clients a bank-operated environment for storing cryptocurrencies and other tokenized assets.

    What Absa Digital Asset Custody Offers

    Absa Digital Asset Custody is designed for custody rather than trading. Clients can store digital assets without relying on retail exchanges or managing private keys through their own systems. The platform includes private-key management tools, approval workflows, transaction controls, and audit trails. These features help institutions meet security, governance, and regulatory requirements when holding digital assets.

    Ripple Provides the Underlying Technology

    Ripple’s custody technology powers the platform, allowing financial institutions to control private keys, define approval policies, and manage digital assets within their existing operating structures. Absa remains responsible for its service and customer relationships.

    “By leveraging Ripple’s custody technology, Absa is setting a new benchmark for secure, compliant, and bank-grade digital asset management on the continent,” said Reece Merrick, Ripple’s managing director for the Middle East and Africa.

    Absa and Ripple first announced their partnership on October 15, 2025, with Ripple identifying Absa as its first major custody partner in Africa.

    Built for Corporate and Institutional Clients

    The service is aimed at corporate and institutional clients, not retail users. It enables organizations to set stronger approval controls, defined access rules, and clear records while keeping custody within a bank-led service.

    Absa has not disclosed which cryptocurrencies or tokenized assets are supported, nor has it published fees or named early customers. Prospective clients must contact Absa’s sales team for eligibility and service terms.

    What’s Next for Absa and Ripple

    Absa plans to expand the platform as institutional demand evolves, while monitoring South Africa’s regulatory framework. Future additions will depend on customer needs, compliance requirements, and operating conditions.

    Ripple continues to grow its institutional custody business. In February 2026, it added Securosys hardware security modules and Figment staking capabilities, giving financial institutions more options for protecting private keys and managing proof-of-stake assets.