BeInCrypto Global

The ETH/BTC ratio has broken out of a long-term declining channel and reached a 7-month high near 0.0334. Bitcoin dominance broke out of its own downtrend in the same week.
Both moves cannot be bullish for altcoins at once. One points to capital rotating into Ethereum. The other points to capital concentrating in Bitcoin at the expense of everything else.
ETH/BTC Reaches 7-Month High After 32% Rally
Ethereum (ETH) has gained 32.28% against Bitcoin (BTC) since the June low at 0.02525. The ratio peaked near 0.0334 last week on TradingView’s weekly Binance chart. That marks its highest level since January 2026.
That move broke the descending parallel channel that had capped the pair since August 2025. The breakout came at the end of June, and the ratio has held above the channel since.
ETH/BTC weekly chart / Source: Tradingview
Resistance now sits at 0.03213, the 0.382 Fibonacci retracement of the decline from 0.04327. Above that, the 0.5 level at 0.03426 marks the next barrier.
Support at 0.031 matters more. The level marks the April 2026 swing high for Ethereum against Bitcoin. Losing it would put the breakout in doubt.
The weekly Relative Strength Index (RSI) reads near 60 and continues to rise. Ethereum traded at $2,472 and Bitcoin at $78,827 at the time of writing, placing the ratio near 0.0313.
Bitcoin Dominance Broke Out in the Same Week
The bullish reading weakens once Bitcoin dominance enters the picture. The metric tracks Bitcoin’s share of total crypto market capitalization, and it closed last week at 60.15%.
That close cleared a descending trendline drawn from the June 2025 high. It also cleared the upper boundary of a triangle formation. Dominance gained 0.91% on the week.
BTCD weekly chart / Source: Tradingview
The last resistance sits at 60.50%, a zone that capped the metric in April and May 2026. Support rests at 58.54%, with stronger support at 58% tested in June and July.
Rising dominance and a rising ratio can coexist. Together, however, they suggest capital is moving into Ethereum and Bitcoin while smaller altcoins lose share.
Traders Are Positioned for an Altseason That Has Not Arrived
The Altcoin Season Index sits at 39, down from roughly 67 in early August. Blockchain Center derives the reading from how many of the top 50 coins beat Bitcoin over 90 days.
Altcoin season requires 75. The current figure, therefore, points away from a broad rotation.
Altcoin Season Index / Source: Blockchaincenter
Derivatives tell a different story. Glassnode reported that 85% of altcoins now show funding rates above their mean. That marks the highest print since Bitcoin traded at record levels.
“The altcoin market has entered a state of optimism. 85% of alts reached funding rates above their mean… In an alt season, these conditions can last for many weeks.”
Altcoins funding rate / Source: X
Positioning is not performance. Traders have crowded into altcoin longs while spot returns still trail Bitcoin. Meanwhile, bitcoin itself trades roughly 37% below its October 2025 record.
Historically, altcoin seasons follow new Bitcoin highs rather than drawdowns. That context weakens the case for altseason, though it does not close it.
Three outcomes remain. A weekly close above 0.03426 on ETH/BTC, paired with a rejection at 60.50% dominance, would suggest rotation has genuinely started.
A dominance break above 60.50% while the ratio stalls would mark an ether rally instead. A drop below 0.031 would suggest the entire move was a relief bounce.

Criminals now use AI every day to clone voices, write flawless phishing emails, and scale scams globally. Meanwhile, some jurisdictions ban their own investigators from touching AI tools, a Recoveris executive warns.
The warning came during BeInCrypto’s Market Intelligence Council panel on crypto crime. Sol Cinosi of Recoveris and Nick Pailthorpe of Kodex agreed on the diagnosis. The AI crypto crime fight is becoming an arms race, and one side shows up unarmed.
Grammar Checks Don’t Catch Phishing Anymore
Cinosi, a former Buenos Aires prosecutor who now leads government and corporate affairs at Recoveris, said during the experts panel that AI is changing the game for both sides. On the criminal side, it scales sophistication and reach at once.
“The old advice to look for poor grammar in phishing emails doesn’t really work anymore. AI clones voices, generates deepfakes, and automates scams at a scale that simply wasn’t possible a few years ago.”
The numbers back him up. Crypto scams took an estimated $17 billion in 2025, according to the Chainalysis 2026 Crypto Crime Report.
Scam operations with on-chain links to AI vendors extracted $3.2 million each, 4.5 times as much as those without. AI does not just make scams cheaper. It makes them dramatically more profitable.
Annual cryptocurrency scam loss / Source: Chainalysis Some Police Are Forbidden From Using AI
The other side of the race looks very different. Cinosi named it as his single biggest worry.
“My biggest concern right now is law enforcement, because many agencies are still behind in AI adoption. Some don’t yet have clear policies, and in certain jurisdictions, investigators are forbidden from using AI tools.”
The result is an asymmetry that regulators are only starting to acknowledge. Europol’s IOCTA 2026 report calls it a “velocity gap” and says enforcement now depends on bridging it with technology. Cinosi put it more bluntly.
“If criminals are using AI every day and investigators, particularly law enforcement, aren’t, we are creating an uneven playing field. That’s not a race we’re going to win.”
Notably, he argued the barrier is often softer than a formal ban. Many investigators simply fear the tools, assuming they lack permission to use powers they already have. In his view, the gap is as much about capacity building as regulation.
AI as the Investigator’s Force Multiplier
The irony is that tracing technology has never been better. Cinosi said Recoveris now follows funds across blockchains, bridges, and even mixers with high confidence. Technology rarely stops an investigation anymore. People do.
“For investigators, AI helps solve a capacity problem by processing enormous amounts of data faster and identifying patterns and connections that would take much more time to find manually.”
That capacity matters because crypto adoption is growing faster than the number of experts. Experienced blockchain investigators remain scarce, so every hour AI saves goes where machines still fail.
“It allows us to spend more time doing what humans still do best, which is applying judgment and making critical decisions.”
AI Crypto Crime: The Training Gap on the Front Line
Pailthorpe, who spent 20 years in UK policing before joining Kodex, sees the same gap from the street level. Crypto no longer appears only in crypto cases. He meets it daily in counterterrorism and human trafficking investigations.
“What I’m seeing is more and more investigators going out to more and more scenes and realizing that crypto is part of that investigation, and not necessarily knowing what to do when they come across a wallet.”
The industry response is constant training. Pailthorpe runs webinars and police station visits almost daily, while exchanges on Kodex provide educational material to verified officers. Still, he refused to end on alarm.
“There’s that challenge of the criminals being one step ahead, but I am confident in lots of police agencies around the world.”
The tools exist, and the trainers are working. Whether policy lets investigators use the same weapons as their adversaries may decide who wins the race.

Nvidia (NVDA) earnings land on Wednesday, August 26, and the stock fell 2.91% to $208.48 on Monday, its seventh straight losing session before the report.
The slide looks strange because prediction markets give the report a near-certain beat. However, money flow, options, and betting data point to the real answer. Traders do not fear the result. They fear what follows it.
Big Money Is Leaving Every Chip Stock
The clearest warning sits in Chaikin Money Flow (CMF), a proxy for institutional buying and selling pressure. As of August 25, the 20-day CMF reads negative for all ten major chip names, which suggests big money may be quietly exiting the sector. The sector-wise institutional outflow trend, right before Nvidia earnings, is no coincidence.
Additionally, Nvidia sits at the bottom of the table at -0.155, weaker than AMD at -0.149 and TSMC at -0.06. Nvidia leads the group, and its seventh straight drop, the longest losing streak since 2022, has spread caution across the sector.
Chip Stocks Money Flow Readings: Charlie Quant Lab
Outflows alone, though, do not prove traders expect a fall. The options market does.
Nvidia Options Traders Are Paying for Crash Insurance
Barchart data shows Nvidia’s put-call ratio for volume climbed from 0.52 on August 12 to 0.65 by Monday’s close. The ratio compares bets on a fall (puts) against bets on a rise (calls), so a climb means more downside protection.
Meanwhile, the open interest ratio held near 0.82, so standing positions have not changed. Options also price a 7% post-earnings swing, far above the 2.8% average move of the past four quarters.
NVDA Put Call Ratio Chart: BeInCrypto
That hedging makes more sense next to the betting data, where the result itself is treated as settled.
Polymarket Says the Beat Is Already Decided
Polymarket traders price a 97% probability that Nvidia beats estimates, extending a streak of 14 straight beats. Consensus sits near $92 billion in revenue, above management’s $91 billion guide, as covered in BeInCrypto’s earlier Q2 earnings preview.
The real disagreement is the data center revenue number. The contract for above $85 billion repriced from 55% to 91% in one August 15 session, then eased to 79% by Tuesday. Above $90 billion trades at 23%, leaving over half the probability inside a $5 billion band.
Polymarket NVDA Earnings Odds: BeInCrypto
History explains the caution. Nvidia fell after each of its last four earnings beats, with drops between 0.79% and 5.46%.
Nvidia Beat And Retreat: BeInCrypto
Yet Wall Street has not moved an inch.
Wall Street Refuses to Blink
TipRanks shows all 27 top-ranked analysts rating Nvidia a Buy, with a $304.13 average price target that implies roughly 46% upside. On Monday alone, four firms reiterated Buy calls.
NVDA Analyst Ratings Forecast: TipRanks
KeyBanc set $330, Rosenblatt $325, J.P. Morgan $280, and Cantor Fitzgerald $350, echoing an earlier $350 price target from Bank of America.
Nvidia Buy Ratings: BeInCrypto
In contrast, Michael Burry, the investor who predicted the 2008 housing crash, says Nvidia anchors a circular financing web. In plain terms, AI firms fund each other’s chip orders, so demand can look stronger than it is. Wednesday’s number must settle the gap between those views.
Nvidia Stock Levels to Watch on Wednesday
The daily chart frames the trap. Monday’s drop broke the 0.382 Fibonacci retracement at $211.63. Sell volume has also surged since August 14, the same window during which the fresh hedges went on.
The line to defend is $201.59, the 0.618 retracement, just 3% below Monday’s close and well inside the 7% implied move. That is the slip, the money flow, and options data appear positioned for, even after a beat. Losing it opens $194.45 and the $185.35 floor.
Nvidia Price Analysis: TradingView
A data center print above $90 billion could surprise and put $227.88 back in play.
NVDA Scorecard: BeInCrypto
After the Nvidia earnings print, $201.59 separates a rebuild toward the $236.46 record from the breakdown that options traders have already paid to insure against.
Analyst’s View: Sellers are not waiting for Wednesday. Seven straight red closes on rising volume points to distribution rather than hedging. A break under the $206 zone, immediate support, on the report would likely trigger profit booking across the chip sector.

JPMorgan kept its $240 target on SpaceX (SPCX) on Tuesday, nearly 80% above Monday’s $135 close. Analyst Doug Anmuth’s reason has little to do with rockets.
The bank is betting on artificial intelligence (AI). At $240, SpaceX would be worth over $3 trillion. Grok, its AI model, and newly acquired Cursor would carry much of that weight.
SpaceX (SPCX) Stock Performance. Source: Yahoo Finance Why JPMorgan Calls SpaceX Stock an AI Bet
SpaceX closed its Cursor purchase on August 14. The coding platform generated about $4 billion in annual recurring revenue as of June. Roughly 75% of it came from business customers.
That matters for two reasons. Companies already paying for Cursor can be sold Grok next. Meanwhile, millions of real coding sessions become training data.
SpaceX wanted a coding asset badly. It also approached Cognition, a rival startup that rebuffed its buyout approach this month.
“SpaceX’s AI ambitions are coming into sharper focus, and we are increasingly positive on Grok,” Anmuth wrote in a note.
The results are already showing, JPMorgan says. Cursor data now feeds Grok’s supplemental training, and recent model performance has visibly improved.
$SPCX – JPMORGAN BULLISH ON SPACEX AS GROK MOMENTUM BUILDSJPMorgan reiterated an Overweight rating and $240 price target on SpaceX, citing growing confidence in Grok’s AI prospects.The firm highlighted SpaceX’s acquisition of Cursor, which reportedly generates around $4…
— *Walter Bloomberg (@DeItaone) August 25, 2026
Grok 4.6, Grok Bot, and a September Share Unlock
Grok 4.6 shipped on August 12. Its debut added $500 billion to SpaceX’s market value.
JPMorgan says the model now sits on the Pareto frontier. In plain English, nothing smarter costs less, and nothing cheaper is smarter. That mix tends to win developers and enterprises.
More is coming. The bank expects new models almost monthly through December, with Grok 5 due before year-end. It also flags the Grok Bot launch, an agent that handles workplace tasks for enterprise customers.
Elon Musk has pledged SpaceX engineering data for future training runs. JPMorgan believes more than two decades of rocket-building knowledge could give Grok an edge in real-world engineering.
The bet is not risk-free, however. SpaceX’s AI unit lost $1.26 billion last quarter and absorbed 86% of capital spending.
Supply is the nearer test. Nearly 370 million shares unlock on September 9 and 10, which could swell the float by about 20%. SPCX has struggled since its record June debut, trading for $137.85 as of this writing.
JPMorgan expects enterprise Grok sales to become a bigger slice of SpaceX’s AI revenue over time. The near-term question is simpler. Can the AI story outweigh a wave of new sellers?

Lisk relaunched on Tuesday as a money operations platform for finance teams, merging bank and stablecoin balances in one workspace. The new Lisk enters a fintech market where rivals hold billion-dollar war chests.
Founder Max Kordek unveiled the product as the Lisk Chain heads for an October 31 shutdown. Early Access opened the same day for businesses handling both fiat and stablecoins.
What the New Lisk Actually Does
The platform puts accounts, payments, and approval rules in one workspace across entities and currencies. A bank transfer and a stablecoin deposit land as one balance. Businesses receive virtual accounts with real bank details and can pay out to external bank accounts.
Lisk does not become a bank. Money moves through regulated providers, including Bridge, a Stripe company. The platform is free on its Professional plan through 2026, a sign Lisk is buying adoption before charging for it.
“It’s the product we wish we had years ago,” Kordek wrote in Tuesday’s launch announcement.
The market behind the pivot is real. B2B stablecoin payments hit $226 billion in 2025, up 733% in a year, per a McKinsey and Artemis Analytics study.
Corporate treasuries have spent the past year replacing wires with stablecoins for cross-border settlement.
The Early Access page does not disclose licensing, custody arrangements, or what the product will cost after 2026.
The Competition Has a Head Start Worth Billions
Lisk’s pitch lands in fintech’s most crowded lane. Ramp raised $750 million in June at a $44 billion valuation. Stripe paid $1.1 billion for Bridge, the very provider Lisk routes money through.
Stripe acquired Bridge for $1.1 billion.I'll explain why, for a company, you probably never heard of 🧵
— Yash (@yashhsm) October 22, 2024
Lisk’s key supplier, in other words, belongs to a rival.
Kordek argues the incumbents built for fiat first or crypto first, never both. The new Lisk targets the multi-entity company holding fiat and stablecoins side by side.
The premise carries risk, however. Incumbents can add stablecoin rails faster than a newcomer can win the trust of finance chiefs. The banking wedge may also narrow.
The US Federal Reserve has proposed direct payment accounts for crypto firms, easing the exclusion Lisk is built around.
There is also a résumé problem. Lisk is asking businesses to trust it with treasury and payments weeks before retiring its second blockchain. Winning CFOs may prove harder than winning developers ever was.
After 10 years, former $4B crypto project Lisk is shutting down its blockchain to pivot into a software business.A wind-down proposal seeks to burn 100M $LSK (25% of the total supply) and dissolve the DAO. The token will now transition into a loyalty asset on Base and… pic.twitter.com/n7pE6MUYsk
— BeInCrypto (@beincrypto) August 25, 2026
The closest precedent is not encouraging either. EOS raised about $4 billion in crypto’s biggest token sale. It rebranded to Vaulta in March 2025 to chase Web3 banking.
Vaulta’s token is down 85% over the past year, per Coingecko data.
Vaulta (A) Price Performance. Source: Coingecko What LSK Holders Get in the New Lisk
LSK becomes the platform’s loyalty token. Businesses earn rewards for using Lisk and for referrals, rolling out in phases, per the token FAQ. Paying fees in LSK comes later, with no date attached.
The design leaves open questions. The announcement describes rewards and fee payments, not revenue sharing, and governance itself ends with the DAO.
Holders now own exposure to a startup without any of a shareholder’s rights.
The DAO treasury tells another story, that after the 100 million LSK burn, roughly 47 million LSK moves to Lisk Ltd. This is according to the cessation documents. Holders vote to dissolve the DAO, and the company inherits what is left.
The scale gap is stark. LSK traded near $0.08 as of this writing, down 5% on the shutdown news. With a market cap of about $20.3 million, the project is roughly 0.05% of Ramp’s private valuation.
Lisk (LSK) Price Performance. Source: BeInCrypto
The next signals are the DAO vote and the first Early Access cohorts. If businesses show up, LSK gets its first real utility in years. If they do not, holders own loyalty points to a product nobody adopted.
The new Lisk has left itself no chain to fall back on.

Bitcoin (BTC) tested $80,000 this week, and three of the largest meme coins rallied, followed by outsized weekly candles. Only Dogecoin (DOGE), however, broke the resistance that mattered.
BTC trades near $80,700 after a weekly gain of roughly 25%. Dogecoin, Pepe (PEPE), and Pump.fun (PUMP) each posted double-digit gains, yet their weekly charts sit at very different points.
Meme Coins Rally: Dogecoin Breaks a 20-Month Downtrend
DOGE gained 34.22% last week, rebounding from the support zone near $0.0556 that had held since late June. The move broke the descending trendline drawn from the December 2024 high at $0.485.
That line had capped every rally for 20 months. Weekly volume on the breakout was the heaviest since May 2026, which suggests real participation rather than a thin bounce.
DOGE weekly chart / Source: Tradingview
The weekly Relative Strength Index (RSI) sits near 50. That neutral reading indicates the advance consumed almost none of the available upside.
DOGE now trades at $0.0929, retesting the broken line from above. A weekly close above it would confirm the end of the downtrend flagged in July. The next resistance stands at $0.1476, the 0.786 Fibonacci retracement, roughly 59% higher.
PEPE Is Repeating a Setup That Already Failed
PEPE rose 59.53% last week, a larger gain than DOGE produced. The chart, however, has shown this pattern once before.
In December 2025, PEPE printed a 78.18% weekly candle that also broke its descending trendline. Price then faded and spent more than seven months trapped between $0.00000223 and $0.0000044.
PEPE weekly chart / Source: Tradingview
PEPE now trades at $0.00000425, up 6.1% over 24 hours and still inside that range. Weekly RSI reads near 56 and volume spiked, so buyers are clearly present. The token is testing range resistance for the second time rather than breaking it.
Above $0.0000044 sits the December 2025 swing high at $0.00000726, then the 0.236 Fibonacci level at $0.0000084. Heavy exchange outflows recorded in August suggest holders are willing to wait.
Pump.fun Leads the Gains but Is Already Stalling
PUMP delivered the strongest move of the three, gaining 95.84% last week. That was the best weekly performance among top meme coins.
The rally runs deeper than one candle. PUMP has climbed roughly 370% from its July low at $0.001151. It cleared the 0.236 and 0.382 Fibonacci levels in a single week, on its heaviest volume since February.
PUMP weekly chart / Source: Tradingview
Momentum has now turned. PUMP trades at $0.005018, down 6.2% over 24 hours after rejecting the 0.5 Fibonacci retracement at $0.005066. Weekly RSI peaked at 77 and has eased to about 70, leaving PUMP the only overbought name of the three.
The golden pocket near $0.005989 caps further upside, while $0.002999 marks first support. The token remains the clearest beneficiary of the current meme coin season, and also the most extended.
Traders now watch three levels. DOGE needs to hold its trendline retest, PEPE needs a weekly close above $0.0000044, and PUMP needs to defend $0.002999. With capital also rotating into RWA tokens, meme coin strength may prove selective rather than broad.

Lady Gaga AI biotech venture Outer Bio stepped out of stealth this week. The startup keeps donated human skin alive for four weeks, then trains machine learning models on how that tissue reacts.
Michael Polansky, the singer’s fiancé, co-founded the company and runs it as chief executive. Gaga, born Stefani Germanotta, sits on the board. Backers have committed roughly $23 million.
Lady Gaga AI Biotech Venture Turns Surgical Waste Into Training Data
Skin removed from the body normally dies within a week. Outer Bio’s platform, named Yuna, stretches that window to four weeks. Researchers can therefore track slow biology such as collagen breakdown, inflammation, and cellular aging.
Tissue reaches the lab within hours of surgery, mostly leftover material from cosmetic procedures. Donors span every age group, both sexes, and all six Fitzpatrick skin tones.
That matters because regulators have turned against the old default. The FDA issued a roadmap in April 2025 to cut unnecessary animal testing, then reported first-year progress this April. More than 90% of drugs that clear animal studies still fail in human trials, the agency notes.
Europe moved earlier. A full marketing ban has covered animal-tested cosmetics across the bloc since March 2013. Brands selling there have needed human alternatives ever since.
Polansky argues that simplified lab models flag toxicity without showing whether a compound actually works. Living tissue answers the harder question, because the drug meets the organ it targets. Slow damage and slow repair both become visible.
So far, the platform has logged more than 10,000 treatments across 300 donors, and each sample yields over 30,000 measurements. Published validation work shows Yuna reproducing inflammation, aging, and sun damage.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights
Silicon Valley Bets on Data It Cannot Compute
Polansky started the company quietly in 2022, after years running Sean Parker’s family office. He argues that biology, not compute, now caps progress in AI.
“AI models are improving far faster than the experiments that validate and train them.”
Michael Polansky, co-founder and CEO of Outer Bio, in the company’s announcement.
Investors in the Lady Gaga AI biotech startup include Wing Venture Capital, Initialized Capital, and SV Angel. Lightspeed Venture Partners joined too, alongside Polansky’s own firm, Hawktail.
That sum stays small next to the rest of the field. Chai Discovery collected $400 million in July for AI molecular design. OpenEvidence took $250 million in January.
Lady Gaga AI biotech startup Outer Bio against 2026 health AI funding rounds, Source: BeInCrypto
Scarce biological data keeps drawing that money. Pfizer and Anthropic already treat AI in healthcare as core infrastructure, while crypto fortunes bankroll crypto-funded anti-aging studies. Meanwhile, Midjourney’s medical scanner chases a billion body scans a year.
Ownership of that data matters as much as the models. Elon Musk feeds SpaceX engineering data into Grok, and Amazon enrolled Twitch creator archives into AI training by default.
Outer Bio now sells platform access to biopharma teams and consumer brands. However, the harder test comes later. Skincare has gained just one proven anti-aging mechanism in three decades, and richer data alone will not change that.

Prediction markets have become one of the fastest-growing fronts in US trading, and sports contracts are now at the center of a legal fight over regulatory control.
On July 31, New York sued Kalshi in state court, seeking more than $36 billion in damages and an order barring the exchange from offering event contracts nationwide.
On August 11, the Commodity Futures Trading Commission responded with emergency authority after Kalshi notified the agency of a market emergency, arguing federally regulated derivatives markets require national treatment.
On August 12, the New York City Council opened an investigation into the marketing practices of several prediction-market companies, including Polymarket, Kalshi, Coinbase and Gemini Titan, examining allegations involving young users, influencer marketing and misleading promotions.
CME has also sued the CFTC over its approval of perpetual futures for Kalshi and Coinbase, alleging competitive injury from rules opening new products to rivals.
Indeed, exchange competition, sports law and federal preemption are increasingly overlapping the same market.
BeInCrypto spoke with Fernando Lillo Aranda, CMO at Zoomex, and Federico Variola, CEO of Phemex, about regulatory classification, the role of sports contracts, retail safeguards and growing competition between exchanges.
The Product-Specific Legal Boundary
US law gives the CFTC authority over designated contract markets, while sports contracts have created a difficult boundary with state gambling law. In June, the CFTC proposed a new framework for event contracts involving gaming and other enumerated activities under the Commodity Exchange Act. The proposal would allow a 90-day review and assess products case by case using public-interest factors.
Court rulings remain divided:
The Third Circuit sided with Kalshi in its dispute with New Jersey, ruling that federal law preempts the state’s attempt to block the exchange’s event contracts.
Judges in Nevada, Maryland and Ohio allowed state gaming authorities to continue enforcement actions against Kalshi while the underlying legal disputes proceed.
A Tennessee court also ruled in Kalshi’s favor, leaving a federally regulated market subject to different judicial interpretations across states.
One way to draw the line is by looking at what the contract is actually used for:
Contracts linked to payroll data, inflation or commodity prices fit within derivatives markets because businesses and investors can use them to manage financial risk;
Sports contracts have a different purpose, with most participants trading based on the outcome of a game rather than hedging an existing exposure.
Kalshi operates as a CFTC-regulated designated contract market, while sportsbooks fall under state gambling laws. Although many retail users approach both products in similar ways, the legal framework remains different.
Retail Protection Depends on Education as Much as Regulation
Fernando Lillo Aranda, CMO at Zoomex, believes prediction markets can protect retail users without restricting access. He argues that exchanges should focus on transparency, market integrity and helping users understand the products they trade.
“The priority should be transparency, fair market design, and clear risk disclosure. Ultimately, the objective should be to empower users through knowledge rather than limiting innovation.”
Lillo Aranda also says sports, politics, weather and live-event markets present additional risks around manipulation, misinformation and emotionally driven trading because many newcomers approach them as entertainment rather than financial products.
“Many first-time participants approach prediction markets as entertainment rather than as financial products with real risk. Financial literacy is one of the strongest forms of consumer protection. Well-regulated prediction markets can coexist with strong user safeguards if exchanges combine transparency, responsible product design, and continuous education with high standards of market integrity.”
Exchange Competition Within the Regulatory Debate
Federico Variola, CEO of Phemex, believes the legal disputes extend beyond consumer protection and into competition between exchanges.
“Perpetual futures and event-based contracts have emerged as some of the most successful and prominent financial instruments in recent years, and controlling access to these products is obviously extremely lucrative.”
Variola says established exchanges and newer platforms are increasingly competing over who gets to offer these products, with regulation becoming part of that battle.
“We have seen platforms such as CME and Kalshi rely on regulation to maintain a degree of exclusive access to certain products. At the same time, there is a strong push toward decentralization from platforms such as Hyperliquid, Lighter, and others.”
The commercial incentives are substantial. CME says its event contracts surpassed 100 million trades within eight weeks of launch, while Coinbase has integrated Kalshi’s event contracts into its own platform. At the same time, CME has challenged the CFTC in court over approvals allowing Kalshi and Coinbase to offer perpetual crypto futures, arguing the regulator unfairly expanded access to competing products.
Variola expects centralized and decentralized exchanges to continue developing in parallel rather than one replacing the other.
“I believe the decentralization of these products is important, particularly because they have already become familiar to the vast majority of crypto users. I hope future regulators take this into consideration when deciding how these instruments should be approached.”
A Fight Over Market Definition
Prediction markets have become a test of how US regulators classify event contracts. Sports contracts present the most difficult example because one product can trade on a regulated exchange while attracting users for reasons associated with sports betting.
Competition between exchanges adds another element. CME, Kalshi, Coinbase and decentralized platforms are all seeking a share of the event-contract market, while state gaming authorities continue to defend their licensing and consumer-protection powers. Each regulatory decision influences both market access and the balance between federal and state oversight.
Future rules are likely to focus on individual contract types, supported by consistent settlement standards, market-surveillance requirements and consumer protections suited to each category. Sports prediction markets have grown into a significant segment of the derivatives industry, giving these decisions lasting importance for exchanges, regulators and market participants.

Germany’s foreign intelligence service is hunting hackers at Gamescom. The Bundesnachrichtendienst (BND) opens its career stand at the Cologne gaming expo on Wednesday, where visitors can work through a five-step cybersecurity challenge.
Opening Night Live starts the week tonight. Meanwhile, the show floor opens Wednesday and runs through Sunday, hosting more than 1,600 exhibitors from 67 countries.
Why the BND Hunts Cyber Talent at Gamescom
The agency wants programmers, analysts, and cyber specialists. Therefore, it goes where those people already spend their weekends. Roughly 350,000 visitors passed through the halls last year.
The exhibitor list explains the pull. Xbox, Nintendo, Sony, and Electronic Arts (EA) all take floor space, alongside HoYoverse, CD Projekt Red, SEGA, Bandai Namco, and NetEase Games.
Highlights Gamescom 2026. Source: Gamescom Global
Featured premieres run from Persona 4 Revival and Marvel Rivals to Nintendo Switch Sports Resort. Consequently, the BND fishes in a pool of several hundred thousand technically fluent visitors.
Organizers sold out the show floor for the first time, filling 233,000 square meters. Saturday tickets vanished before the doors even opened.
🌍 Wherever you are in the world, there’s a time for you to tune in!#gamescom #OpeningNightLive is happening August 25 — check the times below and set your reminder. ⏰🇺🇸 11 AM PT / 2 PM ET🇧🇷 3 PM BRT🇬🇧 7 PM BST🇪🇺 8 PM CEST🇹🇭 1 AM ICT (Aug 26)📺 Don’t miss the show.… pic.twitter.com/mu3Tsvf9qm
— gamescom (@gamescom) August 21, 2026
“We see a significant overlap between gamers and key talent we are looking for: they are tech-savvy, enjoy stepping into different roles.” Julia Linner, BND press spokesperson, as reported this week
The timing matters. European services now face state-backed intrusion campaigns that grow faster than their hiring pipelines. Chinese groups, for instance, doubled their attack volume after handing routine work to artificial intelligence (AI) tools.
North Korea illustrates the other half of the problem. Researchers recently built a fake DeFi startup and hired three suspected North Korean developers to watch them from the inside.
What the 5-Step Cyber Challenge Actually Tests
Reports on the agency’s plans describe five stations built around practical security tasks. Recruiters watch how candidates reason, rather than simply who finishes first.
The pitch itself stays plain. The agency’s own event listing invites visitors to stop by and discover career options at Germany’s foreign intelligence service.
The competition sits a few halls away, however. Xbox, Nintendo, and EA hunt the same engineers, and a federal pay scale rarely outbids them. So the agency sells the work instead of the package.
Defensive talent stays scarce across finance as well. Crypto platforms lost $75.87 million across 40 hacks in June, security firm PeckShield reported. Binance’s security chief similarly points to phishing and stolen credentials, not quantum computers, as the tools draining wallets today.
Gaming supplies its own warning, however. About 93% of Web3 gaming projects now sit effectively dead, a 2026 study by trading firm Caladan found. Technical talent clearly moves fast when a sector stalls.
The BND never publishes hiring figures, so the results will stay invisible. Still, the stand shows how openly Germany now competes for the same people that criminal networks keep courting. Whether other European services book their own Gamescom stand next year will show if the experiment worked.

On June 5, 2025, a Tron wallet held $37.3 million in USDT when Tether began blacklisting it. The freeze took 5.7 minutes to pass through the issuer’s multisignature wallet. Two minutes before the final approval, the entire balance moved.
So, there is always a dramatic race between networks like Tether and criminals, often decided in seconds. Latest research from blockchain analytics firm BitOK found that Tether now completes a typical USDT blacklist faster than a year earlier.
The first approval can still reveal the target address on-chain before enough signers complete the freeze.
BitOK classified 107 events as clean interceptions in the latest 12-month period, with $127.6 million leaving targeted addresses between submission and execution. That was 6.1 times the gross amount measured in the preceding year. The figure counts activity across blacklist events and may include the same principal more than once.
How Tether Races to Freeze Illicit USDT
The stakes extend well beyond one analytics report. USDT is the world’s largest stablecoin, with a current market value of about $183 billion. Issuer-controlled freezes have become part of law-enforcement recoveries.
The US Department of Justice credited Tether’s assistance in a $225.3 million fraud case in 2025, while the Tether-backed T3 Financial Crime Unit says it has frozen more than $300 million across 23 jurisdictions.
A USDT Freeze Starts With a Public Warning
Tether can stop USDT held at an address by adding it to the token contract’s blacklist. On Ethereum and Tron, the order passes through a multisignature wallet: three of six owners must approve on Ethereum, while two of three approve on Tron.
The first signer submits the address. The remaining signers confirm it. Until the approval threshold is reached, the address and pending action are public, while the USDT can still be transferred.
BitOK examined activity from May 1, 2024, through May 9, 2026. Across 7,562 blacklisted addresses, the median submission-to-execution window fell from 3 hours 10 minutes to 1 hour 46 minutes on Ethereum.
Tron’s median dropped from one hour and 57 minutes to one hour and 30 minutes.
Neither multisig changed its owners or approval threshold during the study. The improvement came from faster coordination between signers, leaving the underlying sequence intact.
Figure 1. Median time between a blacklist submission and execution, by month. Source: BitOK The Sixfold Jump Comes With a Counting Problem
BitOK used a strict definition for a clean case: at least 95% of the starting balance left during the window, and no more than 5% remained when the freeze executed. It found 93 such events on Tron, representing $75.25 million in gross outflows, and 14 on Ethereum worth $52.38 million. The earlier period produced $20.9 million across 62 events.
These are event totals, rather than unique losses. In one Ethereum sequence, roughly $3.83 million moved through several addresses as Tether followed it. Five blacklist events generated $19.14 million of measured outflow even though BitOK placed the directly observed principal between $3.83 million and $7.66 million.
On-chain timing also cannot identify every actor or prove why each transfer occurred. The strongest evidence for automation comes from repeated movements that beat final signatures by seconds. The dataset and calculation scripts are public on GitHub, allowing the classifications to be checked independently.
Figure 2. Longer windows had more outflows, while several large transfers occurred in short windows. Source: BitOK Even Five Minutes Can Be Enough
The $37.3 million Tron case shows how little time a prepared transfer needs. The USDT left the targeted address and reached the SunSwap V3 router, where BitOK traced swaps that delivered about 133.7 million TRX. Once converted, Tether’s USDT contract could no longer freeze those funds directly.
Ethereum produced a tighter race. In July 2025, five related transfers of about $3.83 million each moved 24 to 96 seconds before separate freezes executed. Another address moved $27.12 million with 528 seconds to spare.
Longer waits increased the chance of some outflow on Tron. Large dollar movements clustered at the other end: $44.94 million of the $75.25 million in clean Tron outflows occurred in windows shorter than one hour.
Figure 3. Monthly lifecycle of USDT targeted for blacklisting on Tron. Source: BitOK Figure 4. Monthly lifecycle of USDT targeted for blacklisting on Ethereum. Source: BitOK Tether Has Already Tested a Faster Lane
The same dataset shows that Tether can nearly remove the wait when an operation is coordinated in advance.
In March 2026, Ethereum’s median window fell to zero minutes, with some submissions and executions landing in the same block. Tron’s median fell to 1.6 minutes that month. Over the past year, about 16% of Ethereum freezes and 17% of Tron freezes were executed in under two minutes.
BitOK calls this an urgent mode and says the pattern is consistent with off-chain preparation, although the blockchain cannot reveal Tether’s internal process.
The study also found seven instances in 2025 where blacklisting and destruction occurred in one block, totaling $32.02 million. Almost all of that came from one $31.77 million operation.
The research recommends collecting signatures privately before submitting an atomic transaction, or creating a formal emergency route for sensitive cases. Tether has already shown that faster coordination is possible.

The cryptocurrency market has climbed roughly 22% over the past week, with major assets such as Bitcoin (BTC) and Ethereum (ETH) reaching multi-month highs.
Improvements across several demand indicators have accompanied the rally. However, each measure still has a caveat, suggesting the recovery has strengthened but has yet to receive broad confirmation.
Stablecoin Flows Move Toward a Net Inflow Shift
One sign of improving liquidity is emerging from stablecoin flows. Stablecoin netflows to exchanges indicate how much stablecoin liquidity is potentially available for trading. Sustained inflows can increase the pool of capital available to buy crypto, while outflows reduce that immediately available liquidity.
Analyst CW8900 said net inflows began to fall after April. Outflows dominated through the months after that, and BTC slid toward roughly $58,000.
That pattern has since reversed, according to the analyst. Outflows have shrunk, inflows have begun, and a full shift to a net inflow trend is approaching.
“As long as funds flow in, the market will maintain a bullish trend. And when the funds inflow trend stops, it will undergo a correction,” the post read.
Follow us on X to get the latest news as it happens
ETF Inflows Broaden Across Major Altcoins
Institutional demand has also picked up as the market recovered. On August 24, spot Bitcoin funds absorbed $337.56 million, while Ethereum products added $115.57 million, continuing a streak of inflows.
Smaller categories joined the move. Solana (SOL) funds drew $33.49 million, their largest daily total since December 15, 2025, per SoSoValue data. XRP (XRP) products took $13.82 million.
Those daily figures extend a week that BeInCrypto reported as the strongest for Bitcoin and Ethereum funds since October 2025.
However, analyst Darkfost measured that recovery against a much deeper hole. The analyst noted that the recent inflows have yet to offset the broader trend. ETFs remain net sellers for 2026, with their holdings down by roughly 92,000 BTC since the start of the year.
📊 While Saylor has paused his purchases for several weeks now, ETFs are continuing to accumulate.Over the past 30 days, ETF demand has grown by roughly 26 000 $BTC.This marks a clear progression, helping to structure short-term demand.👉 That said, since the start of 2026,… pic.twitter.com/02CV3fuexA
— Darkfost (@Darkfost_Coc) August 24, 2026
US Crypto Demand Drought Narrows but Stays Below Zero
The Coinbase Premium Index provides another test of the rally. The index measures the price difference between an asset trading on Coinbase Pro and Binance.
Positive readings indicate US buyers are paying up, while negative readings indicate the bid is fading. Both major assets have climbed toward zero.
Bitcoin Coinbase Premium Index. Source: CryptoQuant
Ethereum sat at -0.004 and Bitcoin at -0.014, recovering from roughly -0.10 in mid-August. Yet, neither has crossed the line.
CryptoQuant data shows both readings have stayed negative since early May. Moreover, history warns against reading too much into a single flip. The Bitcoin premium reached about 0.0027 in early May before declining again.
Taken together, the indicators show a market with improving liquidity, stronger ETF demand, and recovering US buying pressure. But with stablecoin flows only approaching a sustained inflow, ETFs still negative for the year, and Coinbase premiums below zero, the rally has not yet secured confirmation from all major demand measures.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Analyst and former banker Felix Prehn published a thread on August 24 arguing that most retail investors missed the 500% to 1,000% gains already seen in Palantir, Intel, and Seagate.
He now points to four companies sitting at what he calls the true bottleneck of artificial intelligence: electricity.
This article is not financial advice. Stock prices are volatile, past performance does not guarantee future results, and readers should conduct their own research or consult a licensed advisor before making any investment decision.
Why Power Companies Became the New AI Trade
Prehn’s thesis centers on a structural shortage rather than a speculative narrative. Big Tech companies are signing long-term contracts to secure nuclear and other generation capacity, even as hundreds of billions of dollars pour into chips and data center construction.
If you missed Palantir at $20, Intel at $45 or Seagate at $95, you're not alone.Each stock ran 500-1000% in the next 12 months, and most retail investors missed it.I went hunting for the next stocks that could go on a similar run as these, and came across these 4:🧵
— Felix Prehn 🐶 (@felixprehn) August 24, 2026
Without reliable, clean power, he argues, spending cannot fully materialize into operating capacity. The four companies below sit directly in the path of that demand, each tied to concrete contracts with major AI infrastructure buyers rather than speculative exposure to the sector.
Constellation Energy (CEG)
Constellation owns the largest nuclear fleet in the United States, including Three Mile Island, which was restarted under the name Crane Clean Energy Center. In the second quarter of 2026, the company signed 920 megawatts of new long-term nuclear power purchase agreements, averaging 18.5 years in duration, including a deal with Walmart.
Management raised its adjusted operating earnings guidance to $11.50 to $12.50 per share, with the CEO describing existing plants as the bedrock for powering data centers during this early phase. Shares trade near $273, down roughly 34% from a 52-week high of $412.70, according to TradingView data.
Follow us on X to get the latest news as it happens.
Constellation Energy (CEG) Price Performance. Source: TradingView Talen Energy (TLN)
Talen owns the Susquehanna nuclear plant and holds a major long-term contract with Amazon Web Services covering up to 1,920 megawatts. In its second-quarter results, the company raised adjusted EBITDA guidance to $2.025 billion to $2.225 billion and free cash flow guidance to $1.2 billion to $1.35 billion.
Talen also closed its Cornerstone acquisition and advanced a pipeline of roughly 4 gigawatts in data center options. Shares trade at $305, correcting from an all-time high near $451 reached in October 2025.
Talen Energy (TLN) Price Performance. Source: TradingView Vistra (VST)
Vistra holds long-term contracts with Meta and Amazon and recently launched Helix Digital Infrastructure alongside NVIDIA, KKR, and the Kuwait Investment Authority, with an initial commitment of up to $1 billion. In the second quarter, the company posted more than 30% growth in ongoing operations adjusted EBITDA and reaffirmed its full-year 2026 guidance.
It also received FERC approval for its acquisition of Cogentrix. Shares trade near $135, well off a 52-week high of $219.82.
Vistra (VST) Price Performance. Source: TradingView GE Vernova (GEV)
GE Vernova sells gas turbines, generation equipment, and grid infrastructure, with a backlog that reached $176 billion. AI data center orders more than doubled during the first half of 2026 compared to all of 2025, while its gas turbine backlog hit 116 gigawatts, with management expecting to surpass 125 gigawatts by year-end.
The company recently launched a new MV-UPS system built specifically for AI factories and signed battery storage contracts in Australia. Shares trade around $942, near an all-time high of roughly $1,196 reached in July 2026.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights.
GE Vernova (GEV) Price Performance. Source: TradingView What Investors Should Weigh Before Following This Thesis
All four companies share the same underlying catalyst: growing, contracted demand for the clean, reliable power that AI infrastructure requires. That structural setup differentiates them from purely speculative AI plays tied to chip demand or software hype alone.
Prehn himself flagged the key risk directly. If AI spending slows, or if the profits these buildouts assume fail to materialize on schedule, share prices across this group could fall sharply, given how much of their recent value already reflects future expectations.
He recommended having a clear exit strategy in place before entering any of these positions. As with any concentrated thematic bet, diversification and position sizing matter as much as the underlying thesis itself.
None of the information above constitutes financial advice. Readers should independently verify current prices, company fundamentals, and risk factors, and consider consulting a licensed financial advisor before making investment decisions based on this analysis.

Leveraged exchange-traded funds (ETFs) tied to South Korea’s two biggest chipmakers shed close to $1 billion in August.
The withdrawals mark the first monthly outflow since the products launched in late May. The reversal comes as enthusiasm around the AI trade has weakened and regulators have introduced measures aimed at curbing speculative demand.
Leveraged Chip ETFs in South Korea Snap Inflow Run
Data compiled by Bloomberg Intelligence shows $601 million left the funds tracking SK Hynix, while Samsung-linked products lost $381 million. The ETFs aim to deliver twice the daily move of the underlying stock.
The reversal follows a brutal July for Korean equities. The KOSPI sank 22% that month. Samsung Electronics fell 21.5% over the same period.
SK Hynix dropped 35.5%, deepening losses for retail traders who had piled into double-leveraged wrappers.
Officials called an emergency meeting after 864.5 trillion won left the market across two sessions. Lawmakers blamed single-stock leveraged ETFs for amplifying the slide.
Regulators responded by raising the minimum deposit for new investors. They also mandated a five-day mock trading session. Trading volumes in the products cooled sharply afterward, according to Bloomberg.
Declining Trading Volume of Leveraged ETFs. Source: Bloomberg
Follow us on X to get the latest news as it happens
Retail Money Rotates Into ELS
Both chipmakers have steadied since. Samsung is up 3.63% in August, and SK Hynix has gained 2.19%, though Samsung slid 8.7% on Monday after its record shareholder return plan disappointed the investors.
Samsung and SK Hynix Stock Performance in August. Source: TradingView
Cooling ETF demand has not ended Korean risk appetite. Instead, mom-and-pop traders moved into equity-linked securities (ELS), advertising annualized coupons of 40% to 50%.
About 3.5 trillion won, or $2.5 billion, of ELS products were sold in July, according to the Korea Financial Investment Association. This was the highest monthly total since April 2023, led by notes tied to Samsung and SK Hynix.
The structured notes carry their own history of losses, having burned Korean buyers during the 2016 Brexit vote, the 2020 oil crash, and the China equity slump between 2021 and 2024.
Whether the coupon chase proves steadier than the leverage trade depends on how the chipmakers hold their August gains.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Meta Platforms is reportedly planning to launch a consumer AI agent, known internally as Hatch, within the coming weeks. Documents reviewed by The Information also show an October target for a new model, Watermelon.
The release anchors Chief Executive Mark Zuckerberg’s ambition to monetize Meta’s AI investments and diversify revenue. The AI spending has weighed substantially on the company’s cash flow.
Meta Plans Subscription AI Agent Hatch as It Seeks To Diversify Revenue
Meta has weighed a tiered subscription for Hatch, according to the documents. Premium access could cost up to $199.99 per month and include higher usage limits. The documents put the release in late August or early September.
The stakes show up in Meta’s own numbers. The firm reported $60.8 billion in second-quarter revenue. Advertising delivered $59.4 billion of that, more than 97% of the total.
Everything else is quite small. Reality Labs brought in $431 million, while other revenue reached $1.01 billion.
Meanwhile, the AI buildout continues to grow. Meta raised the floor of its 2026 capital expenditure range to $130 billion from $125 billion. The ceiling stayed at $145 billion.
Capital expenditures reached $31.08 billion in the quarter. Operating cash flow of $31.86 billion left just $784 million in free cash flow. The figure was $8.55 billion a year earlier.
Investors have not rewarded the buildout so far. META closed Monday at $559.02, valuing the company at nearly $1.42 trillion, and the shares have fallen by over 15% this year.
Meta Platforms (META) Stock Performance Year-To-Date. Source: Google Finance
A youth-safety trial in Oakland has also added pressure. Bank of America keeps a Buy rating and an $810 target. Mizuho is wary, comparing the case to the tobacco litigation of the 1990s.
Hatch would give Meta a consumer revenue line that does not depend on advertisers. Whether subscribers pay enough to matter against a $145 billion spending year is the open question.
Follow us on X to get the latest news as it happens
What Meta’s AI Agent Will Do
Hatch has been trained to work across DoorDash, Etsy, Reddit, Yelp, and Outlook. Early prototypes show a customizable dashboard featuring tools and skills created by AI agents.
Meta is also preparing a WhatsApp platform that lets users integrate and interact with third-party AI agents. The company could begin testing the platform with a limited group of users as soon as this week.
The company has shipped several models this year. Muse Spark arrived in April, version 1.1 in July, and version 1.2 alongside the Muse Code agent in August.
Whether Watermelon will join the Muse family remains unclear. Meta is estimated to report its third-quarter earnings on October 28, the same month Watermelon is due. Neither product will have had long to prove its economics by then.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

BlackRock led net inflows across both Bitcoin (BTC) and Ethereum (ETH) exchange-traded funds (ETFs) on Aug. 24, with a $209 million Bitcoin ETF inflow from its iShares Bitcoin Trust (IBIT) and $90.92 million from its Ethereum fund, ETHA.
Spot Bitcoin ETFs recorded $337.60 million in total net inflows for the day, per CoinGlass data. Spot Ethereum ETFs added $116 million, marking a sixth straight day of gains.
BlackRock Leads Bitcoin ETF Inflows
IBIT’s $209 million accounted for roughly 62% of the day’s total Bitcoin ETF inflow, CoinGlass data shows. The fund extended a stretch of strong demand that included a $606 million Bitcoin inflow just days earlier.
BlackRock Leads $338 Million Bitcoin ETF Inflow as Ether Funds Add $116 MillionU.S. spot Bitcoin ETFs recorded $338 million in net inflows on Aug. 24, led by BlackRock’s IBIT with $209 million, while spot Ether ETFs attracted $116 million, with BlackRock’s ETHA accounting for… pic.twitter.com/vd38qplHpR
— Wu Blockchain (@WuBlockchain) August 25, 2026
Total net assets across spot Bitcoin ETFs stood at $79.16 billion as of the latest update. BTC traded over $80,000 at the time of writing.
Bitcoin is holding above $80,000/ Image Source: BeInCrypto Ether ETFs Post Sixth Straight Day of Gains
BlackRock’s ETHA supplied $90.92 million of the day’s $116 million Ethereum ETF haul, about 78% of the total, per SoSoValue. Grayscale’s Ethereum Mini Trust added $12.50 million, the second-largest contribution.
The result builds on the biggest Ether ETF inflow in 10 months, recorded earlier in August. Ether changed hands near $2,486, up 2.4% over 24 hours.
BlackRock is the world’s largest asset manager, with $15.3 trillion in assets under management as of June 30, 2026. It led inflows across both crypto ETF categories on Aug. 24.
That dual dominance suggests BlackRock remains the primary entry point for institutional crypto exposure. The pattern held through most of August.

Chinese state-affiliated hackers now run twice as many attacks as they did before handing mundane work to DeepSeek and open-source artificial intelligence (AI) systems, according to Taiwanese threat intelligence firm TeamT5.
Attribution remains imprecise. The firm cannot tie every intrusion to a specific system, though it said that DeepSeek remains a popular choice among hackers.
Why Cheap AI Beats Frontier Models for Attackers
The finding inverts a common assumption that the risk of offensive AI lies mainly with the most advanced systems. Instead, operators are now scaling output using relatively weaker tools.
Cost and permissiveness drive that choice. Moonshot’s Kimi K3 is more powerful. Yet, TeamT5 has logged no incidents involving it and considers its running costs prohibitive for hackers.
Charles Li, chief analyst at TeamT5, framed the trade-off directly.
“DeepSeek is the AI of choice for Chinese hackers because it’s relatively powerful with very low cyber guardrails. Western models are highly sought-after but their guardrails are much more strict and require a lot more effort to bypass,” Li said.
Follow us on X to get the latest news as it happens
How Hackers Use AI
TeamT5 obtained scripts and logs, placing DeepSeek across multiple attack stages. A group called Grimfengxi used it to generate exploit code. Teleboyi used it to gather 1,000 IP addresses and map a target’s domains.
Huapi hit a Taiwanese company’s email system with a Chinese model that researchers believe was DeepSeek. Western tools appear too.
TeamT5 said a group tracked as Slime22 breached a Taiwanese technology firm’s systems, installed Kali, and directed Claude Code to run lateral movement. Operators bypassed safeguards by claiming to be engineers conducting authorized tests.
Meanwhile, CyCraft traced a 10-person Chinese startup selling intrusion software for 300,000 to 500,000 yuan, or roughly $44,500 to $74,000. At least four hacking groups bought it. The company also used ChatGPT during an attack.
A spokesperson for OpenAI said the firm is committed to identifying, preventing, and disrupting attempts to abuse its models.
Meanwhile, Chinese groups are not alone in this shift. North Korea’s Kimsuky is also testing local models.
Anthropic reached a broader conclusion in June, finding that AI now handles advanced attack work for hackers who lack the skill to do it themselves.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Nikita Bier says X trade buttons arrive soon, letting users trade tokens straight from their timelines. He also rejected claims that the platform quietly suppresses crypto accounts.
The former X head of product replied on Tuesday to a post blaming him for throttling Crypto Twitter. He credited his own features, then pointed to Washington rather than X for the rally.
X Trade Buttons Will Sit Next to Live Token Charts
An X user opened the exchange with a jab, writing that the “Bull Market is allowed to commence because Nikita is no longer shadow banning CT.” CT is shorthand for Crypto Twitter, the trading crowd on the platform. A shadowban quietly limits how many people see an account without telling its owner.
Bier answered directly. He said he built Cashtags, the tool that pulls Solana and Ethereum price charts into posts, and that trade buttons follow soon. He also claimed X gave crypto traders nothing at all before he joined.
What are you talking about? I literally built Cashtags, allowing people to add Solana and Ethereum charts directly into posts—with trade buttons coming soon. You can even paste contract addresses for newly minted tokens.There wasn’t a single feature on X for crypto traders…
— Nikita Bier (@nikitabier) August 25, 2026
Nikita Bier. Source: X
He also flagged contract addresses, the string of characters that identifies a token on its blockchain. Traders can already paste one straight into a post on X. That matters because scam tokens routinely copy the names of real projects, so buyers check the address before they send funds.
X trade buttons would close the last gap. Today a user spots a token on X, then leaves for a wallet or an exchange. A one-tap order removes that detour.
X first rolled out Cashtags in April to iPhone users in the US and Canada. The pilot drew an estimated $1 billion in volume within 48 hours. The company then added live stock charts to the same feature in May.
Bier Credits Treasury Buybacks for the Bull Market
However, Bier refused the idea that X itself sparked the current rally.
Instead, he pointed to the US Treasury buying back bonds and debasing the dollar. Treasury Secretary Scott Bessent at least doubled planned purchases of long-dated government debt on Aug. 20. Critics have since attacked the $950 billion buyback plan and how Washington intends to fund it.
Bitcoin Price Performance. Source: BeInCrypto Markets
Markets moved fast either way. Bitcoin (BTC) traded near $80,643, up 4.25% on the day and 24.5% over the past month. Gold meanwhile logged its best month since 1999.
Bier no longer runs product at X. He stepped back this month and stayed on as an adviser, which hands the rollout to his successor.
Therefore the timing question for X trade buttons stands. X has shipped charts, volume data and address lookups so far. Yet the buy button remains the piece that would turn a timeline into an order book.
Neither Bier nor X has named a launch date. Cashtags also still covers only a few markets, so a buy button needs a wider rollout first.

Roughly $16.8 million flowed into 30 cryptocurrency addresses tied to members of Iran’s Mabna Institute, according to blockchain intelligence firm TRM Labs.
The US Treasury listed those addresses as part of a wider action against Iran. One defendant’s wallets received 92% of that total.
Iran’s Mabna Institute Wallets Took In $16.8M, TRM Labs Says
The designated addresses span Bitcoin (BTC), Ethereum (ETH), and TRON (TRX). They belong to four of the 17 defendants also named in a superseding indictment unsealed by the Department of Justice (DOJ). Activity on the wallets dates back to January 2018.
According to TRM, Keyvan Fayaz controlled 10 of the addresses. Those wallets received a combined $15.5 million between January 6, 2018, and August 20, 2026. That figure represents 92% of the network’s total on-chain volume.
TRM said the concentration suggests Fayaz may have functioned as a treasury for the operation. Addresses belonging to Behzad Mesri, separately accused of breaching HBO, show layered transfers between his own wallets. Hundreds of thousands of dollars eventually reached a deposit address at a large centralized exchange.
However, little value remains. TRM puts the combined residual balance across all 30 addresses at $202,662, or roughly 1% of what passed through them.
Follow us on X to get the latest news as it happens
Treasury Targets Iran’s Digital Assets Sector
Meanwhile, the DOJ unsealed a 14-count superseding indictment on August 18. Prosecutors allege the Mabna Institute, an Iran-based company, ran cyber intrusions since at least 2013 for the Islamic Revolutionary Guard Corps (IRGC), targeting 144 US universities and 178 foreign ones.
Furthermore, the Office of Foreign Assets Control (OFAC) designated nearly 60 Iran-linked targets on Monday. This follows earlier designations of Iranian platforms, including Shelbit earlier this month.
“We are launching an economic onslaught against Iran’s financial connections around the globe. Our objective is to sever every economic lifeline that sustains this tyrannical regime until Tehran stands alone,” Secretary of the Treasury Scott Bessent said.
Treasury also issued five sectoral determinations under Executive Order 13902 covering digital assets, technology, gold, aviation, and shipping.
The determination reaches beyond the listed addresses. Under Executive Order 13902, OFAC can now designate persons operating in Iran’s digital assets sector, and the authority also covers non-Iranian persons. It does not automatically block every firm active in the sector.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Bitcoin (BTC) treasury firm Strive purchased 1,110 BTC for roughly $81.5 million last week, lifting its treasury to 21,356 coins.
The purchase cut Strive’s average cost to $93,257 per coin from $94,345. However, its Bitcoin reserve still carries a paper loss of more than $292 million.
Strive Keeps Buying While Its BTC Reserve Stays Underwater
The 1,110 coins mark Strive’s largest purchase since June 1, when it added 2,499 Bitcoin. It disclosed the purchase in an 8-K filing with the Securities and Exchange Commission (SEC) on August 24.
The coins were acquired between August 17 and 21. The company paid an average of $73,409 per coin, inclusive of fees and expenses. Strive made two smaller purchases earlier in August.
It added 147 coins in the first week and 79 in the second. Those buys cost $14.5 million combined, less than a fifth of the latest purchase.
The firm has now spent $1.99 billion across 29 separate purchases since September 2025. It has never sold a coin. The reserve is currently valued near $1.70 billion, or 14.66% below cost.
Strive Bitcoin Treasury. Source: Strive Strategy Tracker
The purchase arrives as CEO Matt Cole argued that the downturn has ended. He cited Bitcoin breakouts against both the US dollar and gold.
“My conviction is very strong that the Bitcoin bear market is over… seeing both relationships now turn higher together gives me more confidence in the next 12 to 18 months…” he said.
Follow us on X to get the latest news as it happens
Strategy Moved the Opposite Way in August
Strive’s spending contrasts sharply with Strategy, which raised $3.28 billion by selling stock in August and bought no Bitcoin.
Strategy has sold 6,916 coins since its last purchase on June 22. Its average cost of $75,385 sits below current levels, so that position remains profitable.
Strive’s does not. Bitcoin traded at $79,787 at press time. Strive therefore needs a 17% gain to reach its $93,257 break-even price. Thus, Cole’s conviction now carries a measurable price tag.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

Cosmos Labs confirmed an ongoing security incident affecting users of the Cosmos EVM module. It advised chains in contact with it to ask validators to halt block production.
Three networks have now disclosed impact. KiiChain and TAC froze their chains after attackers drained accounts, while MANTRA restarted its mainnet.
3 Chains Traced Incidents to Cosmos EVM
Three networks disclosed security incidents within days of each other. All three named the Cosmos EVM module, a component that lets Cosmos SDK chains run Ethereum-style smart contracts.
MANTRA was first. BeInCrypto reported that the team halted the chain as a precaution amid a security incident in an upstream dependency. The team said two MANTRA-managed wallets were affected, and user balances were never impacted.
The network later informed users that the vulnerability was in the Cosmos-EVM module and that it had been fixed in version 8.4.0, allowing the network to resume normal block production.
Follow us on X to get the latest news as it happens
MANTRA Chain is producing blocks again.The vulnerability in the Cosmos-EVM module has been fixed, the network has resumed, and no user funds were affected.Thank you to everyone for your patience throughout the incident.Review the full history of incident status updates… pic.twitter.com/IDVpw7H7Tp
— MANTRA | The EVM L1 for RWAs (@MANTRA_Chain) August 22, 2026
KiiChain then disclosed an exploit. The team said that on August 22, an attacker repeated the same technique 18 times, draining 148,326,583.15 KII before validators halted the chain at block 9355723.
“The vulnerability is in Cosmos code, not KiiChain code. It sits in the shared Cosmos EVM module (cosmos/evm), which KiiChain runs unmodified,” the team said.
The chain remains halted. KiiChain said the network will resume through a coordinated binary upgrade at a predetermined block height, with all validators applying the update simultaneously. The process will not require an on-chain governance proposal.
TAC halted the same day at block 24,671,475 after an attacker drained a single account. The team said the defect sits in the shared module rather than in TAC-specific code.
Cosmos Labs has pointed teams with questions to its security contact and said it will publish an incident report once the situation is resolved. It has not yet described the cause.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights