BeInCrypto Global

President Donald Trump will award Elon Musk the National Medal of Science on Thursday. Musk is one of six tech leaders honored, while OpenAI’s Sam Altman and Anthropic’s Dario Amodei, the best-known AI lab chiefs, are not.
The medals are America’s highest honors for science and technology. Trump has never presented either one across his two terms.
Who Is Getting Trump’s Science and Technology Medals
Musk, Google co-founder Sergey Brin, Nvidia CEO Jensen Huang, and AMD CEO Lisa Su will receive the science medal. Dell Technologies CEO Michael Dell and Microsoft CEO Satya Nadella will receive the technology medal.
The awards will be handed out at the White House “Science: A New Golden Age Summit.” The event will focus on superintelligence, meaning AI that outperforms humans at most tasks.
“These recipients are helping ensure America keeps leading the world in innovation,” White House spokesperson Liz Huston told Fox.
The medal also caps Musk’s return to Trump’s side. The pair split in 2025 after Musk called Trump’s tax bill a “disgusting abomination.” He now co-leads the Pentagon’s Project Meridian study of future warfare.
Why Altman and Amodei Are Missing From the List
The White House has not explained the picks. Its statement named only the six winners.
*President Trump is set to award six prominent technology leaders the nation’s highest science and technology honors Thursday during the White House’s “Science: A New Golden Age Summit". • National Medal of Science: Elon Musk, CEO of $TSLA Tesla and $SPCX SpaceX; Sergey Brin,… pic.twitter.com/DEW4N7spm4
— Jesse Cohen (@JesseCohenInv) October 7, 2026
The list arrived the same week OpenAI posted proofs for hundreds of unsolved math problems, worked by an internal ChatGPT model. Altman said AI is opening a “new era of discovery.”
Anthropic’s year with Washington has been rockier. In March, the Pentagon labeled it a supply-chain risk after it refused to let its Claude models power autonomous weapons. An appeals court upheld the Pentagon’s Anthropic blacklist in September, days before Trump hosted Amodei for dinner.
Altman skipped Trump’s September 29 AI lunch for OpenAI’s own conference, sending president Greg Brockman. Musk, Nadella, and Su attended.
Amodei has urged a slower pace for the most powerful AI. Trump has rejected that, arguing it helps China, which has cut America’s AI lead to 3% on benchmark scores, Bloomberg Intelligence estimates.
Thursday’s ceremony will put the chip, cloud, and rocket builders on stage. The heads of the companies behind ChatGPT and Claude will not be among them.

Wall Street went from celebration to caution in less than 24 hours.
On Tuesday, the S&P 500 climbed 0.6% to a record close of 7,818.93. The Nasdaq also reached a new high as falling bond yields and another surge in AI stocks pushed investors back into equities.
By Wednesday, that momentum had disappeared.
Industrials were the weakest S&P 500 sector. Homebuilders fell 2.9% and chip stocks lost 1.2%. Nvidia slipped around 0.7%, while SpaceX fell 2.5%.
Caterpillar plunged nearly 6% after US regulators launched an inquiry into competition in the agricultural equipment market.
Nvidia Stock Turns Red on Wednesday, September 7, After Hitting Record Highs. Source: Yahoo Finance Why Did US Stocks Fall? Treasury yields surged again
The 30-year US Treasury yield reached a 24-year high, while the 10-year briefly moved above 5.3%. Higher bond yields give investors an increasingly attractive alternative to stocks.
They also reduce what investors are willing to pay today for future corporate profits, putting particular pressure on expensive growth stocks.
Oil returned above $100
Brent crude remained around $100 as Middle East supply concerns returned. Expensive energy can push inflation higher and make it harder for the Federal Reserve to lower interest rates.
The Fed offered little relief
Minutes from September’s meeting showed officials remain concerned about inflation. Markets currently see only a small chance of another hike in October, but investors still expect rates to stay restrictive.
Investors took profits after the record
The market had avoided the correction many investors expected this summer. Wednesday gave traders an excuse to lock in gains after another record close.
BREAKING: September Fed Meeting Minutes show that most Fed officials expect another interest rate hike by year-end.All 19 Fed officials backed the September interest rate hike.Almost all Fed officials see inflation risks tilted upward, with some warning AI could push demand…
— The Kobeissi Letter (@KobeissiLetter) October 7, 2026
Will Wall Street Bounce Back?
For now, major Wall Street firms are largely expecting a slowdown rather than a collapse.
Goldman Sachs, JPMorgan and Morgan Stanley all have 8,000 year-end targets for the S&P 500.
Citi and UBS see 8,100. From Wednesday’s 7,801 close, that implies only modest upside.
BofA is more cautious at 7,400, while Wells Fargo sees 7,700.
That leaves the market sitting close to where many banks think it should finish the year.
The next move may depend heavily on whether Treasury yields and oil retreat — and whether the coming earnings season can justify stocks trading near record levels.

In April 2021, a flaw in a crypto platform’s code let one man walk off with $53.3 million. He spent part of it on Pokémon cards. On Wednesday, a New York jury called it theft.
Jonathan Spalletta, 36, of Maryland, faces up to 30 years in prison. His lawyers argued he only used tools the platform left open to anyone.
The Bug That Paid Out $53 Million
Uranium Finance was a crypto exchange with no company in charge. Software called a smart contract held users’ coins and paid them out by fixed rules.
On April 8, 2021, Spalletta found a flaw in those rules. He used it repeatedly to collect about $1.4 million he was not owed, the Justice Department said.
Two weeks later, he described it in writing.
“I did a crypto heist of $1.5MM a couple of weeks ago . . . There was a bug in a smart contract, and I exploited it . . . Crypto is all fake internet money anyway.”
Prosecutors say he then pressured Uranium into letting him keep $386,000 as a “bug bounty,” the reward for reporting flaws. The government called it a sham.
On April 28, he hit a second flaw. He drained about $53.3 million from several pools, and Uranium shut down.
From Crypto Heist to Pokémon Cards
He first ran the money through Tornado Cash, a service that hides where coins came from.
Then he went shopping. A first-edition Pokémon base set cost about $750,000, and a sealed booster box $257,500.
A Black Lotus, a famously rare Magic card, cost $500,000. A Roman coin marking Julius Caesar’s murder cost $601,545.
He also paid $137,500 for a scrap of Wright brothers airplane fabric that Neil Armstrong carried to the moon.
The jury took about two hours to convict, Inner City Press reported. BeInCrypto covered the original March charges.
“As Spalletta’s many victims know, those words could not be further from the truth,” said US Attorney Jamie McDonald in a statement.
Agents seized the Black Lotus and the moon fabric from his home, plus about $31 million in crypto.
The $350 million September Bitget breach ranks as 2026’s largest hack, per DefiLlama data.
A judge will now decide Spalletta’s sentence.

Robinhood has added $25 million worth of Bitcoin (BTC) to its corporate balance sheet. The trading platform framed the purchase as a statement of commitment to crypto, not a bet on returns.
Less than a year ago, however, Robinhood’s incoming finance chief publicly doubted that holding Bitcoin was a sensible use of company cash.
What Changed Behind the Robinhood Bitcoin Buy?
Johann Kerbrat, Robinhood’s senior vice president and general manager of crypto and international, disclosed the purchase on Wednesday. He spoke during a livestreamed interview broadcast on X.
In November 2025, Shiv Verma, then Robinhood’s incoming chief financial officer (CFO), weighed the same idea. Verma also named alignment with the crypto community as a benefit, but raised a sharper question.
“Is it the best use of our capital? There’s a lot of different things you’re doing, you know, from new products, for growth, investing in engineering.”
Shiv Verma, Robinhood, November 2025 interview
Verma, who has since stepped into the CFO role, said at the time that the company would keep reviewing the pros and cons.
How Much Does $25 Million Mean for a $100 Billion Company?
Kerbrat acknowledged the stake is small. He said it would do little to change the trajectory of a company valued near $100 billion.
At Bitcoin’s current price of $85,582, the purchase equals roughly 292 BTC. Robinhood has not disclosed the exact number of coins.
The stake also represents about 0.025% of Robinhood’s market value.
By comparison, Strategy spent $28.7 million on 334 BTC between October 1 and 4, according to its filing. That purchase lifted its holdings to 848,000 BTC.
Meanwhile, Robinhood’s own crypto business has cooled. Cryptocurrency revenue fell 38% year-over-year to $100 million in the second quarter, according to company results.
Total revenue, in contrast, rose 32% to a record $1.31 billion.
Robinhood (HOOD) shares closed Tuesday at $112, down 1.85%. The stock extended losses in Wednesday’s pre-market session, according to TipRanks.
Is Bitcoin Only One Piece of a Wider Crypto Bet?
The Bitcoin purchase sits alongside a broader on-chain expansion. Robinhood Chain, the company’s layer-2 network built on Arbitrum technology, launched its public mainnet on July 1.
The network now holds $1.047 billion in total value locked (TVL), according to DefiLlama. TVL measures the assets deposited in a network’s applications.
Robinhood Chain total value locked since its July mainnet launch / Source: DefiLlama
The chain crossed $1 billion roughly 80 days after launch. Growth has since slowed, with TVL holding between $1 billion and $1.05 billion since late September.
Robinhood also plans perpetual futures for eligible US customers, with up to 10x leverage on BTC and ETH. The company unveiled them last week alongside its in-app AI agents.
Unlike treasury firms that issue stock to keep stacking coins, Robinhood is spending a token sum to signal loyalty. Its next quarterly filing should confirm the exact holdings and show whether further purchases follow.

Ethereum researcher Justin Drake on Wednesday laid out how holders can protect their coins if AI cracks wallet security. His main step is moving funds to brand-new addresses that have never sent a transaction.
The move needs no new software or wallets. However, Drake said a break could come within months in the worst case, while warning that a rushed exodus would do more harm than good.
How Drake Says to Protect Your Coins
Drake recommends 4 steps, starting with the largest and most experienced holders.
Move most funds to a never-used address.
An address that has never sent coins keeps its owner’s key hidden.
Switch addresses after every spend.
Once coins leave an address, the rest should go to a fresh one. It can come from the same seed phrase, the backup word list that restores a wallet.
Do not rush or panic.
Drake called the step simple and preventive, not an emergency.
Know where you stand.
Wallets under 50 Bitcoin (BTC) have partial cover, he said, from about 20,000 early exposed addresses holding 50 BTC each.
“Don’t rush. While I believe there is cause for action a rushed migration would do more harm than good,” he warned.
Why a Fresh Address Works as a Defense
Every crypto wallet has a private key, a secret code that approves payments, and a matching public key.
A never-used address shows only a scrambled fingerprint of the public key, called a hash. The full key appears on the blockchain, crypto’s public ledger, at the first spend.
Drake fears AI could help attackers work backward from a visible public key to the private one. He pointed to OpenAI’s release this week of 722 AI-written math papers.
None of the results he named deals with wallet math. No published research shows Bitcoin or Ethereum (ETH) cryptography has been broken.
Still, about 6.04 million BTC already sits behind visible keys, according to analytics firm Glassnode. BeInCrypto cited that figure earlier Wednesday in coverage of Europol’s quantum wallet warning.
Drake named Binance, Robinhood, Bitfinex, and Tether as firms that could harden their offline reserve wallets. He also said Ethereum’s switch to standard SHA or BLAKE hash-based security must now speed up.

Most Federal Reserve officials said another interest rate hike would likely be appropriate this year, minutes of the September 15–16 meeting showed on Wednesday. US stocks and gold barely moved on the release. Bitcoin was the exception.
The minutes cover the meeting where the Federal Open Market Committee (FOMC), the Fed’s rate-setting panel, lifted its benchmark rate by 0.25 percentage point to a range of 3.75% to 4.00%. The 12–0 vote delivered the Fed’s first hike since 2023.
Gold, Stocks, and Bitcoin Reaction After FOMC Minutes. Source: TradingView What the Fed Minutes Say About the Next Rate Hike
“Most” participants saw another increase in 2026 as likely appropriate, according to the minutes. “Some” said the risks to inflation had tilted further toward prices rising faster than expected.
Several officials judged that the current rate was not restrictive, or only mildly so. In Fed terms, restrictive means borrowing costs high enough to slow spending and cool inflation.
🇺🇸 Key takeaways from today’s Fed Minutes:• All 19 officials supported September’s rate hike• Most expect another hike in 2026• Inflation risks remain tilted higher• Many saw risk-management reasons to raise rates• Several see rates as only mildly restrictive•… pic.twitter.com/ir54tOV0Xr
— CryptoGoos (@cryptogoos) October 7, 2026
Several participants also said the scale of the artificial intelligence (AI) buildout, the wave of spending on data centers and chips, continued to surprise them.
Why Stocks and Gold Barely Moved on the Fed Minutes
The S&P 500, the benchmark index of 500 large US companies, sat at about 7,801 in the first five minutes after the 2 p.m. ET release, down 0.02%, TradingView data showed.
Gold held near $4,110 an ounce. It had touched about $4,125 roughly 45 minutes before the minutes came out.
Little in the document was new.
The view that one more hike was coming had already been published on September 16, when the Fed’s projections showed the median official expecting one more quarter-point increase by year-end. Sixteen of 18 officials pencilled in at least one.
The minutes also predate the September jobs report. Employers added about 29,000 jobs that month against forecasts near 90,000, and unemployment rose to 4.2%.
Traders had already cut the odds of an October hike to roughly 20%, from about 55% a week earlier. Banks such as J.P. Morgan expect one more December hike instead.
Why Bitcoin Was the Only Market That Reacted
Bitcoin (BTC) rose from about $83,159 to $83,306 on Binance in the first five minutes after the release, a 0.18% gain. That was the largest move among the three markets.
The bounce followed a slide. Bitcoin had traded near $83,600 around 12:30 p.m. ET before falling to about $83,050 shortly before 2 p.m. ET, the same chart showed.
Unlike the stock market, Bitcoin trades around the clock and had already sold off into the release. Rate decisions have been driving crypto flows: investors put $3.55 billion into crypto funds in the week after the September hike, the largest weekly crypto inflows of 2026.
What Comes Before the October 28 Fed Decision
Officials have split in public since the meeting. New York Fed President John Williams called one more hike this year reasonable. Dallas Fed President Lorie Logan argued at least two may be needed. Governor Michelle Bowman said she sees no urgent need to act.
Long-term Treasury yields recently hit their highest since 2002, a move tied to the $400 billion metals selloff earlier on Wednesday.
The Fed meets next on October 27–28, days before the US midterm elections. September consumer price data, due October 14, will be the last major inflation reading before that decision. Track the Bitcoin price today as the data lands.

Cardano’s booth at TOKEN2049 did not feature its biggest and most popular project, Midnight. The privacy-focused blockchain is a major part of the Cardano ecosystem, yet it wasn’t featured anywhere in the branding.
Charles Hoskinson has publicly criticized this and called it a result of governance politics at the project.
The timing made the omission harder to ignore. Hours earlier, Midnight announced plans to support Solana wallets.
Midnight has become the where's waldo of Cardano pic.twitter.com/pvge2BB2T0
— Charles Hoskinson (@IOHK_Charles) October 7, 2026
Cardano’s Biggest Project Is Missing
Hoskinson walked viewers through the Cardano display at TOKEN2049, which featured projects from across the ecosystem. Even SNEK, the Cardano meme coin, made the wall.
Midnight did not. This is odd because Midnight is Cardano’s largest project and NIGHT already trades on Binance as a Cardano native token.
He also suggested the Cardano Foundation helped decide what appeared on the booth, joking that Midnight may simply be so private that it refused to reveal itself.
The dispute quickly became more serious.
“The booth is a public asset for Cardano. It was funded by the Cardano community.”
— Charles Hoskinson
That matters because Cardano’s treasury is ultimately funded and governed through ADA holders and their delegates.
Midnight Finds a Warmer Welcome on Solana
The booth dispute also exposes a broader political divide inside Cardano.
Hoskinson has repeatedly clashed with the Cardano Foundation over governance and the direction of the ecosystem. That tension resurfaced in June when a crypto fund founder called for Hoskinson’s removal, triggering a fierce community debate.
Meanwhile, Midnight is widening its reach.
At a recent Solana Summit, the team said Solana users would eventually be able to make private swaps without installing separate software.
Hoskinson publicly welcomed the move in a quote post.
Welcome to Midnight, Solana https://t.co/4ejFp6n5LL
— Charles Hoskinson (@IOHK_Charles) October 6, 2026
The irony is difficult to miss. A project built inside the Cardano ecosystem is now being welcomed by Solana while its own logo is missing from Cardano’s biggest conference booth.

Two trading firms hired to keep Hunter Biden’s LAPTOP meme coin stable came out millions of dollars ahead after its launch-day crash, according to a report Biden released Wednesday. It does not name either firm.
The report, by Groom Lake, finds the founders never sold. It ties the crash to a thin launch pool, heavy early trading and one firm pulling its cash, with what it calls moderate confidence.
“We hired two of the biggest market makers in the business to avoid exactly this, and we gave them the capital to stock both sides,” Biden stated.
Why Did a $6 Trade Move LAPTOP More Than $7,400 of Selling?
LAPTOP, a meme coin on Coinbase’s Base network, launched September 9. Its main trading pool, which sets the price automatically, held $35,663 but only 29,885 tokens.
That imbalance meant a $6.02 purchase lifted the price 5%. Pushing it down 5% took $7,376 of selling.
Across 668 other launches the report checked, buying and selling moved prices almost equally.
The price jumped from 5 cents to $316.75 in under two minutes, then LAPTOP’s two-minute peak gave way to a 98% drop within the hour.
Hunter Biden’s Laptop (LAPTOP) Price Performance. Source: BeInCrypto
The “$300 billion” valuations multiplied that price by 1 billion tokens.
What Did the Market Makers Do During the Crash?
Market makers are firms hired to keep buy and sell orders flowing. According to the report, the first received $500,000 but put about $5,244 into the main pool.
Then, 84 seconds after the peak, it withdrew. Cash for sellers near the market price fell from $16,158 to zero.
The report says its positions ended $685,873 ahead, though its loan terms reserve some fees for the lender. Trading the report links to the second firm netted $2.18 million.
The team says the first firm has not explained itself. A day after launch, Biden had denied the scam claims and blamed bots. Now he wants the coins destroyed.
9/The biggest winners? The market makers.Market Maker 1's DEX positions ended up about $686,000 ahead. DEX trading linked to Market Maker 2 netted more than $2.1 million.I think the market maker who screwed up the launch should buy it all back and burn it.
— Hunter Biden (@HunterBiden) October 7, 2026
Biden, who called Donald Trump’s TRUMP coin “max extraction,” also targeted paid crypto promoters known as KOLs.
“Hell, I’d make it work just to piss off Don and Eric. And every KOL who called it a scam because we refused to pay them,” he added.
The founders’ 300 million tokens stay locked for six months, then release over two years. Next week, the team plans to burn most unclaimed airdrop tokens. Neither firm the report points to has been named.

Lumentum has quietly become one of the biggest winners of the AI infrastructure boom. Shares have surged about 570% over the past year, leaving NVIDIA, Dell, and even Micron behind.
The stock rose another 3.82% on October 6 to close at $1,133.40, just below its 52-week high of $1,137.20.
The scale of the rally stands out even in a market obsessed with AI. Over the same 12 months, Micron gained about 457% and Dell rose roughly 308%. NVIDIA gained around 28%.
That means Lumentum delivered more than 21 times NVIDIA’s return. Its market value has now climbed to roughly $101.7 billion.
Lumentum Stock Price Over the Past Year. Source: Yahoo Finance Why Did Lumentum Stock Rally So Much?
Lumentum makes lasers and optical components used inside data centers, telecom networks and AI infrastructure.
As companies build larger AI systems, they need faster connections between servers and data centers. That has pushed more money into optical networking suppliers.
The numbers have followed.
Fiscal fourth-quarter revenue reached $1.01 billion, up 109.3% year over year. Earnings of $3.23 per share also beat expectations.
Citigroup has pointed to demand for optical circuit switches as another major growth driver.
$COHR and $LITE are both up over 10% after a report said U.S. could restrict Chinese optical transceivers starting at 3.2T without touching today’s 800G and 1.6T buildout.That makes ~65% U.S. content workaround really interesting since it could shift more of the next optics… pic.twitter.com/E58YbFPtx0
— Shay Boloor (@StockSavvyShay) October 1, 2026
Lumentum Price Forecast: What is Wall Street Targeting?
After a 570% rally, analysts are sharply divided.
The average target among 23 analysts is $1,078.75, already around 5% below the latest share price.
Citigroup is far more bullish, with a $1,400 target implying roughly 24% upside.
Rosenblatt sees $1,300, while Stifel recently raised its target to $1,232.
UBS sits at the other end with an $820 target, implying roughly 28% downside.
Earlier this year, $LITE CEO Michael Hurlston reported that Lumentum was under-shipping optical demand by 30%.At the Global Photonics Economic Forum, he revealed that the projected deficit has now expanded to 70%:"Next year, with the advent of CPO and NPO in 2027, our… https://t.co/iH5mS9Gi2o pic.twitter.com/0xgDZdi6p0
— KawzInvests (@KawzInvests) October 3, 2026
The rally has already priced in enormous growth. From here, Lumentum will need earnings to keep catching up with its share price.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights.

Math tests have been the benchmark method for testing the capabilities of AI models like ChatGPT and Claude. OpenAI says its models are now performing so well that existing math tests are becoming less useful. So researchers tried something harder.
They gave an unreleased model around 4,000 unsolved math problems. The results were surprising and concerning.
The AI produced 722 manuscripts, grouped into 372 families of related results. OpenAI says each result used about three hours of reasoning compute on average.
“Some pretty exciting days ahead for the mathematical community!” said Stefano Gogioso, a member of BeInCrypto’s Future Tech and AI Experts Council
Is AI Becoming Too Powerful Too Fast?
This is the bigger story. Frontier AI is beginning to move beyond answering known questions and into generating possible answers to unknown ones.
That could sharply increase how much intellectual work a researcher, engineer or analyst can attempt.
But output is not truth. Many of OpenAI’s papers have computer-checkable Lean proofs. Others do not. OpenAI warns that some unverified results “could have issues.”
That creates a new bottleneck: humans may struggle to check research as quickly as AI can produce it.
We’re releasing a broad range of new mathematical results produced by an internal frontier model.We’ve been consulting with the independent Advisory Group on Mathematics and Artificial Intelligence at the Institute for Advanced Study, and we have drawn on their advice and…
— OpenAI (@OpenAI) October 6, 2026
Could an AI Now Perform Predictive Analysis and Make Investment Decisions?
Possibly, but finance is harder in a different way.
A mathematical proof can eventually be shown right or wrong. Markets are noisy and constantly changing.
Recent finance benchmarks still show frontier AI struggling with complex investment research, while studies of market timing find limited predictive advantage.
The near-term opportunity is deeper analysis rather than perfect prediction.
An AI capable of hours of sustained reasoning could examine filings, earnings calls, macro data and competing scenarios simultaneously, then test far more hypotheses than one analyst could.
That may be the bigger signal from OpenAI’s experiment. AI is becoming capable of producing serious analytical work at extraordinary volume. The next problem is deciding which of it deserves to be trusted.

Gold and silver lost about $400 billion in paper value within minutes on Wednesday as US bond yields hit their highest level since 2002. The figure reprices metal already owned. It does not measure money pulled from markets.
Gold slid from about $4,120 to near $4,066 an ounce in a 15-minute window. Likewise, silver fell from roughly $60.13 to $59.00 before both metals bounced.
Gold and Silver Prices Against US 10-Year Yields. Source: TradingView How a $54 Gold Drop Becomes a $400 Billion Loss
The headline number comes from multiplication. The World Gold Council estimates about 216,000 tonnes of gold have been mined in history. That is close to 7 billion ounces.
A $54 drop across that stock marks it down by roughly $375 billion. Silver, valued near $3.65 trillion in BeInCrypto’s $1 trillion metals selloff report, adds about $67 billion for its $1.13 fall. Together, that lands near $440 billion.
BREAKING: Over $400 BILLION has been wiped out from gold and silver in just 10 minutes. pic.twitter.com/O2tVWHgicJ
— Bull Theory (@BullTheoryio) October 7, 2026
However, owners who did not sell still hold the same ounces. Only the price used to value them changed.
Why Rising Bond Yields Pulled Gold and Silver Lower
The 10-year Treasury yield, the interest rate on 10-year US government debt, reached 5.35% on Wednesday. The 30-year yield touched 5.70%.
U.S. 30-YEAR YIELD HITS NEW 24-YEAR HIGHThe 30-year Treasury yield climbed to 5.706%, its highest since 2002, as the global bond selloff resumed ahead of Fed minutes.The 10-year yield rose to 5.323%, while Brent crude topped $101 amid renewed geopolitical pressure.Danske…
— *Walter Bloomberg (@DeItaone) October 7, 2026
Gold and silver pay no interest. When bonds pay more, holding metal means giving up that income.
Meanwhile, traders were bracing for a $39 billion sale of 10-year notes and minutes from the Federal Reserve’s September meeting, when it raised rates to 3.75%–4.00%.
CME FedWatch data showed a 69% chance of another hike in December.
Target Rate Probabilities for December 9 Fed Meeting. Source: CME FedWatch Tool
“Here we go again. Traders have reacted to rising bond yields by selling gold and silver…Yet falling bond prices and rising yields are extremely bullish for precious metals,” Schiff wrote.
Gold is down 7.2% over the past month. Analysts hold that 2026 could be its most volatile year since 1982. The bond auction and Fed minutes later Wednesday are the next tests for yields, and for the metals that move against them.

Future quantum computers could crack the private keys that guard cryptocurrency wallets, Europol said in a report published Wednesday. For Bitcoin wallets whose public keys are already visible, the EU police agency says no cryptographic fix exists.
Europol rejects claims that quantum computing will collapse crypto. Its warning is narrower, and falls on individual holders rather than the network.
Why Europol Says Wallets, Not the Blockchain, Are the Weak Point
Every Bitcoin (BTC) wallet runs on a pair of linked codes. Anyone can see the public key, while the private key authorizes spending.
A powerful enough quantum computer could work backward from the public key to the private one, letting an attacker move the funds, Europol’s cybercrime unit said.
The hashing math that chains Bitcoin’s blocks together is far harder to break, the report finds.
For coins with public keys, the report offers one way out.
“Once a public key is exposed, it remains permanently vulnerable to future quantum decryption.”
Owners must move those funds to new wallets before any attack arrives, Europol says.
How Many Bitcoin Are Already Exposed, and How Long Would a Rescue Take?
Analytics firm Glassnode put the figure at 6.04 million BTC in May, or 30.2% of all coins issued. At Bitcoin’s current price near $83,050, that stash is worth roughly $502 billion.
Most of it is exposed through reused addresses and exchange wallets, Glassnode found. The rest sits in older script formats, including Satoshi Nakamoto’s early coins.
Whether a holder’s coins sit in that group depends on how they are stored, according to Glassnode.
Most standard Bitcoin addresses show only a scrambled fingerprint of the public key, keeping the key itself hidden. Those coins are not counted as exposed.
That changes once the owner sends a payment from the address. The transaction reveals the key, and any coins left behind at that address lose their cover, Glassnode said.
Coins held on exchanges often land in the exposed group, because platforms tend to reuse addresses. Taproot addresses, a format added in 2021, and Bitcoin’s earliest addresses show the key by default.
Quantum-safe signatures are 10 to 120 times larger than today’s, Europol notes, which could clog blocks and raise fees.
One study the report cites estimates a full migration would need at least 76 days of the network’s entire capacity. Spreading it across a quarter of each block stretches that to about 300 days.
A second Europol report warns attackers could store encrypted data now and unlock it once quantum machines exist. It found no clear evidence of this at scale.
Europol set no date for such machines. It argues that upgrades on a network with no central operator take years, and some developers already back plans to freeze vulnerable coins.

Developed with input from members of the BeInCrypto Expert Council, including experts from AWS and Cloudflare, our new report looks at what the observable activity shows, how much money is moving, and where the hype runs ahead of the data.
By BeInCrypto Research
AI agents that can pay for things on their own are one of the loudest and most repeating stories in crypto right now. Stablecoins, new payment protocols and big-name launches keep making headlines. But how much of it is actually happening?
That is the question behind The State of AI Agent Payments 2026, the new report from BeInCrypto Research. We are launching it at TOKEN2049 Singapore, taking place 7-8 October at Marina Bay Sands.
The report looks at real-world use, not forecasts. It follows the money on-chain, reviews what products allow agents to do, and asks a simple question for businesses: what needs to be true before this goes mainstream?
What we found
Most x402 payments are tiny. In a 35-day sample of x402 payments on Base and Solana, we recorded 6.4 million transactions worth just $119,947 in total. 90.8% of them were worth less than one cent. That is exactly the kind of payment cards were never built for, and it explains why stablecoins keep coming up.
Lots of transactions does not mean lots of demand. Solana accounted for 94.5% of the sample’s transactions, and most of that came in just eight days. On 26 August alone, 145 addresses made 1,283,926 payments to just three recipient addresses. We don’t call that wash activity or establish intent, but it shows why transaction counts alone say little about real business demand.
What gets paid for is small and practical. The most-requested paid x402 resource was a two-cent token safety check, with 63,319 calls in the catalogue’s sampled period. Most of the top services supplied information or completed small software tasks. These are calls recorded, not confirmed purchases.
People still hold the reins. Of 19 AI products and workflows we reviewed, 14 act within limits set in advance by a person or business. None of the 19 sets its own goals. Five of the seven trading platforms we looked at explicitly block agents from withdrawing funds.
AI trading is still a small slice. Across four selected Hyperliquid connections, applications labelled as AI handled just 0.55% of $653.08 million in trading value. Almost all of it, 99.4%, went through CCXT, a standard connection tool that makes no AI claim. AI may also trade through CCXT unlabelled, so this is not a market-wide figure.
When things go wrong, they go wrong fast. We documented two incidents, including a public message that triggered a $175,000 transfer from an AI-linked wallet. Together, they sent $281,200 to attacker-controlled wallets before any recoveries.
The money is following the pipes. We reviewed 36 funding rounds and acquisitions worth $4.25 billion in disclosed value. Around 92% went to payment and cash-management businesses, not to trading agents.
Get the report
Download The State of AI Agent Payments 2026 here: research.beincrypto.com/agentic-payments-2026
Get a hard copy of the report at the BeInCrypto NewsDesk located at PB4-45 (Level 4, Marina Bay Sands) and answer our ‘Question of the Day’ survey while you’re there. We’ll collect answers across both days and feed them into the next edition, so you could end up in the research, not just reading it.
BeInCrypto is part of the BeInNews Academy Ltd, an independent media group covering the convergence of finance and digital assets.We help professionals act with confidence in a complex and fast-changing industry through our newsroom, research reports, events, expert councils and multimedia studio.

BitMine Immersion Technologies, the largest company holder of Ethereum (ETH), will stop buying once it owns 5% of all ETH. Chairman Tom Lee told the Token2049 conference in Singapore on Wednesday the firm is about 100,000 ETH away.
BitMine began buying on June 30, 2025, with ETH near $2,500. After 15 months and 6 million tokens, ETH trades at $2,581.
How Close Is BitMine to the 5% Mark?
BitMine reported 6,016,414 ETH as of October 4, about 4.9% of supply. The remaining 89,000 ETH would cost about $230 million at the current ETH price.
Ethereum Price Performance. Source: TradingView
“We only need to get another 100,000 ETH to get to 5%,” Lee stated at the Token2049 in Singapore.
Lee said ending purchases also means BitMine no longer needs to raise money to fund them. BeInCrypto reported in June that Lee had signaled the aggressive buying was ending.
BREAKING: Tom Lee says BitMine $BMNR will stop accumulating $ETH once it reaches 5% of Ethereum’s total supply.BitMine currently holds 6,016,414 ETH, equal to roughly 4.9% of total supply. pic.twitter.com/3sc3b3gWeo
— Crypto Rover (@cryptorover) October 7, 2026
How Much Did BitMine Move the ETH Price?
In August 2025, BitMine added about 1.1 million ETH, more than any other treasury company that month, according to VanEck. ETH gained 15.8% that month and peaked near $4,953.
VanEck said the stablecoin story and treasury-company buying drove that month’s $4 billion in Ethereum fund inflows.
In 2026, BitMine kept buying every week. ETH still fell below $2,000 by June, down 44% for the year at the time.
By then, BitMine’s 5.41 million ETH sat about $9 billion below what it paid.
Recent fund flows run the other way. US spot Ethereum exchange-traded funds (ETFs) lost about $408 million in six days, per SoSoValue data. That equals roughly 152,000 ETH, 10 times BitMine’s last weekly buy of 15,112.
Ethereum ETF Flows. Source: SoSoValue
BitMine’s own stock has already beaten ETH this year. Its 5.07 million staked ETH will still earn about $363 million a year. The next weekly holdings update should show how fast the final purchases arrive.

Financial advisors have an outdated idea of how wealthy women get rich and what they want from financial advice. Many advisors still picture female wealth as something that comes through inheritance or marriage.
A new BlackRock survey shows 8 in 10 women attribute their wealth to their professional careers. Only about half of advisors thought the same.
Advisors were far more likely to point to marriage and inheritance. Roughly six in 10 named each source, at least twice the share of women who did.
“Women are an increasingly powerful force in US wealth, and their story is about more than inheritance and wealth transfer,” BlackRock said.
Bar chart comparing what wealthy women and advisors say makes advice worth paying for. Source: BlackRock Report Women Want Investment Returns. Advisors Expect Reassurance
Advisors also misunderstand what these women want them to do with the money.
Half of the women surveyed ranked better investment returns among the top three reasons financial advice is worth paying for. Only 16% of advisors expected that.
Tax results showed a similar divide. Some 41% of women ranked better tax outcomes in their top three, compared with just 12% of advisors.
Among women with at least $5 million, 56% prioritized investment results and 47% tax results.
Bar chart comparing what contributes to women’s wealth. Source: BlackRock Report
This is important as 35% of women with at least $2 million have no advisor at all, according to BlackRock. Their biggest unmet need was better coordination between accountants, lawyers, and financial planners.
The market is becoming harder to ignore. McKinsey projects women will control $34 trillion in US assets by 2030, or around 38% of the total.
“The findings challenge some long-held assumptions: women are actively creating wealth, focused on growth and financial outcomes,” BlackRock added.
The survey covered 1,067 women and 409 financial advisors between July 29 and August 23. BlackRock noted that women could select all sources of wealth, while advisors could choose only three.

The first decade of crypto can be seen as a sequence of bubbles, hacks, memes, and projects that disappeared almost as quickly as they appeared. I see a decade of financial experimentation.
An open, global laboratory operating 24/7, with an unusual combination of innovation, speculation, fraud, successes, and failures.
Not everything created in that environment will survive. In fact, most of it will not. But what survives will change the way the financial industry operates.
What Has Survived So Far
Some of the mechanisms first tested in crypto are already being adapted and incorporated by traditional finance: faster settlement, programmable assets, continuous markets, new forms of collateral, automated market makers, contracts without expiration, and the ability to combine different financial services within the same infrastructure.
The fight is not between crypto and traditional finance. It is over the financial infrastructure that will emerge from their convergence.
The task is to absorb, adapt, and evolve the innovations that work — without importing the risks and deficiencies of the original laboratory.
Crypto Experiments that Survived Crypto Tested What TradFi Couldn’t At the Same Speed
In the more open and less regulated parts of crypto, the industry operated as an open-air laboratory. For better and for worse.
Protocols could be launched quickly. Users could test products without going through long institutional approval cycles. Markets could operate continuously. Developers could combine different applications and build new structures on top of existing contracts.
Some experiments were useful. Others were fragile. Many were simply speculation with a layer of technology on top.
But even failed experiments helped answer questions that traditional finance rarely gets to test at the same speed: What happens when settlement is effectively continuous? How does a market work without a central exchange?
Can collateral management be automated? What happens when a financial contract can execute certain conditions on its own? How does liquidity behave when any participant can create a pool?
Crypto did not invent every financial innovation of the last decade. But it was where many important ideas could be tested before being accepted by traditional finance.
Automated market makers, or AMMs, are one example. Instead of organizing trading through a central limit order book, an AMM uses liquidity pools and a formula to determine relative prices between assets. One participant supplies liquidity to the pool; another trades against that liquidity; and the contract executes the transaction automatically.
The mechanism has obvious advantages. It can operate without a traditional dealing desk, integrate with other protocols, and allow an exchange transaction to be executed directly on the network’s infrastructure.
But it also has known problems: manipulation risk, exposure to execution order and front-running.
That combination is precisely what makes the laboratory interesting. Innovation does not appear separately from risk. It appears together with risk.
The market’s job is to determine whether a mechanism can be redesigned, regulated and used in an environment where risks and responsibilities are understood and managed.
The same logic applies to other experiments: global markets operating 24/7, perpetual futures, stablecoins, tokenization and composability. Not all of these formats will be transferred directly to banks or exchanges, but the benefits they bring are clear.
Ideas from Crypto that TradFi Adopted Convergence is Not a Conversion of Faith
For a long time, the debate was framed as a conflict between two incompatible systems.
On one side, traditional finance: slow, expensive and excessively intermediated. On the other, crypto: fast, global, open and decentralized.
That polarized analysis is unhelpful because it turns a question of architecture and incentives into a question of identity.
Traditional finance has real challenges. Fragmented systems create reconciliation work, delays, operating costs and dependence on multiple intermediaries.
At the same time, it carries functions that are essential to a well-functioning market: governance, the legal definition of ownership, controls, custody, investor protection, the prevention of illicit activity, risk management and mechanisms for operating through periods of stress.
Crypto, in turn, showed that some of these functions can be performed differently. It also showed the cost of ignoring others.
Convergence, therefore, will not be a conversion of traditional finance to crypto — or the other way around. It will be a selection process.
The market will absorb whatever solves concrete problems:
• Faster settlement
• Programmable assets and money
• Process automation
• Continuous operation
• New forms of margin and collateral
• Greater interoperability
• Global distribution
• Fewer reconciliations and manual steps.
At the same time, those solutions will have to work with governance, legal certainty, compliance and risk management.
This is the least exciting part of the narrative, but it is what will determine how convergence takes place. A protocol can be technically elegant and economically irrelevant. An asset can be tokenized and remain illiquid. A transaction can settle instantly and still carry credit, counterparty, or ownership risk.
Tokenization does not turn a bad asset into a good one. It only changes how that asset is represented, transferred, and potentially integrated with other processes.
The Problem is Not Just Digitization. It is Integration.
The last four decades have been defined by the transition from an analog system to a digital one.
Records stopped being physical. Orders became electronic. Communication became faster. Information began to circulate at a speed that would have been difficult to imagine at the beginning of my career.
But digitization does not mean integration.
Financial infrastructure is still made up of different systems that need to communicate with one another. A transaction can pass through trading, confirmation, messaging, clearing, custody, ownership registration, money transfer and reconciliation. Each stage can be digital and the process as a whole can still remain fragmented.
That is where tokenization may be more important than the simple creation of new assets.
The BIS describes tokenization as recording rights to real or financial assets on a programmable platform when those rights previously existed on a traditional ledger.
The potential lies not only in creating a digital representation. It lies in combining messaging, reconciliation, and asset transfer within the same operation.
In practice, this could allow money, a financial asset, and contractual conditions to coexist in the same environment. An asset transfer could be conditional on payment. Collateral could be adjusted automatically.
A financing transaction could incorporate margin rules. A distribution could be programmed according to predefined criteria.
The most important change, therefore, may take place in the rails and plumbing of the financial industry.
Not only in the assets that appear on an investor’s screen, but in the infrastructure that allows those assets to be issued, traded, financed, collateralized, transferred, and settled.
Elements that Will Drive the Next Phase of Finance The Internet is a Good Example
The internet also began surrounded by experiments, fragile business models, and exaggerated expectations.
Many companies disappeared. Some ideas looked promising but never found an economically sustainable application. Others were absorbed by companies that did not even exist when the technology was first being developed.
What remained was not a list of the first projects. It was the protocols, connectivity, distribution and, above all, the behaviors that the new infrastructure made possible.
The analogy with crypto is useful for that reason. The value of the first decade does not have to lie only in the tokens and companies that dominated the previous cycle. It will lie in
the mechanisms that survive the market’s test.
Meme coins, NFTs, lending protocols and different exchange models were part of that process. Some will remain niche products. Others may disappear. Many will be rebuilt in regulated environments and integrated into the existing financial industry.
The point is not to predict which token will rise or fall. It is to understand which economic functions will continue to make sense once they are subjected to scale, governance, regulation, and liquidity.
That distinction matters to anyone investing, building products or setting strategy at a financial institution. Technology can be useful without the token associated with it capturing value.
A protocol can generate volume without generating sustainable returns for its participants. An application can be innovative and still fail to find product-market fit.
In finance, utility is not enough. You need to understand who pays, who captures value, who takes the risk, and what happens when the incentive disappears or something goes wrong.
The Transition Creates Opportunities and Destroys Established Positions
The biggest opportunities tend to appear when infrastructure changes.
That was true of the internet. It was also true in other technological and financial transitions: when the old standard still dominates, but the new standard is beginning to alter costs, behavior and business models.
The problem is that transitions also destroy established positions.
Companies that depend on a long chain of intermediation may lose relevance if certain steps become automated. Professionals who only execute repetitive processes may be replaced or see their work reconfigured.
Institutions that treat digital assets as an isolated category may discover too late that the technology has begun to affect traditional products such as payments, funds, foreign exchange, credit, custody and capital markets.
On the other hand, it is not enough to learn how to code a smart contract or understand the mechanics of a blockchain. The next phase will require a combination of market knowledge, technology, regulation, risk and distribution.
The next generation of financial professionals will not be defined only by their ability to trade an asset or use a tool. It will be defined by their ability to understand the interaction between the asset, the infrastructure, and the rules that make an operation possible at scale.
That is one of the subjects I intend to follow in this column: not only what changes in markets, but which business models and skills become valuable when infrastructure changes.
The Challenges
It would be a mistake to turn this thesis into a narrative of inevitable progress.
Institutional adoption of tokenization is still limited. Many projects remain in testing. Secondary-market liquidity is insufficient for a range of applications. Interoperability between networks has not been solved. Legal and operational standards are still being built.
Regulation has also advanced unevenly.
Many applications do not solve a meaningful problem. They were created because they could be created, not because there was sufficiently strong economic demand. It is a solution looking for a problem.
That is why the thesis does not depend on the entire crypto industry surviving. It depends on some of the mechanisms developed in that environment proving useful enough to be absorbed by regulated and economically sustainable structures.
The filter will be tough.
The next decade will probably be less about choosing a side and more about understanding which rails will carry money and assets.
That is where the change happens.

One year ago, Bitcoin (BTC) was worth more than ever. Four days later, one post from the White House erased $19 billion in bets.
Traders call that crash 10/10. It turns one on Saturday, and Bitcoin has still not recovered. It trades at $85,558, about 32% below its record.
Bitcoin Price Performance Since October 10, 2025. Source: TradingView How One Tariff Post Wiped Out 1.6 Million Traders
On October 6, 2025, Bitcoin hit about $126,200, effectively marking a new all-time high for the pioneer crypto. Four rays later on October 10, President Donald Trump announced a 100% tariff on Chinese goods. Crypto prices fell within hours.
Cryptocurrency prices tumbled on Friday after Trump said he would impose an additional 100% tariff on China and impose export controls on software. The declines precipitated — and then were made worse by — what data tracker Coinglass described as “the largest liquidation event in… pic.twitter.com/euSrIC6CwF
— Ajay Bagga (@Ajay_Bagga) October 11, 2025
Many traders had bet with borrowed money (leverage). When prices drop far enough, exchanges close those bets automatically. This is called a liquidation.
Data tracker Coinglass counted more than $19 billion in liquidations, its largest ever. About $7 billion vanished in one hour. Reportedly, over 1.6 million traders were hit.
The slide kept going, until the Bitcoin price ultimately bottomed near $58,600 on July 1.
Why Diesel Is the New Threat This October
This year, the danger sits at the pump. The US war with Iran is in its eighth month. Two things have happened this week that highlight growing tensions.
Iran lawmakers seek to impeach Foreign Minister Abbas Araghchi over U.S. contacts.
Iran oil Minister Mohsen Paknejad resigned.
Gulf crude is flowing again, at 91% of pre-war exports in September. Fuels such as diesel reached only 60%, Vortexa data shows.
Oil flows from the Gulf reached more than 80% of their pre-war levels in September, globally easing pressure on prices. Naveen Das of Kpler told Reuters high freight rates have encouraged shipowners back to the Gulf, but disruption is far from over pic.twitter.com/PkJ15xlXD8
— Reuters (@Reuters) October 6, 2026
US diesel averaged $6.41 a gallon on September 30, AAA says. That is about 73% higher than a year ago.
“Diesel is the fuel that moves nearly everything you buy,” said Patrick De Haan, head of petroleum analysis at GasBuddy.
On Monday, Trump said the strait no longer drives fuel prices. He blamed refineries instead.
“What’s driving up Gasoline is no longer the Strait of Hormuz… but the word, ‘Refineries,'” the President wrote on Truth Social
Minutes after that post, a tanker in the strait was struck and caught fire, BeInCrypto reported on Monday.
The Group of Seven (G7) has agreed to release 100 million barrels of emergency oil. Even so, Brent crude, the global oil benchmark, still traded near $102 on Monday.
Saturday brings the 10/10 anniversary with oil above $100 and diesel above $6.

Sui and Alibaba Cloud are teaming up to introduce a new way AI agents could pay for cloud services. Give AI a budget, and it could pay for the cloud services it needs without asking you to approve every purchase.
The approach builds on a model already documented in BeInCrypto’s State of AI Agent Payments 2026 report. Of 19 products and agentic workflows reviewed, 14 could act within limits set by people without needing approval every time.The two firms revealed the plan at Sui Basecamp, the network’s annual conference in Singapore. Still, neither side has shared a launch date, pricing, or which services come first.
What Changes When an AI Agent Holds the Budget?
Under the plan, Alibaba Cloud services will plug into Sui Agent Payments, Sui’s toolkit for AI agent payments. As a result, agents would pay for each request as they use a service.
Imagine this checkout experience: you set a budget, your agent does the rest.Sui and @alibaba_cloud are collaborating to bring Alibaba Cloud services to Sui Agent Payments, with per-call payments in stables on Sui.Announced live at Sui Basecamp: https://t.co/z23CpE8MM8
— Sui (@SuiNetwork) October 7, 2026
That model differs from card checkouts, which assume a person confirms every purchase. Instead, software sets the pace of spending within limits a human defines.
However, Sui and Alibaba Cloud have not explained how users brief their agents beyond setting a budget. A spending cap limits how much an agent can pay, but not how it spends.
One Bug, and AI Could Burn All the Cash
Consider an agent stuck in a loop, sending the same paid request over and over. Each call would settle automatically, so one bug could burn through the entire budget before anyone steps in.
Neither firm has said whether the system adds rate limits or alerts to stop that.
Mysten Labs, the developer behind Sui, built this year’s event around the so-called agentic economy. Its co-founder, Adeniyi Abiodun, summed up the thesis ahead of the conference.
“Economic activity is shifting from people clicking checkout buttons to agents transacting on their behalf, continuously and at machine speed.”
Adeniyi Abiodun, Co-Founder and Chief Product Officer at Mysten Labs, via Sui
Most Agentic Payment Workflows Work Within Human Set Limits. Source: BeInCrypto Research Sui and Alibaba Cloud Enter a Crowded Agent Payments Race
The deal lands as rival chains chase the same niche. In September, Cardano joined the x402 standard, which lets agents pay online services in ADA.
Meanwhile, Wall Street has taken notice. A BlackRock research paper named stablecoins as the leading candidate for machine-native money.
SUI 1-month price chart. Source: BeInCrypto
On the market side, Sui’s native token, SUI, has gained about 43% over the past month, BeInCrypto price data shows.
The bigger test reaches beyond one network. If Sui and Alibaba Cloud push per-call stablecoin payments to real scale, other cloud providers may face pressure to follow.
For now, adoption remains early. BeInCrypto’s State of Agentic Payments research examines how much real volume sits behind the headline numbers.

Samsung Electronics will publish preliminary Q3 earnings guidance on Thursday. Analysts expect operating profit of 106.1 trillion won ($79.1 billion), which would mark a fourth straight record quarter.
Demand for artificial intelligence (AI) infrastructure is still outrunning memory supply. However, memory price gains slowed during the quarter, leaving investors questioning whether chip margins have topped out.
What to Watch as Samsung’s Preliminary Q3 Numbers Land
Samsung’s guidance filings list only consolidated sales and operating profit. A detailed report follows in late October.
Thursday’s release will therefore answer the profit question first. The other four points rest on analyst estimates, industry data and the market’s reaction.
1. A Record Profit Built on Trimmed Forecasts
Profit is the key headline number, and it is expected to reach nearly nine times the 12.17 trillion won Samsung earned a year earlier.
That would top the record second-quarter operating profit of 89.5 trillion won ($62 billion) Samsung posted in July.
However, that figure reflects a downward revision. The LSEG SmartEstimate, based on 21 analysts, has dropped 7.7% since the end of August.
Two factors sit behind the cut.
Slower memory price gains. Memory chip prices kept climbing in Q3, though the pace eased.
A stronger won: The currency rallied 14.3% against the dollar in Q3, its biggest quarterly gain since early 1998. That lowers the value of Samsung’s overseas earnings when they are brought home.
2. Memory Margins Stuck at 76%
SK Securities analyst Han Dong-hee puts Samsung’s Q3 memory operating margin at 76%. That would leave it unchanged from the April to June period.
US rival Micron, meanwhile, expects its gross margin to ease to 86.3% from 87% this quarter. The company partly blamed employee compensation costs.
Follow us on X to get the latest news as it happens
3. Where DRAM Prices Head in Q4
TrendForce expects the slowdown in memory price gains to continue into Q4. It projects conventional DRAM contract prices will rise 10% to 15% over Q3, after a roughly 60% jump in the second quarter.
Avril Wu, TrendForce’s senior vice president for research, linked part of the slowdown to long-term supply deals.
“With ceiling-price mechanisms built in, the rate of price increases has slowed down…long-term agreements represent an increasingly higher proportion of suppliers’ total ‌output. With ⁠ceiling-price mechanisms built in, the rate of price increases has slowed down,” Wu said.
Samsung said in July it aims to secure long-term contracts covering about two-thirds of its memory output.
4. Samsung Chips Away at SK Hynix in HBM
SK Hynix remains the largest supplier of high-bandwidth memory (HBM), which AI data centers depend on. Samsung has been gaining on its rival this year.
Samsung had fallen behind after delays in qualifying its products for Nvidia. This year, larger shipments of its latest HBM4 chips have helped it recover ground.
J.P. Morgan’s estimates show how far that recovery could go. The bank sees Samsung’s HBM market share rising to 34% in 2026, up from 20% last year. Over the same period, SK Hynix’s share is forecast to fall to 46% from 60%.
5. How the Samsung Stock Takes the News
Samsung’s last preliminary guidance, on July 7, showed record Q2 operating profit. Shares still fell 6.92% from the prior close to 296,000 won. CNBC linked the drop to investor worries over capital spending and demand.
After the full results on July 30, the stock slipped only 0.72% from the prior close to 207,000 won.
Samsung Stock Performance. Source: Google Finance
On October 7, Samsung traded at 271,750 won at 1:45 p.m. Seoul time, down 0.09% on the day. That leaves it about 27% below its 52-week high of 374,500 won.
Still, shares are up 111.48% so far in 2026. Thursday’s preliminary figure will show whether another record, this time against lowered estimates, meets a similar reaction.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights

China’s central bank added 740,000 ounces of gold to its reserves in September, its largest monthly purchase since October 2023. The buying came as gold fell over 6%.
The People’s Bank of China (PBOC) lifted its holdings to 77.47 million fine troy ounces from 76.73 million in August. September also marked its 23rd straight month of gold buying.
How China’s Gold Buying Picked Up Pace Through 2026
The September addition works out to roughly 23 tonnes. It matched October 2023, when holdings also rose by 740,000 ounces, State Administration of Foreign Exchange (SAFE) records show.
Purchases have grown every month since March, when the PBOC added 160,000 ounces. September was the seventh straight month of larger additions. In August, China added 650,000 ounces while gold climbed nearly 10%, BeInCrypto reported.
China’s Monthly Gold additions, October 2023 to September 2026. Source: BeInCrypto/SAFE
The World Gold Council (WGC) placed China’s 2026 purchases at 80 tonnes through August. That ranked second only to Poland, which bought 98 tonnes. Adding September lifts China to roughly 103 tonnes for the year.
Follow us on X to get the latest news as it happens
Gold’s Slide Shrank the Value of China’s Reserves
Gold closed September at $4,157.14, down from $4,449.24 at the end of August, market data shows. That drop erased about 72% of August’s $404 gain.
The September close also left gold 21.2% below its February close of $5,278.51, the highest monthly close of 2026.
China’s reserve valuation moved lower too. The value of its gold reserves fell to $323.52 billion from $350.08 billion. The PBOC still held more gold than a month earlier.
SAFE also reported China’s foreign exchange reserves at $3.4 trillion at the end of September. That figure was down $38.1 billion from the end of August.
China’s Gold Buying vs. Gold Price in 2026. Source: BeInCrypto/SAFE, TradingView
The decline has carried into October. At press time, gold traded at $4,122.25, down 0.84% from its September close. It has dipped as low as $4,103.52 this month.
The PBOC’s October data will show whether purchases keep growing while prices sit below their August level.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights