OpenAI CEO Sam Altman has ruled out a 2026 initial public offering, citing unfinished safety and alignment work while leaving the company without a confirmed listing date.

Fortune reported on Sept. 12 that Altman called the current period an “ill-advised moment” to go public and said OpenAI felt no pressure to pursue an offering.

Asked whether an IPO had moved from 2026 to 2027, Altman replied, “I would say not 2026.” He said OpenAI had substantial work remaining on safety, alignment and cooperation between governments and the AI industry.

His answer rules out a listing this year but does not establish a 2027 offering. OpenAI has not publicly disclosed a listing date, stock exchange, ticker, price range or number of shares to be offered.

OpenAI IPO has no confirmed 2027 date

Some reports have described 2027 as the earliest possible date for an OpenAI IPO. Altman did not commit to that timetable during the interview. He said the company would go public when its business was ready and when conditions surrounding the technology supported that decision.

The distinction is important to the factual record. OpenAI did not have a publicly announced 2026 offering that it formally postponed. The company remains privately held, and no publicly available registration statement identifies an active OpenAI stock sale.

Media reports have previously linked the company to a possible listing carrying a valuation of up to $1 trillion. OpenAI has not confirmed that estimate. Any valuation attached to a future offering would depend on its financial results, investor demand, capital structure and the terms disclosed at the time.

A traditional U.S. IPO would require a registration statement containing business, financial, management and risk information. The SEC says Form S-1 is the basic registration form available to companies and must contain a prospectus with audited financial statements.

OpenAI has not announced when it might begin such a process. Altman’s comments leave 2027 possible, but describing that year as a confirmed IPO schedule would go beyond his statement.

Safety work takes priority over listing preparations

During the Fortune interview, Altman connected the decision to rapid advances in artificial intelligence and unresolved questions about controlling increasingly capable systems. He said safety standards were not yet ready for the industry to push capabilities much further without additional safeguards.

OpenAI has discussed pausing at certain capability levels, Altman said, giving researchers and institutions time to improve alignment measures. He did not identify a model release, capability threshold or binding protocol that would trigger a pause.

The CEO called for cooperation among competing AI developers and governments. Any international arrangement remains prospective because OpenAI has not released a signed agreement, participating organizations or enforcement terms.

Altman framed the IPO decision as part of OpenAI’s ability to place its mission ahead of immediate shareholder returns. A public company would face reporting duties and market expectations, though he did not claim those requirements made responsible AI development impossible.

No direct market reaction exists for OpenAI shares because the company is not publicly traded. Private-market transactions and reported valuation estimates do not provide the continuous price discovery associated with a listed stock.

Independent evaluator proposal awaits details

Altman’s remarks followed a proposal from Anthropic CEO Dario Amodei calling for slower development of frontier AI systems. Amodei urged laboratories to provide qualified independent evaluators with access resembling that available to employees.

Altman publicly supported the evaluator proposal and said OpenAI would adopt a similar measure. The company has not yet published the evaluators’ selection process, technical access, confidentiality rules or authority over model deployments.

Independent access could involve exposure to internal models, testing tools and security-sensitive information. OpenAI has not specified how it would separate external review from access controls protecting proprietary technology and user data.

Elon Musk supported Amodei’s warning with the brief statement, “Dario is right.” His endorsement did not include a technical framework or a commitment describing how xAI would apply the proposed controls.

The XRP Ledger has processed 3,254 transactions in one ledger, setting a reported single-ledger record on Sept. 14.

Validator operator Vet reported the figure after reviewing the ledger and said most entries were one-drop XRP payments. He described the activity as a possible throughput test, although the sender’s purpose has not been confirmed.

A drop is one-millionth of one XRP, making it the smallest unit recorded by the network. The high transaction count therefore represented many small transfers, not an unusually large amount of XRP moving between accounts.

The ledger index and initiating account were not identified in Vet’s public post. Without those details, the record claim relies on his analysis and cannot be compared through the post alone with every earlier ledger in XRPL history.

Tiny XRP payments dominated the record ledger

Most of the 3,254 transactions were simple payments carrying one drop of XRP, according to Vet. Simple native-asset transfers require less processing work than transactions involving decentralized exchange orders, NFTs or cross-currency payment paths.

“I don’t know why this person is doing these transactions, but it looks like throughput testing, probably,” Vet said. The description remains speculative because the account owner has not publicly explained the activity.

Transaction count does not show how much computational work a ledger required. A ledger containing thousands of direct XRP payments can place a different load on validators than one containing fewer trades, token operations or complex payment paths.

“Not all transactions are equal in load footprint,” Vet said. He estimated that 500 simple XRP payments could create less stress than 200 transactions that require extensive decentralized exchange processing.

Official XRPL documentation states that each validated ledger records the transactions applied to the preceding ledger state. The associated metadata provides the result and effects of each included transaction.

Transactions with a tesSUCCESS result completed their requested action. Entries carrying a tec result remain recorded and consume a fee, even when they fail to perform the requested operation. The reported total of 3,254 therefore describes included transactions, not necessarily 3,254 successful transfers.

XRP Ledger capacity uses an adaptive target

The XRP Ledger does not use one permanent transaction limit for every ledger. Its servers adjust operating conditions in response to transaction volume, network latency and consensus performance.

Vet said the network can raise its soft transaction target when a heavily loaded ledger closes within the expected period. When close times move beyond the preferred range, the network can reduce the target to help validators return to normal timing.

XRPL documentation says servers exchange proposals until trusted validators agree on a transaction set. Each server then calculates the new ledger state and distributes a signed validation containing the resulting ledger hash.

A supermajority of trusted validators must agree on the same hash before the ledger becomes validated. Once validated, its transactions and resulting state become final parts of XRPL’s ledger history.

The 3,254-transaction result consequently provides evidence that validators agreed on a ledger carrying that number of entries. It does not establish a new permanent throughput rate, because sustained capacity depends on transaction complexity, hardware, network conditions and consecutive ledger close times.

Throughput measured from one ledger differs from transactions per second over an extended period. A short burst can place many pending payments into a single ledger, while the following ledgers may return to normal activity.

No performance report from Ripple, the XRP Ledger Foundation or the network’s reference software maintainers had confirmed a permanent capacity change following the record. No service interruption or failed consensus round was reported in connection with the burst.

Recent activity has included heavier payments and trading

The record occurred after a period of increased XRPL payment and trading activity. In related coverage, XRP Ledger order-book volume rose 79% year over year during the second quarter of 2026, according to an Evernorth report.

Average daily order-book volume reached 3.57 million XRP during the quarter, while the number of daily traders fell from 1,864 to 1,111. Evernorth said average volume per trading account nearly tripled during the same comparison period.

Stablecoin transfers have created another source of network use. As crypto.news reported, RLUSD generated approximately $9 billion in first-half transfer volume on XRPL during 2026.

Such activity is separate from the one-drop transfers identified in the record ledger. No evidence cited by Vet connected the 3,254 transactions to RLUSD, institutional settlement, exchange trading or customer payments.

The sender could have been testing transaction submission, ledger packing or another technical process. The available account pattern does not confirm whether the activity came from a developer, institution, automated service or individual user.

BatchV1_1 moves through the amendment process

The record arrived while validators were considering protocol features introduced with version 3.3.0 of rippled, the network’s reference server software. The XRP Ledger Foundation released version 3.3.0 on Aug. 6.

Its proposed features include BatchV1_1, ConfidentialTransfer, DynamicMPT, PermissionDelegationV1_1 and Sponsor. Each feature follows the XRPL amendment process before it can become active across the main network.

BatchV1_1 would allow multiple transactions to be bundled and processed together. Official XRPL records say it replaces the earlier Batch amendment after developers found a critical bug in the original implementation.

The feature does not explain the 3,254-transaction ledger because BatchV1_1 had not completed mainnet activation when the activity occurred. Its presence in the server release means validators can review and vote on the amendment.

XRPL’s amendment rules require more than 80% support from trusted validators for two continuous weeks. If support falls to 80% or lower before the period ends, the countdown resets.

ConfidentialTransfer would introduce shielded balances and transfer amounts for Multi-Purpose Tokens while providing viewing mechanisms for authorized parties. DynamicMPT would permit issuers to change selected token settings unless they make those properties permanently immutable.

PermissionDelegationV1_1 replaces an earlier delegation feature that developers disabled after finding a critical bug. The updated amendment would let XRPL accounts assign limited permissions to other accounts after validator approval.

No activation date is guaranteed for amendments still under voting. Validator operators can change their votes, and the network checks amendment support around flag ledgers, which occur approximately every 15 minutes.

Key takeaways

• Cardano trades near $0.207 after falling more than 8% during the previous week.
• ADA’s long-to-short ratio of 0.88 indicates bearish positioning among derivatives traders.
• The funding rate turned positive at 0.0052%, showing a mild bullish bias despite the elevated short positioning.
• ADA must hold the $0.199-to-$0.200 support zone to avoid a decline toward $0.195, $0.173, or $0.150.

Cardano (ADA) traded near the critical $0.200 support zone on Monday after declining more than 8% during the previous week.

Although ADA remains above two important short-term moving averages, mixed derivatives data and sell-side pressure from large traders suggest that recovery attempts could encounter resistance at higher levels.

ADA long-to-short ratio signals bearish sentiment

Cardano’s derivatives indicators present a cautious and somewhat conflicting outlook.

CoinGlass data shows ADA’s long-to-short ratio at 0.91, close to its lowest level in a month. A reading below one indicates that traders hold more short positions than long positions, reflecting expectations of further price weakness.

However, Cardano’s funding rate turned positive on Monday, reaching 0.0052%. Positive funding means long-position holders are paying shorts, suggesting that some traders are positioning for a price recovery.

The contrast between the bearish long-to-short ratio and positive funding rate points to uncertainty rather than a clear directional consensus.

CryptoQuant’s market summary also signals caution. Large whale orders are appearing in ADA’s futures market, but sell-side activity remains dominant. 

Both spot and futures markets are also showing signs of increased trading activity or “heating,” while several other indicators remain neutral.

Together, these factors suggest volatility could rise around the current support area. However, the dominance of large sell orders leaves Cardano exposed to further downside if buyers fail to defend $0.200.

Cardano holds above key moving averages

ADA traded at approximately $0.207 on Monday, remaining slightly above its 50-day exponential moving average at $0.199 and its 100-day EMA at $0.200.

Holding above these indicators gives Cardano’s short-term technical structure a mildly constructive tone despite the broader downward trend.

The Relative Strength Index stands at 50, indicating balanced momentum and consolidation. Meanwhile, the Moving Average Convergence Divergence indicator remains slightly negative, showing that bullish momentum has not yet strengthened enough to confirm a recovery.

Cardano’s immediate resistance sits at the 50% Fibonacci retracement level of $0.213. A move above that level could allow ADA to test the 61.8% retracement at $0.231.

Additional barriers are located at $0.236, the 200-day EMA near $0.240 and the horizontal resistance level at $0.245. ADA would need to break decisively through this cluster to improve its medium-term outlook.

A sustained move above $0.245 could bring the more distant $0.299 resistance level into focus.

Conversely, losing the 50-day and 100-day EMAs near $0.200 would expose the 38.2% Fibonacci retracement at $0.195. A deeper correction could then target the structural support levels at $0.173 and $0.150.

Key takeaways

• Pi Network traded above $0.097 on Monday after recording gains for two consecutive weeks.
• The Pi Core Team released SoloHost updates and Pi Desktop version 0.6.3 to improve reliability, app discovery, and developer tools.
• PI remains above its 50-day EMA at $0.094, while the RSI near 60 signals improving bullish momentum.

Pi Network (PI) extended its recovery on Monday, trading above $0.097 following two consecutive weeks of gains.

New ecosystem updates and improved developer tools are strengthening the network’s utility. However, PI continues to trade below major long-term exponential moving averages, leaving its broader technical outlook bearish despite improving momentum.

Pi Network releases SoloHost and Desktop updates

The Pi Core Team recently released updates for SoloHost alongside version 0.6.3 of Pi Desktop.

SoloHost updates have been released alongside Pi Desktop version 0.6.3! This improves app discovery, reliability, and developer tools.

Pioneers can more easily discover actively used SoloHost apps. Reliability improvements and developer resources, such as AI agents… pic.twitter.com/67SXkf4CMp

Advertisement

— Pi Network (@PiCoreTeam) September 10, 2026

According to the project, the updates improve application discovery, platform reliability, and the tools available to developers building within the Pi Network ecosystem.

Continued development could support greater application activity and expand PI’s utility. These improvements have accompanied the token’s recent recovery, with PI gaining 1.68% last week before extending its advance on Monday.

However, ecosystem developments will need to translate into sustained user activity and demand for the token to support a stronger long-term recovery.

PI holds above the 50-Day EMA

Pi Network traded near $0.097 on Monday, remaining slightly above its 50-day exponential moving average at $0.094.

Holding above this indicator gives PI’s short-term outlook a mildly bullish tone. The Relative Strength Index stands near 60, indicating that buying momentum is improving without reaching overbought territory.

The Moving Average Convergence Divergence indicator is also marginally positive, supporting the possibility of further short-term gains.

Nevertheless, the recovery remains tentative because PI continues to trade below its 100-day and 200-day EMAs. These indicators currently stand near $0.105 and $0.138, respectively, preserving the token’s broader bearish structure.

PI’s first major resistance is located at the 100-day EMA near $0.105. A sustained break above this level could strengthen the recovery and open the way toward the horizontal resistance at $0.118.

Beyond that, the 200-day EMA near $0.138 represents a more substantial obstacle. Reclaiming this indicator would be necessary to improve PI’s medium- to long-term technical outlook.

Failure to overcome the $0.105 resistance could leave the token consolidating around its current level or expose it to renewed selling pressure.

Immediate support sits at the 50-day EMA near $0.094. Holding this level would preserve PI’s improving short-term structure and allow buyers another opportunity to challenge overhead resistance.

A decisive decline below $0.094 could weaken the recovery and bring the horizontal support at $0.075 into focus.

If selling pressure intensifies, PI could revisit the former trendline-break area near $0.045, which represents a deeper structural support level.

US spot Bitcoin exchange-traded funds (ETFs) saw a sharp weekly reversal, posting $462.7 million in net outflows after three straight weeks of inflows. The shift came across all four trading sessions from Tuesday through Friday, according to data from Farside Investors, with the selling extending the slide into a fourth consecutive day.

Ether ETFs moved in the opposite direction over the same period, recording nearly $197 million in net inflows. Farside Investors reported that US spot Ether products drew $196.9 million during the four-day window, even as flows were choppy earlier in the week.

Key takeaways

• US spot Bitcoin ETFs recorded $462.7 million net outflows for the four-session week, reversing a three-week inflow streak.
• Bitcoin ETF outflows were broad-based across sessions, with Thursday’s $282.7 million withdrawal the largest daily outflow since July.
• Despite the weekly reversal, spot Bitcoin ETFs remain net positive for September, with $307.3 million in inflows through Friday.
• US spot Ether ETFs delivered $196.9 million in net inflows, led on Friday by BlackRock’s iShares Ethereum Trust ETF with $148.8 million.

Bitcoin ETFs reverse course after three-week inflow run

According to Farside Investors, Bitcoin ETF outflows unfolded across every session from Tuesday to Friday. The withdrawal week followed what the market data described as the strongest three-week inflow stretch of 2026, which ended after funds shed $166.8 million during the first two days of the holiday-shortened week.

The pressure intensified on Thursday, when US spot Bitcoin ETFs posted net outflows of $282.7 million—reported as the largest daily withdrawal since July. Friday’s outflow slowed to $13.2 million, but SoSoValue data indicates it still extended the streak of negative daily flows to four trading days.

That combination—one very large day followed by continued (though smaller) withdrawals—helps explain why the weekly total turned decisively negative. Investors watching ETF flows typically treat these reversals as short-term signals of changing risk appetite, especially when they break prior momentum rather than merely pausing inflows.

Which funds drove weekly Bitcoin outflows

Farside Investors’ breakdown shows that ARK 21Shares’ Bitcoin ETF led the weekly withdrawals with $234.2 million in net outflows. Grayscale’s Bitcoin Trust ETF followed with $129.1 million.

BlackRock’s iShares Bitcoin Trust ETF also saw withdrawals, losing $52.5 million over the week. Fidelity’s Wise Origin Bitcoin Fund recorded net outflows of $50.7 million. Together, the results suggest the reversal was not isolated to a single product—multiple major issuers posted negative weekly flow.

While daily volatility can be normal for ETF baskets, the fact that several large operators posted sizable weekly declines is notable for traders who monitor whether outflows are concentrated (often linked to specific investor behavior) or distributed across the complex.

Broader flow picture: still net positive for September

Even after the weekly reversal, spot Bitcoin ETFs remain in positive flow territory for the month. According to the report, through Friday these products have accumulated about $307.3 million in net inflows for September.

That matters because it changes how the week’s news may be interpreted. A negative week can reflect temporary positioning or macro-driven caution, but a still-positive month indicates that large inflows have not fully disappeared across the broader period. For market participants, the key question going forward is whether the ETF complex can stabilize its daily flows before monthly net gains start to erode.

Ether ETFs turn positive with strong Friday inflows

On the Ether side, US spot Ether ETFs recorded $196.9 million in net inflows over the same Tuesday-to-Friday period, per Farside Investors. The week’s flow pattern looked more uneven earlier, with $24.3 million in outflows on Tuesday, $34.7 million in inflows on Wednesday, and $29.9 million in outflows on Thursday.

Friday marked the turning point. The Ether funds drew $216.4 million in net inflows, flipping the four-day total into positive territory despite the earlier back-and-forth.

Product leadership also differed by day. BlackRock’s iShares Ethereum Trust ETF drove Friday’s inflows with $148.8 million, while 21Shares Core Ethereum ETF added $29.1 million. In practice, such leadership changes can help investors gauge where incremental demand is showing up within the Ether ETF lineup.

Overall, the contrast between Bitcoin’s outflows and Ether’s inflows in the same calendar window underscores how different investor demand can be across the two major spot ETF ecosystems. Rather than assuming flows will always move together, traders often watch whether capital rotates between Bitcoin and Ether depending on positioning, risk appetite, and broader market sentiment.

Looking ahead, the next signals to monitor are whether Bitcoin ETF outflows persist after the Thursday-heavy withdrawal day and whether Ether’s strong Friday inflow momentum can sustain through the following week—because the month-to-date net picture remains supportive for Bitcoin while Ether’s reversal will be tested by the next few sessions’ flow consistency.

A Hong Kong man in his 70s has lost more than HK$13 million ($1.67 million) after a self-described cryptocurrency investment expert contacted him through WhatsApp and directed him to a fake trading app.

Hong Kong police said the case was among more than 40 investment scams recently reported to authorities, with victims losing a combined total exceeding HK$50 million.

The latest case began when the elderly man received an unsolicited WhatsApp message from someone claiming to be a cryptocurrency investment expert from Singapore. The person introduced him to what was presented as a crypto platform offering favorable exchange rates, low fees and withdrawals at any time.

Following the instructions he received, the man opened a cryptocurrency wallet and purchased Tether (USDT) and Ethereum (ETH). He was then told to download an investment application supplied by the scammer and transfer the cryptocurrency to designated wallet addresses as investment capital.

Fake crypto investment app showed profits before withdrawals failed

Once the funds had been transferred, the fraudulent application displayed what appeared to be continuing investment profits, according to police.

Seeing his account balance rise inside the app, the victim became less suspicious and continued sending more cryptocurrency to the wallets provided by the scammer. The fraud only became clear when he tried to withdraw his funds and was repeatedly prevented from doing so under different pretexts.

By the time the victim realized the investment platform was fraudulent, his losses had exceeded HK$13 million.

Police disclosed the case through their CyberDefender social media page while warning residents against unsolicited investment advice and promises of large profits. Authorities urged investors not to download investment applications from unknown sources and to verify platforms through official channels before transferring funds.

The case follows another incident involving a Hong Kong retiree who was targeted through a similar approach earlier this year. In March, crypto.news previously reported that a 66-year-old retired man lost HK$6.6 million across three cryptocurrency scams after fraudsters approached him while posing as investment experts.

One of the schemes began with a WhatsApp message in September 2025. The victim was directed toward cryptocurrency investments and ultimately lost his savings after transferring funds under the scammers’ instructions.

Hong Kong crypto scams have used fake platforms to build trust

Fake balances and apparent investment returns have surfaced repeatedly in crypto fraud cases in Hong Kong.

In August, a woman in the city reportedly lost around $3.3 million after an online romantic partner directed her to a fraudulent crypto platform. The platform displayed supposed returns of more than 800% before withdrawals were blocked.

Hong Kong authorities recorded 25 romance-linked investment fraud cases during the week ending July 30, with combined reported losses approaching $9 million.

A separate investigation into the Fun Coffee investment scheme has involved a larger group of victims. By Aug. 6, police had received 255 reports connected to the Fun Coffee crypto scam, with reported losses reaching approximately HK$104 million.

Participants in the scheme were instructed to download an application, register accounts and transfer cryptocurrency, primarily USDT, to designated wallet addresses. Investors were offered different deposit plans carrying advertised annual returns of roughly 197% to 278%, according to police analysis.

Some users were initially able to withdraw small amounts, which investigators said reduced suspicion and encouraged larger deposits. Withdrawals stopped after the application ceased operating on July 20, while customer service channels stopped responding.

USDT remains common in investment scam payments

USDT has frequently appeared in crypto investment fraud because victims can be instructed to buy the stablecoin before transferring it directly to wallets controlled by scammers.

A Sept. 4 analysis from the U.S. Treasury’s Financial Crimes Enforcement Network linked approximately $12.7 billion in suspicious financial activity to digital asset investment scams largely associated with overseas scam compounds.

FinCEN reviewed 33,904 Bank Secrecy Act reports filed between September 2023 and December 2025. Money services businesses, most of them cryptocurrency companies, accounted for 55% of the reports and identified $5.5 billion in suspicious activity, while banks reported another $6.4 billion.

The agency found that scammers used at least 22 digital assets. Proceeds were commonly converted into stablecoins and almost exclusively into USDT before being transferred through decentralized finance protocols or overseas exchanges.

Hong Kong authorities have meanwhile continued warning residents about fraudulent websites and applications designed to imitate legitimate financial services. In July, Hong Kong Interbank Clearing Limited identified counterfeit websites using virtual wallets and cash reward offers to obtain users’ personal and banking information.

The fraudulent sites falsely presented themselves as connected to official services and attempted to persuade users to complete purported identity verification procedures. HKICL said the websites had no connection to the clearing company.

Police, in their latest warning, told residents not to trust people presenting themselves as investment experts with supposed methods for generating large profits. Investors were advised to avoid unknown investment applications and confirm a platform’s authenticity through official sources before committing funds.