Visa and financial operations platform Ramp are introducing AI agents to automate corporate bill payments.
Editorial
This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.
Leveraging Visa’s Intelligent Commerce and Trusted Agent Protocol, the two companies are replacing manual workflows with automation and real-time controls to help large, global organizations reduce complexity in accessing, managing and spending funds.
Ramp serves 50,000 corporates with a unified platform that combines corporate cards and expense management, bill payments, procurement, travel booking, treasury, and automated bookkeeping.
The new suite of AI agents will provide Ramp customers with greater payment flexibility and more control over corporate spending, says Colin Kennedy, chief business officer at Ramp.
“The best financial systems don’t add controls after the fact—they build them into every transaction,” he says. “That’s what we’re delivering with Visa.”
On April 2, 2026, Toulouse-based ST Group announced its upcoming listing on Lise, a specialized French stock exchange utilizing blockchain technology for Small and Medium Enterprises (SMEs). The subscription period for the aerospace component manufacturer begins on April 9, marking a shift toward digital financial markets for the French defense industrial base.
The Lise exchange employs distributed ledger technology to streamline the issuance and trading of shares, offering a decentralized alternative to traditional regional markets. ST Group, which has operated since 1998, intends to use the capital to increase production capacity for clients like Airbus and Dassault.
By utilizing a blockchain-based infrastructure, the exchange provides transparent, real-time settlement and reduced administrative overhead for industrial firms. The project gained momentum following a recommendation from the General Directorate for Armament (DGA), highlighting the growing institutional trust in blockchain-driven capital raising.
🧭 FAQs
• What is the primary location of ST Group operations? The firm operates its main manufacturing facilities in the Toulouse region.
• How does blockchain technology benefit the Lise exchange? It provides a transparent and efficient decentralized ledger for SME share trading.
• When can investors participate in the subscription? The period for public subscription opens locally on April 9, 2026.
• Who advised ST Group on this blockchain financing? The French General Directorate for Armament introduced the company to the platform.
The partnership brings leveraged derivatives to one of crypto’s largest consumer distribution channels, targeting emerging-market users priced out of traditional brokerages.
Wallet in Telegram has launched perpetual futures trading through a new integration with Lighter, the Ethereum-based decentralized exchange (DEX), the teams announced Thursday.
The feature enables users to open long and short positions on more than 50 assets — spanning crypto, metals, equities, oil, and ETFs — with up to 50x leverage and a minimum position size of $1, all without leaving the Telegram app.
The Open Platform (TOP), the entity that develops Wallet in Telegram, told Forbes that it evaluated multiple decentralized perpetual exchanges before selecting Lighter, with the decision driven by cost structure, incentive design, and alignment with a retail-heavy audience. Lighter’s zero-fee model for standard accounts was a key factor.
Rather than competing for power users on standalone exchanges, Wallet is targeting a broader audience that may not have previously used derivatives platforms. The wallet has more than 150 million registered users, many of whom were onboarded through earlier gamified mini-app features on Telegram.
Users in the United States and the United Kingdom are excluded from the rollout. The initial focus is on emerging markets where traditional brokerage infrastructure is more limited.
Lighter Struggles Post-TGE
The deal represents a significant distribution channel for Lighter, which has become one of the top perp DEXs by volume since launching its public mainnet in late 2025.
Lighter processed $59 billion in perpetual volume in March 2026, ranking fourth among perp DEXs, according to DefiLlama. That’s down nearly 80% from its peak of $292 billion in November.
The platform runs on a custom zero-knowledge rollup on Ethereum, where every order match and liquidation is cryptographically verified onchain. It raised $68 million in November 2025 from Founders Fund, Ribbit Capital, Haun Ventures, and Robinhood. Since then, Lighter has expanded into spot trading, launched its LIT token, and introduced equity perpetuals.
Still, the exchange trails category leader Hyperliquid, which processed nearly $210 billion in March, by a wide margin.
The platform’s LIT token rallied 5% on the news, but has struggled since its December launch, losing more than two-thirds of its value since January 1.
LIT Chart
Lighter also announced that its Partner Attribution program is now open, allowing developers to integrate the exchange’s perpetuals and spot infrastructure into their own applications.
Kulipa, a Paris-based stablecoin-native card issuing infrastructure platform, has raised $6.2 million in seed funding.
Editorial
This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.
The round was co-led by Flourish Ventures and 1kx, with participation from White Star Capital and Fabric Ventures.
Stablecoins have soared in popularity over the last year and now settle more than $300 billion daily. But, this still represents only a small share of global payment flows, says Kulipa, in part because the infrastructure connecting onchain settlement with regulated card networks remains fragmented and capital-intensive, often relying on prefunded structures and regionally limited licenses.
To address this gap, the startup has built a stablecoin-native issuing infrastructure designed for capital efficiency, seamless compliance, and global scale. The platform enables partners to launch payment programmes funded directly from stablecoin balances, supporting both rapid pre-funded deployments and deeper wallet-native integrations.
By verifying balances and triggering settlement onchain, Kulipa says it reduces reliance on collateral-heavy prefunding and enables partners to scale more sustainably. Cards issued through the platform can be used anywhere major card networks are accepted, including for retail payments and ATM withdrawals. In addition, Kulipa assumes fraud liability on issued programmes.
Kulipa operates a local-first issuing model with regulated coverage across the European Union, Argentina, and Nigeria, with US expansion underway through BIN sponsorship. Since launching its infrastructure in February 2025, the firm has issued more than 120,000 cards and signed 20 customers, including Flutterwave, Solflare, nSave, and Ready.
“Stablecoins have proven their value as a settlement layer, but using them in everyday financial products is still early,” says Kulipa CEO Axel Cateland, who previously led global Apple Pay and Google Pay deployments at Mastercard.
“Card issuance is the bridge between onchain balances and real-world payments. We built Kulipa to give regulated fintech platforms the compliant, capital-efficient infrastructure they need to operate at global scale.”
The amount of Bitcoin supply in profit and loss is now getting closer to levels typical of a bear market, according to a CryptoQuant analyst.
There are currently about 11.2 million Bitcoin (BTC) in profit. The previous bear market recorded 9 million BTC in profit at its lowest point, CryptoQuant analyst “Darkfost” said Thursday.
CryptoQuant data also shows there are about 8.2 million Bitcoin at a loss, with Glassnode data confirming it’s at levels not seen since late 2022.
“This is quite significant, considering that during the last bear market this figure reached about 10.6 million BTC,” Darkfost said.
Analysts have been debating whether Bitcoin has further to fall this year amid growing global turmoil. Bitcoin metrics that show a movement toward previous cycle lows could suggest that a market bottom is getting closer.
“This suggests that the market is reaching a notable level of undervaluation, comparable to the conditions observed during the previous bear market,” the analyst added.
Bitcoin in profit and loss at bear market lows. Source: CryptoQuant
Analyst sees increasing market stress, not undervaluation
However, Andri Fauzan Adziima, research lead at the Bitrue exchange, argued the data signals “increasing market stress, not immediate undervaluation.”
True capitulation bottoms saw deeper pain, he told Cointelegraph. The supply in loss in 2022 was greater than 50% and the supply in profit was around 45% or lower, while metrics such as net unrealized profit/loss (NUPL) and market value to realized value ratio (MVRV) were at “extremes.”
“Current data points to early/mid-bear transition (potential structural bottom near $55,000), with more downside or consolidation likely before a full reset.”
Related: Bitcoin’s drawdown is ‘less dramatic’ this cycle, Fidelity says
Data also shows Bitcoin has declined by about 52% from its all-time high this cycle, much less than previous bear markets, which saw 77% to 84% drawdowns from their cycle highs.
Strong dollar hampering recovery
Bitcoin author Timothy Peterson commented on X that Bitcoin “tends to struggle when the dollar is strong, and the Chinese yuan is weak.”
He added that this was due to tighter global liquidity, with higher dollar yields attracting capital into cash and bonds and cautious investor sentiment as China eases policy.
That only changes when US interest rates fall and “dollar yield loses its attractiveness,” which is not likely until the second half of 2026 or more likely 2027, he said.
The US dollar index (DXY) has gained about 5% over the past two months, according to TradingView.
DXY has strengthened since late January. Source: TradingView
Magazine: Your guide to surviving this mini-crypto winter
Cointelegraph is committed to independent, transparent journalism. This news article is produced in accordance with Cointelegraph’s Editorial Policy and aims to provide accurate and timely information. Readers are encouraged to verify information independently. Read our Editorial Policy https://cointelegraph.com/editorial-policy
Google has identified six trap categories—each exploiting a different part of how AI agents perceive, reason, remember, and act.
Attacks range from invisible text on web pages to viral memory poisoning that jumps between agents.
No legal framework yet decides who is liable when a trapped AI agent commits a financial crime.
Researchers at Google DeepMind have published what may be the most complete map yet of a problem most people haven’t considered: the internet itself being turned into a weapon against autonomous AI agents. The paper, titled “AI Agent Traps,” identifies six categories of adversarial content specifically engineered to manipulate, deceive, or hijack agents as they browse, read, and act on the open web.
The timing matters. AI companies are racing to deploy agents that can independently book travel, manage inboxes, execute financial transactions, and write code. Criminals are already using AI offensively. State-sponsored hackers have begun deploying AI agents for cyberattacks at scale. And OpenAI admitted in December 2025 that the core vulnerability these traps exploit—prompt injection—is “unlikely to ever be fully ‘solved.'”
The DeepMind researchers aren’t attacking the models themselves. The attack surface they map is the environment agents operate in. Here’s what each of the six trap categories actually means.
The Six Traps
First there are “Content Injection Traps.” These exploit the gap between what a human sees on a webpage and what an AI agent actually parses. A web developer can hide text inside HTML comments, CSS-invisible elements, or image metadata. The agent reads the hidden instruction; you never see it. A more sophisticated variant, called dynamic cloaking, detects whether a visitor is an AI agent and serves it a completely different version of the page—same URL, different hidden commands. A benchmark found simple injections like these successfully commandeered agents in up to 86% of tested scenarios.
Semantic Manipulation Traps are probably the easiest to try. A page saturated with phrases like “industry-standard” or “trusted by experts” statistically biases an agent’s synthesis in the attacker’s direction, exploiting the same framing effects humans fall for. A subtler version wraps malicious instructions inside educational or “red-teaming” framing—”this is hypothetical, for research only”—which fools the model’s internal safety checks into treating the request as benign. The strangest subtype is “persona hyperstition”: descriptions of an AI’s personality spread online, get ingested back into the model through web search, and start shaping how it actually behaves. The paper mentions Groks “MechaHitler” incident as a real-world case of this loop.
You can see examples of this in our experiment, jailbreaking Whatsapp’s AI and tricking it to generate nudes, drug recipes, and instructions to build bombs
One example of a Semantic attack. Image: Decrypt
Cognitive State Traps are another attack in which malicious actors target an agent’s long-term memory. Basically, If an attacker succeeds in planting fabricated statements inside a retrieval database the agent queries, the agent will treat those statements as verified facts. Injecting just a handful of optimized documents into a large knowledge base is enough to reliably corrupt outputs on specific topics. Attacks like “CopyPasta” have already demonstrated how agents blindly trust content in their environment.
The Behavioural Control Traps go straight for what the agent does. Jailbreak sequences embedded in ordinary websites override safety alignment once the agent reads the page. Data exfiltration traps coerce the agent into locating private files and transmitting them to an attacker-controlled address; web agents with broad file access were forced to exfiltrate local passwords and sensitive documents at rates exceeding 80% across five different platforms in tested attacks. This is especially dangerous now that people start to give AI agents more control over their private information with the rise of platforms like OpenClaw and sites like Moltbook.
Systemic Traps don’t target one agent. They target the behavior of many agents acting simultaneously. The paper draws a direct line to the 2010 Flash Crash, where one automated sell order triggered a feedback loop that wiped nearly a trillion dollars in market value in minutes. A single fabricated financial report, timed correctly, could trigger a synchronized sell-off among thousands of AI trading agents.
And finally Human-in-the-Loop Traps target the human reviewing its output. These traps engineer “approval fatigue”—outputs designed to look technically credible to a non-expert so they authorize dangerous actions without realizing it. One documented case involved CSS-obfuscated prompt injections that made an AI summarization tool present step-by-step ransomware installation instructions as helpful troubleshooting fixes. We’ve already seen what happens when humans trust agents without scrutiny.
What researchers recommend
The paper’s defense roadmap covers three fronts. The first one is technical: adversarial training during fine-tuning, runtime content scanners that flag suspicious inputs before they reach the agent’s context window, and output monitors that detect behavioral anomalies before they execute. Then there’s the ecosystem level: web standards that let sites declare content intended for AI consumption, and domain reputation systems that score reliability based on hosting history.
The third front is legal. The paper explicitly names the “accountability gap”: If a trapped agent executes an illicit financial transaction, current law has no answer for who is liable—the agent’s operator, the model provider, or the website that hosted the trap. Resolving that, the researchers argue, is a prerequisite for deploying agents in any regulated industry.
OpenAI’s own models have been jailbroken within hours of release, repeatedly. The DeepMind paper doesn’t claim to have solutions. It claims the industry doesn’t yet have a shared map of the problem—and that without one, defenses will keep getting built in the wrong places.
Daily Debrief Newsletter
Start every day with the top news stories right now, plus original features, a podcast, videos and more.
Embedded finance provider Cross River has secured a $50 million capital raise to fuel investment in AI and crypto.
Editorial
This content has been selected, created and edited by the Finextra editorial team based upon its relevance and interest to our community.
Founded in 2008 and backed by global investors, Cross River powers lending, payments, card, and crypto solutions for over 100 technology partners, including X, Upgrade, Upstart, Trustly, Stripe, Plaid, DailyPay, Coinbase, Checkout, Bill.com, Best Egg, and Affirm.
The company’s proprietary real-time banking core is the foundation for delivering embedded finance products spanning payments, lending, crypto, and capital markets—for businesses and consumers worldwide.
The funding, supported by existing investor T. Rowe Price, will be used for crypto scaling, global expansion and accelerated uptake of AI capabilities.
Gilles Gade, chairman, founder, and CEO of Cross River, says: “We thank T. Rowe Price for their support of Cross River’s strategy of embedded finance 2.0, the bundling of crypto, lending, payments, and cards on one platform with a sophisticated AI layer to deliver innovative solutions with exceptional compliance and risk management.”
The International Monetary Fund said tokenization has the potential to remove friction and boost transparency in finance, but warned that the technology could also create challenges that affect financial stability.
“The net effect of tokenization on financial stability is uncertain,” the IMF said in a 23-page report on Thursday, stating that “atomic settlement and enhanced transparency reduce some traditional risks, but speed and automation introduce new ones.”
Source: IMF
More than $27.6 billion worth of real-world assets, minus stablecoins, is currently tokenized onchain, data from RWA.xyz shows. Boston Consulting Group estimated in 2022 that the tokenization market could rise to $16 trillion by 2030, while McKinsey & Co in 2024 predicted a more conservative $2 trillion over the same time frame.
The IMF acknowledged that tokenization expands how securities and other financial products are issued, traded, settled and managed but said it shifts risks from the banking system to shared ledgers and smart contract code.
“Stress events in tokenized markets are likely to unfold faster than in traditional systems, leaving less time for discretionary intervention.”
The agency also said tokenization offers opportunities in emerging markets, such as faster cross-border payments and financial inclusion but added that it “raises the risk of volatile capital flows, rapid currency substitution, and erosion of monetary sovereignty.”
Wall Street advocates for tokenization
Blockchain tokenization has been pushed by Wall Street leaders such as BlackRock CEO Larry Fink, who is among those seeking to tokenize everything from stocks and bonds to money market funds and real estate.
The biggest RWA project by total value locked is Securitize — the tokenization platform behind the BlackRock USD Institutional Digital Liquidity Fund — at $3.38 billion, according to CryptoDep, citing data from April 1.
Tether Gold and Ondo Finance are close behind at $3.35 billion and $3.21 billion, respectively.
Source: CryptoDep
The New York Stock Exchange’s parent, Intercontinental Exchange, is also taking action, announcing in January that it would launch a tokenization platform for 24/7 trading and instant settlement of stocks and exchange-traded funds with a blockchain post-trade system.
Related: Liquidity, not novelty, determines tokenization’s value
However, the IMF said legal challenges present another obstacle, stating that without legal clarity over ownership records and settlement finality, tokenized markets risk being “fragmented and peripheral.”
The crypto industry has been developing solutions to address this problem, such as the Ethereum ecosystem’s ERC-3643 permissioned token standard, which ensures that only certain investors have access to tokenized products.
Coinbase Asset Management launched tokenized shares for the Coinbase Bitcoin Yield Fund on Ethereum layer 2 Base on March 20, with the help of financial services firm Apex Group, which implemented the ERC-3643 standard to ensure that token holder identity and eligibility were checked for compliance.
Magazine: Big Questions: Can Bitcoin save you from the dreaded Cantillon Effect?
Cointelegraph is committed to independent, transparent journalism. This news article is produced in accordance with Cointelegraph’s Editorial Policy and aims to provide accurate and timely information. Readers are encouraged to verify information independently. Read our Editorial Policy https://cointelegraph.com/editorial-policy
Metaplanet (3350) continued to scale its accumulation strategy through the first quarter of 2026, acquiring 5,075 BTC for approximately $398 million, implying an average purchase price of about $78,000 per coin.
The Tokyo-based firm has has generated a BTC yield of 2.8% year-to-date.
As of March 31, Metaplanet holds a total of 40,177 BTC, acquired for roughly $3.9 billion, with an average cost basis of approximately $97,000 per BTC.
Metaplanet is now the third largest bitcoin treasury company worldwide, overtaking MARA Holdings after the miner reduced its bitcoin stack significantly.
Twenty One Capital (XXI) holds second place with 43,514 BTC, according to Bitcoin Treasuries, while Strategy (MSTR) is by far and away the largest with over 762,000.
Shares of Metaplanet were down 2%, trading at 302 yen ($1.89).
OpenEden has introduced HYBOND, the first tokenized product tied to BNY Investments’ Global Short-Dated High-Yield Bond strategy, expanding the scope of institutional-grade investments available onchain.
The new token gives qualified investors 1:1 exposure to a managed portfolio of short-dated corporate bonds overseen by BNY Investments, a unit of BNY.
The product introduces higher-yield fixed income exposure to a market segment that has so far been dominated by tokenized cash-equivalent and treasury strategies. Data from rwa.xyz shows over $12 billion of the more than $27 billion in the tokenized real-world asset market are U.S. Treasury debt.
HYBOND is issued by OpenEden Digital Limited, a Bermuda-regulated entity licensed under the Digital Asset Business Act, according to a press release on Wednesday.
While BNY Investments serves as the investment manager for the underlying bond portfolio, it has no direct involvement in the token itself, which is managed and issued by OpenEden.
“Tokenization has proven its product market fit with cash-equivalent and treasury strategies. HYBOND represents the next step by bringing actively managed corporate bond exposure on-chain within a regulated framework,” said Jeremy Ng, OpenEden’s CEO.
BNY and OpenEden previously collaborated on TBILL, a tokenized U.S. Treasury bill product. HYBOND builds on that relationship by pushing into riskier credit instruments, which may appeal to investors seeking greater yield.
As of year-end 2025, BNY oversaw $2.2 trillion in assets under management and more than $59 trillion in assets under custody.
