BeInCrypto Global

L’action IBM a chuté de 21,86 % lors de la séance pré-marché de mardi, à 226,80 $, après qu’International Business Machines a publié des résultats préliminaires du deuxième trimestre inférieurs aux attentes de Wall Street. La baisse a effacé près de 60 milliards de dollars de valeur boursière par rapport à la clôture de lundi à 290,23 $.
La rare annonce préalable est intervenue huit jours avant la conférence téléphonique sur les résultats prévue par l’entreprise. Les investisseurs doivent désormais attendre le 22 juillet pour obtenir les résultats complets et des orientations détaillées actualisées pour l’ensemble de l’année.
IBM a dévoilé un chiffre d’affaires préliminaire de 17,2 milliards de dollars, en hausse de 1 % par rapport à l’année précédente, mais inférieur aux 17,86 milliards de dollars attendus par les analystes interrogés par FactSet. Les bénéfices ajustés, à 2,93 $ par action, ont également déçu, sous la moyenne des analystes de 3,01 $.

Les deux premières semaines explosives de la chaîne Robinhood ont ravivé une question familière. Ethereum est-il mort, ou bien un courtier qui dessert près de 28 millions de clients vient-il d’apporter l’argument le plus solide à ce jour en faveur de l’idée qu’il est en train de gagner discrètement ?
La couche 2 lancée le 1er juillet, construite avec la technologie Arbitrum et utilisant l’ether (ETH) comme jeton de gaz natif. Ethereum finalise ses transactions et conserve ses registres, mais les analystes ne s’accordent pas du tout sur la personne qui capte la valeur.
Pourquoi Robinhood Chain renforce le scénario haussier autour d’Ethereum

L’inflation américaine s’est refroidie plus rapidement que ne le prévoyaient les économistes le 14 juillet 2026, offrant une victoire notable aux actifs à risque, dont la cryptomonnaie.
Le bitcoin a repris 63 000 $ à l’annonce et semblait en bonne voie pour tester 64 000 $ au moment où ces lignes sont écrites.
Le Bureau of Labor Statistics a annoncé un IPC (headline) de juin à 3,5% en glissement annuel, en dessous des 3,8% attendus par le consensus. L’IPC de base, hors alimentation et énergie, s’est établi à 2,6% sur un an, contre des attentes autour de 2,8–2,9%. Les prix “headline” ont également reculé de 0,4% par rapport au mois précédent.

Un nouvel avertissement de bulle liée à l’IA fissure les actions de mémoire. SanDisk, SK Hynix, Micron et Samsung affichent tous des schémas de retournement baissiers après un rallye très chaud en 2026.
Les dégâts ne sont peut-être même pas encore à leur niveau. On dirait moins un mouvement sectoriel unique qu’une remise en cause valeur par valeur, où même Samsung, le leader relatif, est en train de se dégrader.
Un avertissement de bulle liée à l’IA divise la filière des puces
Le déclencheur est venu de Wall Street. Le 1er juillet, l’indicateur de risque lié à la bulle de Bank of America a atteint 0,91 sur 1 pour les valeurs de semi-conducteurs, et l’ETF de puces SOXX a chuté de 6,4 % en une journée. BofA y a vu une poche d’air, pas un krach complet.

Morgan Stanley, Bank of America, Mizuho et Compass Point ont relevé leurs objectifs de cours pour Robinhood Markets (HOOD) entre le 2 juillet et le 10 juillet. La nouvelle fourchette va de 124 à 132 dollars, les résultats du deuxième trimestre étant attendus le 29 juillet.
Quatre desks rivaux qui se déplacent dans la même direction en l’espace de huit jours est inhabituel. Les catalyseurs qu’ils citent sont sur la chaîne, au sein de nouveaux comptes de trading liés à l’IA, et dans une émission obligataire inédite.
Les hausses de l’objectif de cours de Robinhood sont arrivées en grappes
Mizuho a été la première à agir. L’analyste Dan Dolev a relevé son objectif à 130 dollars depuis 115 dollars le 2 juillet, estimant que le courtier pourrait devenir le premier hyperscaler mondial des courtages en ligne, selon CNBC.

Les memecoins ont absorbé plus de 1,2 milliard de dollars de ventes nettes sur Binance depuis que le Bitcoin (BTC) a atteint son record en octobre 2025.
Ce chiffre illustre la pression exercée sur les actifs les plus risqués de la crypto pendant une correction prolongée.
Pourquoi la chute a le plus durement touché les memecoins
La baisse du Bitcoin, depuis son record d’octobre, a entraîné dans sa chute les actifs les plus liés à la spéculation. Les memecoins dominent cette liste.
La vente n’a pas été inversée. L’analyste Darkfost a souligné que le volume net cumulé a diminué régulièrement d’octobre à juillet.

L’argent s’est échangé près de 58 dollars mardi, soit environ 52% en dessous de son sommet historique de janvier à 121,76 dollars, après avoir chuté d’environ 4% lundi. Une nouvelle escalade américano-iranienne autour du détroit d’Ormuz a entraîné la dernière vague de baisse.
Les prix plus élevés du pétrole ont ravivé les inquiétudes concernant l’inflation et renforcé les paris sur une hausse des taux de la Réserve fédérale en septembre. Les opérateurs surveillent désormais le rapport sur l’IPC de juin ainsi que le témoignage devant le Congrès du président de la Fed, Kevin Warsh, pour obtenir des indications.
Blocus d’Ormuz et paris de hausse des taux de la Fed maintiennent l’argent sous pression
Le président Donald Trump a rétabli un blocus visant les navires iraniens traversant le détroit d’Ormuz. Il a également exigé un remboursement auprès des pays qui profitent de la protection américaine de la route maritime.

There is no single best currency for every kind of saving. Our research found three assets with three different jobs: the US dollar for liquidity, gold for long-term insurance, and Bitcoin for high-risk upside.
The Swiss franc was the strongest government-issued currency in the study. Yet even the franc failed to beat US inflation in most long holding periods. Gold protected purchasing power more often, while Bitcoin produced far higher returns over its much shorter history and charged investors with repeated crashes.
The main lesson is simple. A saver first needs to decide what the money must do. Cash for next month’s bills has a different job from wealth meant to last for decades.
Figure 1. The final three combine different strengths: the US dollar for liquidity, gold for insurance, and Bitcoin for high-risk upside. What “Best Currency” Means in This Study
People often use currency, cash, and savings as if they mean the same thing. They do not.
In this study, “money” includes seven government-issued currencies, gold, and Bitcoin. The seven currencies are the US dollar, euro, British pound, Swiss franc, Singapore dollar, Japanese yen, and Chinese yuan.
Gold and Bitcoin are not everyday currencies for most people. We included them because savers use both as alternatives to government money and because neither has a central issuer that can create more supply at will.
The research judged each form of money by four practical questions. Did it hold its value over time? Could a saver sell or use it during a crisis? How severe were the losses along the way? Could an owner move and control it without depending entirely on one institution?
This produces a broader answer than a simple exchange-rate table. A currency can be stable and still lose purchasing power. An asset can deliver a large return and still be unsuitable for emergency savings.
How the Research Was Conducted
We ran three return tests and then built a seven-part scorecard. All figures use data available through July 10, 2026. The charts draw on BeInCrypto Research calculations using the London Bullion Market Association, US Bureau of Labor Statistics, Federal Reserve H.10 exchange-rate series, IMF reserve data, World Gold Council, RWA.xyz, and CoinGecko. Individual claims link to supporting sources in the text.
Test 1: What Happened to $100 from 1971?
The first test starts in 1971, when the United States ended the dollar’s convertibility into gold. That change marked the beginning of the modern system of freely floating government-issued currencies.
We asked what happened if a saver converted $100 into each available form of money in 1971 and held it until the study’s cutoff date. We then translated the final value back into US dollars.
Alongside those returns, we plotted an inflation line using the US Consumer Price Index. It shows how much $100 had to become to buy the same broad basket of goods and services. By July 2026, the answer was about $815.
The test measures the money itself. It excludes bank interest, bond yield, gold storage costs, trading fees, taxes, and other income or expenses. A dollar, therefore, remains a nominal $100 throughout the chart, even though its buying power falls.
This choice matters. A saver who held Treasury bills or an interest-paying account would have done better than someone holding banknotes. The same principle applies to interest earned on deposits in other currencies.
Figure 2. The 55-year test compares the US-dollar value of $100 held in each available form of money with the amount needed to match US inflation. Test 2: A common starting point in 2013
Four of the nine assets could not take the full 1971 test. Bitcoin and the euro did not exist. The Chinese yuan was not freely tradable, and comparable Singapore dollar data starts later.
We therefore repeated the exercise from the end of 2013. This gave all nine assets the same starting date and allowed Bitcoin to complete at least one full 10-year holding period.
In this test, $100 needed to become $144 to match US inflation. Bitcoin reached $8,381 and gold reached $342.
The Swiss franc finished at $110 and was the only fiat currency to beat the dollar, although it still missed the inflation target.
Figure 3. The common 2013 test puts all nine assets on the same starting line. It excludes interest on foreign-currency deposits. Test 3: Rolling five-year and 10-year periods
A single start date can flatter or punish an asset. Someone who bought gold near its 1980 peak had a very different experience from someone who bought it in 2001.
To reduce that start-date problem, we tested every available five-year and 10-year holding period in annual steps. A 10-year window beginning in 1971 was one observation. The next began in 1972, then 1973, and so on.
For each window, we asked a yes-or-no question: did the money rise faster than US inflation? The final percentage is the share of windows in which it succeeded.
Bitcoin beat inflation in all four of its available 10-year windows. That perfect result comes from a short and favourable sample. Gold succeeded in 59% of 10-year windows. The Swiss franc managed 22%, the yen 24%, and the pound and dollar cash recorded 0%.
Figure 4. Rolling windows show how often each available asset beat US inflation over five-year and 10-year holding periods. The Seven-Part Scorecard
Returns tell only part of the story. We also scored each asset from 1, or weak, to 5, or strong, across seven traits: supply discipline, market liquidity, trust, inflation protection, behaviour during crises, portability, and price stability.
The scores combine market data with editorial judgement. They are a framework for comparing different risks, rather than a price forecast or a promise of future performance.
The Swiss franc earned the highest raw score at 30 out of 35. Gold followed at 28. The US dollar and Singapore dollar scored 27.
However, the total alone does not decide the final choice. Two assets can earn similar scores while doing the same job. A useful combination needs assets whose strengths cover different weaknesses.
Figure 5. The scorecard rates all nine assets across seven practical traits. The totals show overall defensibility, while the last column identifies each asset’s most useful role. The Cost of Volatility
We compared annualised US-dollar returns with each asset’s worst peak-to-trough fall from 2014 to July 2026. A peak-to-trough fall, or drawdown, measures how far an asset dropped before it began to recover.
For this chart, short-term US Treasury bills act as a realistic low-risk dollar cash proxy. That is different from the zero-interest cash assumption in the long-run currency tests, and the chart labels the change.
The result shows the trade-off clearly. Assets with small price falls generally produced low returns. Bitcoin produced the largest return and the deepest loss.
Figure 6. The return-versus-drawdown chart shows the price a saver paid for each asset’s return. No asset delivered both the highest return and the smallest loss. The Three Assets that Survived the Test
The final three did not finish first in every category. They survived because each solves a different savings problem.
1. US Dollar: The Operating Money
The dollar remains the world’s main reserve currency. Central banks hold it, companies borrow in it, and a large share of global trade is priced or settled through it.
It accounted for 56.9% of disclosed foreign-exchange reserves at the study’s cutoff date. That network creates constant demand and gives dollar markets unmatched depth.
This is especially important during a panic. Borrowers and investors often need dollars to repay debts or meet urgent obligations, so demand can rise even when the crisis began in the United States.
For a saver, this makes the dollar easy to buy, sell, transfer, and spend in many countries. It also makes dollar cash or short-term US government debt useful for emergency liquidity.
Figure 7. The dollar scores well for liquidity, crisis behaviour, portability, and stability, but poorly for supply discipline and inflation protection. The Dollar’s Weakness is Purchasing Power
The same dollar that worked well during crises performed badly as a 55-year store of value. It stayed at $100 in nominal terms while the inflation hurdle rose to $815.
Since 2013 alone, dollar cash lost about 30% of its US purchasing power. Interest can offset part of that loss, but the result depends on the rate paid and the tax charged on the interest.
The dollar’s global share has also declined. It represented about 71% of disclosed reserves in 1999, compared with 56.9% in the study.
The decline has been gradual. No single rival currency has replaced it. Central banks have spread part of their reserves into gold and smaller currencies.
Figure 8. The dollar remains dominant, but its share of disclosed global reserves has fallen since 1999. Debt and Sanctions Add Longer-Term Risk
US federal debt stood near 120% of annual economic output. Higher debt does not automatically cause a currency crisis, but it can limit policy choices and increase pressure to tolerate inflation.
The freezing of about $300 billion of Russian central-bank reserves after the invasion of Ukraine revealed another risk. Dollar assets held through banks and custodians can be blocked by governments.
That lesson matters most to states and sanctioned entities, but it also explains why some central banks have increased their gold holdings. Gold held directly does not depend on another country’s payment system.
Figure 9. Government debt varies widely across the currencies in the study. High debt can narrow a central bank’s room to raise interest rates or defend its currency. Digital Finance is Extending the Dollar’s Reach
Stablecoins and tokenised US Treasury products have carried the dollar onto blockchain networks. At the cutoff date, the market included hundreds of billions of dollars in dollar-backed stablecoins and about $15 billion in tokenised Treasuries.
These products add new settlement routes. They still depend on dollar assets and therefore extend the dollar system rather than replace it.
Figure 10. Dollar-backed stablecoins and tokenised Treasuries have expanded rapidly, creating new digital rails for the existing currency.
Verdict: The dollar remains the strongest choice for liquidity and near-term obligations. Cash alone is a weak long-term inflation hedge.
2. Gold: The Long-Term Insurance
Gold produced the strongest 55-year result. The notional $100 investment grew to $9,436, far above the $815 inflation line.
Its advantage is simple. No government issues gold, and no central bank can create more of it with a policy decision. It also has a long global market and no issuing government that can default.
Gold therefore protects against a different risk from dollar cash. It can benefit when investors lose confidence in currencies, public debt, or financial institutions.
Figure 11. Gold scores strongly for supply discipline, trust, inflation protection, and global liquidity. Storage and price swings reduce its score. Gold Did Not Protect Savers All the Time
The long-run return hides long periods of disappointment. Gold beat US inflation in 59% of rolling 10-year windows, which means it failed in about four out of ten.
Its worst 10-year period lost about 8.3% a year after inflation. A buyer near the 1980 peak then waited roughly two decades for a durable recovery.
By contrast, gold’s most recent 10-year period returned about 10% a year after inflation. It was the strongest rolling decade since 2011.
Figure 12. Rolling 10-year real returns show that gold’s inflation protection arrived in cycles, while fiat currencies often remained below zero. The Current Rise is Broad, Not Only a Dollar Story
Gold reached records in dollars, euros, pounds, yen, and yuan. When the same asset rises against several major currencies, the move reflects concern about government money more broadly.
That pattern is one reason we describe gold as insurance. It can respond to risks that affect several countries at once.
Figure 13. Gold rose sharply in five major currencies, which suggests a broad repricing rather than weakness in only one currency.
Central banks bought 863 tonnes of gold in 2025. It was the fourth consecutive year of unusually strong official demand.
Those purchases do not guarantee further price gains. They do show that national reserve managers increasingly use gold alongside government bonds.
Figure 14. Central-bank gold buying remained high in 2025 after three stronger years. Storage is Gold’s Main Practical Weakness
Physical gold is heavy, costly to secure, and difficult to move in large amounts. Gold held through a fund or bank is easier to trade, but the saver then depends on a custodian.
Governments have also restricted private ownership in the past. The United States sharply limited private monetary-gold ownership from 1933 until the 1970s.
Verdict: Gold has the strongest long-term record and can insure against broad loss of confidence in government money. It can still fall sharply and remain weak for many years.
3. Bitcoin: The High-Risk Upside
Bitcoin produced the largest return in the common 2013 test. A notional $100 became $8,381 by July 2026, compared with $342 for gold and an inflation target of $144.
Its supply rules cap the total at 21 million coins. An owner can also hold Bitcoin directly without relying on a bank, although doing so safely requires technical care.
Those features address two weaknesses of government money: supply can expand, and intermediaries can block access. They help explain why some investors treat Bitcoin as a digital alternative to gold.
Figure 15. Bitcoin scores highly for supply discipline and portability, but poorly for crisis behaviour and price stability. The Short History Flatters the Result
Bitcoin beat inflation in every available 10-year window. There were only four such windows, all beginning during the asset’s early growth period.
Gold, the franc, the yen, and the pound have decades of data across several inflation and interest-rate cycles. Bitcoin does not. Its result should therefore carry a large confidence discount.
A Saver Had to Survive Repeated Crashes
Bitcoin fell about 82% from its 2013 peak to the 2015 low. It fell 77% during the 2022 bear market and was around 50% below its October 2025 record of $126,080 at the study’s cutoff date.
These losses are too large for money that may be needed at short notice. A saver forced to sell during a downturn may never receive the long-term return shown in the chart.
Figure 16. Bitcoin’s drawdown history shows repeated losses of roughly half or more from previous peaks.
Bitcoin also lost value against gold during the latest decline. One Bitcoin bought about 35 ounces of gold at the end of 2024 and around 15 ounces by July 2026.
This comparison matters because a rise in dollar terms does not always mean an asset is preserving value against other scarce assets.
Figure 17. The Bitcoin-to-gold ratio shows how many ounces of gold one Bitcoin could buy. The ratio fell sharply after the end of 2024. Bitcoin Has Behaved Like a Risk Asset in Panics
Bitcoin fell sharply during the March 2020 market panic. It also received the lowest crisis-behaviour score in our framework because investors often sell it when they urgently need cash.
That does not remove its long-term case. It limits its usefulness as an emergency reserve.
Verdict: Bitcoin offers the strongest upside and the clearest defence against supply dilution. Its short history and extreme drawdowns make it unsuitable for money a saver cannot afford to lose or leave untouched.
Why the Other Currencies Did Not Make the Final Three
Several fiat currencies performed better than the dollar on individual measures. None added a sufficiently different role to the final combination.
Swiss Franc: The Strongest Fiat Currency
The Swiss franc turned $100 in 1971 into $486 and $100 in 2013 into $110. It was the only fiat currency to beat the dollar in the common 2013 test.
Switzerland has a long record of low inflation and strong institutions. Its currency often gains when investors become nervous.
However, the franc still beat US inflation in only 22% of rolling 10-year periods. It also comes from a small market, which limits its ability to absorb global savings at the scale of the dollar.
Policy can create sudden moves. In January 2015, the Swiss National Bank ended its exchange-rate floor against the euro without warning, and the franc rose as much as 30% within minutes.
Figure 18. The Swiss franc was the strongest fiat currency in the study, but it still failed the inflation test in most long holding periods.
Verdict: A high-quality fiat reserve for savers who can access it efficiently. It improves the cash layer but does not replace gold’s inflation role or Bitcoin’s upside.
Singapore Dollar: a Strong Regional Currency
The Singapore dollar finished the 2013 test at $98. That was close to its starting value and much stronger than the euro, pound, yen, or yuan.
The Monetary Authority of Singapore manages the currency against a basket within a policy band. The system has produced stability, but the exchange rate remains a policy choice shaped partly by the needs of an export economy.
The market is also small. The Singapore dollar does not appear as a separate major category in the IMF’s global reserve data.
Figure 19. The Singapore dollar remained stable against the US dollar after 2013, although it did not reach the US inflation line.
Verdict: A disciplined regional store of value. Its size and managed exchange rate limit its role as a global savings anchor.
Euro: Deep Markets, Weaker Returns
The euro is the clear second-largest reserve currency, with 20.3% of disclosed reserves. Its markets are large, liquid, and widely accessible.
Yet $100 converted into euros at the end of 2013 was worth $83 in July 2026. It lost value against the dollar and remained far below the $144 inflation target.
The euro also combines one central bank with 20 national governments and bond markets. During a crisis, investors can question whether the weakest member will receive enough support from the rest.
Figure 20. The euro is liquid and widely trusted, but it lost ground against the dollar after 2013 and carries political fragmentation risk.
Verdict: Useful for spending and saving in the euro area. The research found no evidence that euro cash offers stronger long-term inflation protection than the dollar.
British Pound: a Liquid Currency in Long Decline
The pound remains a major traded currency and one of the world’s larger reserve holdings. London gives it deep financial markets.
Its long-run savings result was the weakest in the 1971 test. The original $100 fell to $53 in dollar terms, and sterling beat US inflation in none of the 46 rolling 10-year windows.
The decline included sudden crises. Sterling came under severe pressure during the 1976 IMF rescue, the 1992 exit from Europe’s exchange-rate mechanism, and the 2022 UK budget shock.
Figure 21. The pound’s market remains deep, but its long-run exchange-rate and inflation record was weak.
Verdict: Highly usable and liquid, especially for UK expenses. Its historical record does not support it as a superior long-term store of value.
Japanese Yen: A Haven that Weakened after 2020
The yen nearly doubled against the dollar over the full 55-year test, turning $100 into $194. For much of that period, investors treated it as Asia’s main safe-haven currency.
The recent result was very different. The yen lost roughly 36% against the dollar from 2020 to July 2026 and finished the 2013 test at $65.
Japan’s government debt is more than twice the size of annual economic output, while the Bank of Japan owns close to half of the government bond market. Raising interest rates enough to support the yen can therefore increase pressure on public finances.
Figure 22. The yen’s long history contains large gains against the dollar, followed by a sharp decline after 2020.
Verdict: Still a large and liquid currency, but its recent policy constraints weaken the case for long-term savings.
Chinese Yuan: Scale with Limits on Access
China has the world’s second-largest economy, and about 30% of its own trade was settled in yuan at the study’s cutoff date.
Global reserve managers remain cautious. The yuan accounted for only 1.9% of disclosed reserves, and $100 converted at the end of 2013 was worth $89 in July 2026.
The main problem is control. China restricts how money moves across its borders. A saver may therefore face rules that limit when or how funds can leave the country.
Figure 23. The yuan benefits from China’s economic scale, but capital controls and limited global reserve use reduce its savings role.
Verdict: Increasingly useful for trade connected to China. Capital controls make it a poor general-purpose global savings currency.
What the Findings Mean for an Ordinary Saver
The first priority is matching the currency to the expense. Someone who pays rent and food bills in euros needs enough euros for those bills. Holding the entire emergency fund in dollars would add exchange-rate risk at the moment the money is needed.
After near-term expenses are covered, the study supports a layered approach. Dollar assets can provide global liquidity. Gold can protect against long-term monetary and political risk. Bitcoin can add a high-risk return source for money that can remain invested through a severe fall.
The research does not identify a universal percentage for each asset. The right mix depends on the saver’s home currency, time horizon, legal access, custody options, and ability to withstand losses.
Before choosing a savings currency, a reader should ask:
When will I need the money? Near-term funds need stability and easy access.
What currency will I spend it in? Matching future expenses reduces exchange-rate risk.
How large a temporary loss can I tolerate? Bitcoin and gold can fall when the money is needed.
Who controls access? Bank accounts, funds, physical gold, and self-custodied Bitcoin carry different custody risks.
The Swiss franc and Singapore dollar can strengthen a fiat savings layer for people with affordable access to them. They remain government-issued currencies and overlap with the dollar’s main role.
The evidence therefore points to three distinct functions. The dollar handles liquidity. Gold provides insurance against a wider loss of confidence in government money. Bitcoin supplies possible upside with a risk level that requires strict limits.
Final Summary
The 55-year test found no asset that was safe, liquid, inflation-resistant, portable, and calm at the same time. Every strength carried a cost.
Gold delivered the strongest long-run protection but failed through long stretches. The dollar stayed liquid in crises while losing buying power. Bitcoin led returns over its short record and suffered losses many savers could not tolerate.
For most readers, the practical answer is based on roles. Match near-term cash to the bills and keep liquid reserves for emergencies. Treat gold as long-term insurance. Bitcoin belongs only in money that can withstand a deep drawdown.

Le Bureau of Labor Statistics (BLS) américain publiera les données de l’indice des prix à la consommation (IPC) de juin mardi. Le rapport devrait montrer un ralentissement de l’inflation des consommateurs, porté par l’apaisement des prix du pétrole brut après l’annonce de cessez-le-feu entre les États-Unis (USD) et l’Iran.
L’indice des prix à la consommation (IPC) mensuel devrait reculer de 0,1 %, après la hausse de 0,5 % enregistrée en mai. Dans le même temps, la lecture annuelle est attendue en repli à 3,8 % contre 4,2 % le mois précédent, niveau le plus élevé depuis mai 2023. Les chiffres de l’IPC hors produits alimentaires et énergie, qui excluent les prix volatils des denrées alimentaires et de l’énergie, devraient progresser de 0,2 % sur un mois et de 2,9 % sur un an, en se stabilisant par rapport à mai.

Anthropic confirmed a massive billing error after its system tried to charge a South Korean free-tier user $16.6 million, despite zero API usage and no registered card.
The phantom invoices blocked his credit card and triggered a four-day support ordeal before the company responded.
Inside Anthropic’s $16.6 Million Phantom Invoice
A phantom invoice is a payment demand generated by a billing system without any underlying transaction or usage. That is exactly what a Korean developer, known as remy_notes, received from Anthropic in early July. The story quickly went viral after he shared the screenshots on Threads.
The first invoice arrived on July 7, totaling $1,669,875. Less than 24 hours later, a second one multiplied the figure almost tenfold to $16,627,739. The escalation occurred without any activity being recorded on the accoun
Follow us on X to get the latest news as it happens.
💸 Anthropic a accidentellement envoyé une facture de… 16,6 millions de dollars à un utilisateur de la version gratuite.Un utilisateur sud-coréen a eu la surprise de recevoir une facture de 16,6 millions de dollars de la part d'Anthropic, alors qu'il utilisait uniquement la… pic.twitter.com/78cddkLCJJ
— Goku 🗞 (@Crypto__Goku) July 14, 2026
The developer initially suspected phishing. However, the emails came from Anthropic’s official domain and used the company’s legitimate Stripe infrastructure.
Furthermore, his dashboard showed zero API usage, no billable keys, and no payment method on file. He audited his own automation scripts and AI agents, finding nothing that explained the charges.
His bank declined two overseas charge attempts on July 8 because the amounts exceeded the card’s per-transaction limits. The repeated attempts still resulted in his primary credit card being blocked.
What Caused the Billing Error, According to Anthropic
Anthropic confirmed the mistake on July 12 and assured that no money was collected. The company attributed the invoices to an incorrect automatic credit reload setting, which it disabled as a precaution.
It also clarified that the incident did not involve unauthorized access. Anthropic has not explained how the reload value reached such an extreme figure.
The resolution, however, was far from smooth. The user spent four days and around 18 emails seeking written confirmation that the invoices were void. Meanwhile, automated replies kept arriving instead of human responses. He eventually asked the company to respond by Monday afternoon, Korean time.
The case also exposed a wider industry problem. Audit startup Vaudit reviewed $34 million in AI invoices from 60 enterprise clients and found roughly $1.7 million in overcharges. That represents an error rate of about 5%, much of it tied to AI billing systems. Clients affected by those errors included major corporations like Panasonic, HP, and Honda.
For users, the lesson is simple. Reviewing billing dashboards, monitoring card notifications, and reporting unusual charges quickly can prevent a glitch from becoming a financial nightmare.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights.

La base de portefeuilles de Chainlink (LINK) s’élargit sur Ethereum (ETH), tandis que son prix continue de subir des vents contraires du marché.
De nouvelles données de Santiment portent le total à un niveau record de 900 000, et une vague de nouvelles intégrations suggère que cette croissance n’est pas le fruit du hasard.
L’adoption de Chainlink augmente tandis que le prix accuse du retard
Le chiffre constitue un record historique pour les portefeuilles LINK non vides sur Ethereum, avec plus de 20 000 ajouts au cours du mois écoulé. D’après les données de Santiment, ces ajouts sont survenus sans percée du prix.
« Ce type de croissance des détenteurs est généralement un signe de confiance à long terme », a déclaré Santiment.

Derive (DRV) a bondi d’environ 30 % après que les plus grandes bourses de Corée du Sud, Upbit et Bithumb, ont annoncé des inscriptions, faisant passer le prix de 0,12 $ vers 0,18 $.
Le double lancement offre au protocole de dérivés DeFi une liquidité nouvelle et une exposition accrue en même temps, dans l’un des marchés crypto les plus actifs.
Derive est un protocole d’options en chaîne et de contrats perpétuels construit sur Ethereum, sous la forme d’un optimistic rollup. La plateforme, connue sous le nom de Lyra Finance avant son changement de marque en 2024, combine de faibles frais, une liquidité profonde et la garde en self-custody. Cette combinaison en a fait l’un des lieux de négociation de dérivés les plus complets de la finance décentralisée.

XRP (XRP) et Ethereum (ETH) ont tous deux affiché un signal baissier à contre-courant, alors que le sentiment social devenait euphorique. Pourtant, leurs marchés dérivés racontaient des histoires opposées : le financement de XRP était négatif, tandis que celui d’Ethereum était positif.
Le partage compte parce que des discussions haussières élevées précèdent souvent des replis à court terme. Quand cela coïncide avec un financement négatif, comme sur XRP, la foule et les traders à effet de levier se retrouvent de part et d’autre de la même transaction.
Le signal de sentiment de XRP et Ethereum
Santiment a suivi le ratio entre les commentaires positifs et négatifs sur les plateformes sociales. La métrique met en balance les publications haussières et baissières pour chaque actif.

Une crise de la mémoire pilotée par l’IA est en train de remodeler le marché mondial des téléphones, et l’action Apple atteint un niveau record. Les livraisons mondiales de smartphones ont chuté de 6,7 % le trimestre dernier, mais Apple a progressé de 15,3 % et a publié des livraisons record.
La raison tient aux coûts. Les puces mémoire se vendent désormais près de trois fois le prix de l’an dernier, si bien que les fabricants de téléphones économiques ont augmenté leurs prix et perdu des acheteurs, tandis que les marques premium, sécurisées par des approvisionnements verrouillés, se sont retirées.
Comment la captation de mémoire par l’IA étrangle les téléphones
La pression commence dans les centres de données d’IA. Les hyperscalers achètent d’énormes volumes de mémoire pour entraîner et faire fonctionner des modèles d’IA, et cette demande a asséché l’offre pour les téléphones et les PC.

Une part importante des produits de la crypto de M. le président Donald Trump a été investie dans des actions et des obligations l’an dernier, selon une analyse de Reuters portant sur ses derniers relevés financiers.
Les déclarations montrent un président qui ne considère pas la crypto comme une réserve principale de richesse personnelle, même s’il la défend publiquement.
Les placements traditionnels de Trump quadruplent
Une analyse de Reuters des avoirs de Trump sur les deux dernières années a révélé que ses portefeuilles d’actions et d’obligations avaient au moins quadruplé. Il détenait entre 703 millions de dollars et 2,6 milliards de dollars de tels instruments à la fin de 2025, contre entre 225 millions de dollars et 608 millions de dollars un an plus tôt.

Most crypto brands disappear because they cannot make anyone feel the difference, not because their technology is weak, according to Jordi Urbea, CEO of Ogilvy Spain. He says sameness, not code, is the real killer.
Urbea spoke with BeInCrypto at the Ibiza Tech Forum 2026. He has spent 25 years helping brands stand out. His verdict on crypto marketing is blunt, and the data backs it up.
Every Crypto Brand Looks the Same
In an expert council interview with BeInCrypto, Urbea argued that crypto advertising has collapsed into one template. Swap the logo, he says, and the message barely changes.
“If you look at the crypto sector and all the advertising, the ads are exactly the same. You change the logo, and it’s the same.”
The numbers explain why sameness spreads so easily. Between 150 and 300 new coins launch every week, and roughly 10,700 remain active. Yet Bitcoin and Ethereum hold close to 75% of the total market value.
So thousands of near-identical projects compete for a shrinking slice of attention. In that crowd, a copied message vanishes on contact.
“It’s very strange to find one company that says, ‘This crypto is completely different.’ The rest are just repeating, message by message. And people say it’s boring, it’s all the same.”
Great Technology, No Story
For Urbea, the failure is rarely technical. He has watched strong projects die for a simpler reason.
“For many years I collaborated with many startups, and most of them disappeared because they couldn’t explain the difference between one brand and another. There are people with amazing technology and amazing ideas, but they don’t have the capacity to explain it.”
Startup data backs him almost exactly. CB Insights found the top reason companies fail is no market need, cited in about 42% of cases. Marketing and go-to-market problems account for a further large share.
Running out of money tops some lists at 70%, yet that is the final symptom. The root cause usually sits upstream, in a value no one managed to communicate.
Communication and market-fit failures top the list, not broken tech
Crypto shows the pattern at an extreme scale. More than 53% of all tokens launched since 2021 have already failed, and 2025 was the deadliest year on record.
Most of those projects were not undone by broken code. They simply never gave the market a reason to remember them.
The Follow-the-Leader Trap
Urbea believes imitation is the mechanism behind the sameness. Teams copy whatever seems to work for a rival.
“In some cases people repeat the formulas that work for others. ‘It goes well for that company, so I’ll repeat it.’ Follow the leader and repeat. But by the tenth message, your brand disappears, your message disappears, and you’re a big ship lost in the night.”
Marketing science adds a useful twist here. Byron Sharp and the Ehrenberg-Bass Institute argue brands grow by being distinctive rather than merely different, because buyers choose fast and rarely study fine detail.
That view sharpens Urbea’s point instead of breaking it. Copying rivals erases the distinctive assets, the voice, colors, and language that let a brand register at all. Without them, recall collapses.
The same logic haunts Web3 marketers who chase trends. When every campaign borrows the same hooks, none of them stick.
Building a Brand Nobody Can Copy
Jordi Urbea has a direct remedy. Stop borrowing formulas and build your own.
“If you create your space, you create your language, you create your own way to work. That is my humble advice.”
The payoff is measurable. Kantar analyzed 40,000 brands and found a strong link between relative uniqueness and the amount consumers are willing to pay. Distinctive brands command higher margins and lower price sensitivity.
Research also shows that fresh, varied advertising lifts recall, while repetition fades fast. A distinct voice is therefore an asset, not a cost.
For crypto founders, the lesson mirrors classic marketing wisdom. Technology may open the door, but identity is what keeps a brand alive.
As automation floods every channel with more content, Urbea’s warning grows louder. In a market of copies, the only safe move is to be impossible to copy.

Les traders de Bitcoin sont en alerte maximale avant la publication de l’indice américain des prix à la consommation (IPC) de juin, prévue le 14 juillet, le BTC évoluant près de 62 000 $ après des mois de volatilité marquée.
De précédentes publications d’inflation ont déclenché des variations à deux chiffres, et le rapport de juin pourrait décider du prochain grand mouvement du marché.
Données de l’IPC et Bitcoin : un schéma d’oscillations violentes
L’indice des prix à la consommation mesure l’évolution des prix des biens et services dans le temps, ce qui en fait le principal indicateur de l’inflation aux États-Unis. Les marchés le suivent de près, car ces données influencent les anticipations concernant la politique de la Réserve fédérale. Une seule surprise dans le rapport peut faire évoluer les paris sur des baisses de taux en quelques minutes.

Le volume de négociation sur 24 heures déclaré par Upbit a bondi de 1 437% aujourd’hui, alors que l’indice KOSPI de Corée du Sud chutait et que les marchés boursiers en Asie s’affaiblissaient.
La flambée d’activité a coïncidé avec une vaste vague de ventes dans la région. Le chiffre d’affaires des transactions a augmenté, même si les principaux indices de référence à Séoul, Hong Kong, Tokyo et Taipei ont tous évolué à la baisse.
Trading sur les Rockets d’Upbit +1 437% alors que le KOSPI chute de 4%
Le volume échangé reflète une hausse de l’activité de trading plutôt qu’un transfert confirmé de capitaux hors des actions. Le total en dollars a atteint 4,24 milliards de dollars au moment de la rédaction, selon les données de CoinGecko.

Le marché européen des crypto-actifs est suffisamment vaste pour que la réglementation modifie la carte concurrentielle. Chainalysis a indiqué que les volumes régionaux de crypto ont retrouvé un pic mensuel de 234 milliards de dollars en décembre 2024, tandis que l’Allemagne, la France, le Royaume-Uni et d’autres grands marchés européens ont chacun reçu des centaines de milliards de dollars de valeur crypto entre juillet 2024 et juin 2025.
MiCAR a désormais placé ce marché derrière un filtre d’autorisation plus strict. La période de transition s’est achevée le 1er juillet 2026 et la CSSF luxembourgeoise a indiqué que les prestataires de services sur actifs virtuels ne sont plus autorisés à offrir des services dans l’UE sans autorisation en tant que prestataires de services sur crypto-actifs, ou CASP.

A new poll shows 79% of Americans bracing for a long US war with Iran. This comes as President Donald Trump told Congress that military action resumed on July 7, a move that frees the military for 60 more days without congressional approval.
The shift lands months before November’s midterms, where pump prices could cost Republicans their grip on Congress.
Most Americans Now Expect a Long War With Iran
The Reuters/Ipsos poll surveyed 1,019 US adults over three days, closing Sunday. Only one in five, 18%, still expect the fighting to end within weeks. The rest see no quick exit.
Furthermore, only 37% backed the strikes, which restarted on June 26 after Washington blamed Tehran for hitting commercial ships.
Earlier surveys told a similar story. A Financial Times poll found 58% of voters judged the war not worth the cost. A separate Generation Lab survey of adults aged 18 to 34 found 77% called the strikes on Iran the wrong move.
Hormuz Blockade Reignites Market Fears
On Monday, Trump vowed to seal off Iranian ports near the strait and skim 20% off every cargo passing through. Tehran had already called the channel shut.
“The Hormuz Strait is OPEN, and will remain OPEN, with or without Iran… The U.S.A. will be, from this point forward, known as ‘THE GUARDIAN OF THE HORMUZ STRAIT’… reimbursed, at the rate of 20% on all cargo shipped,” he said.
Follow us on X to get the latest news as it happens
BREAKING: President Trump has formally notified US Congress that the US is at war with Iran, per Politico.Details include:1. The move gives Trump another 60 days to use the military in the region without congressional approval2. Trump said strikes that began on July 7th…
— The Kobeissi Letter (@KobeissiLetter) July 13, 2026
Traders moved quickly. Crude climbed roughly 4%, and Bitcoin (BTC) slid to around $62,600 amid fears the chokepoint will remain snarled for a while.
The poll also revealed that 6 in 10 respondents expect gasoline to be costlier over the coming year. Pump prices already sit near $3.87 a gallon, far above where they stood before the war.
Meanwhile, the war has weighed on Trump’s standing. He has claimed a 59% approval rating, yet the New York Times tracker pegs him closer to 39%.
The stakes climb as the November midterms near. His sinking approval adds to the pressure on Republicans, who risk losing the House and possibly the Senate in November.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights