SQD says its enterprise arm, SQD 360, has added validated onchain data from 10 blockchain networks into Google Cloud BigQuery via Google Cloud Web3 Blockchain Analytics. The datasets cover each network’s genesis-to-present history, enabling analysts and developers to query blocks, transactions, logs, traces, and state changes in SQL without running their own indexers.
Before loading, SQD applies six cryptographic checks to each block—using comparisons across multiple sources and verifying transaction roots and state roots—to reduce missing, incorrect, or inconsistent records. Once in Google Cloud BigQuery, the data is available in the same cloud environment used for BI, machine learning, and large-scale analytics.
SQD did not disclose which 10 chains are included or a timetable for adding more networks and AI-agent tooling. However, it framed the integration as a step toward broader agent-based access to verified records inside Google Cloud. SQD CEO Wanja Oberhof said partnering with Google Cloud signals “enterprise-grade” blockchain data has arrived.
For crypto traders, this is mainly a data-infrastructure development: it can improve the reliability and speed of compliance, risk monitoring, and onchain research workflows, but it is not an immediate protocol or token-economics change. Overall impact on market stability is likely limited unless wider adoption accelerates demand for SQD services and data-driven monitoring.
Neutral
Google Cloud BigQueryOnchain dataSQDWeb3 Blockchain AnalyticsEnterprise infrastructure
Cosmos Labs has urged Cosmos EVM operators to halt operations after a security incident linked to the shared Cosmos EVM infrastructure. The issue is not chain-specific code, raising systemic risk for multiple Cosmos EVM networks.
Key updates: (1) KiiChain said an attacker drained 148,326,583.15 KII by repeating the same technique 18 times on Aug. 22. The chain halted at block 9,355,723 after internal detection, freezing remaining funds. KiiChain traced the root cause to a Cosmos EVM staking precompile underflow plus two additional undisclosed bugs, and warned vesting-account configurations could be impacted. (2) TAC reported a separate exploit of the Cosmos EVM precompile-layer flaw on the same day, draining 2,985,651,403 TAC from one account. TAC halted its network to stop further loss. (3) MANTRA paused Layer 1 for about 30 hours as a precaution, later saying the threat was contained and user balances were unchanged.
Cosmos Labs said it requested coordinated validator stops across Cosmos EVM teams and will publish an incident report after resolution. KiiChain criticized patch handling, noting a fix posted Aug. 19 was not clearly flagged as critical and no halt recommendation was bundled when it reached affected networks.
For traders, the Cosmos EVM security crisis can increase near-term volatility and liquidity risk around KII and TAC, especially if additional halts or delayed incident details emerge.
Two former Google DeepMind researchers, Rishub Jain and Josh Jacob, launched London-based nonprofit Sampura Research on Aug. 25 to improve AI oversight. The group secured $11M from Coefficient Giving: $7M for year one and $4M pledged for later.
Sampura’s core idea is “Human-AI Complementarity for Scalable Oversight.” It will build evaluation systems that scale with increasingly capable models. Rather than relying only on automated testing or human reviewers, Sampura plans to deploy “judges” combining human evaluators, AI components, or both across the AI lifecycle—training, evaluation, and real-world deployment—to catch failure modes that pure approaches can miss.
The funding targets known evaluation risks such as reward hacking, where models game their metrics (for example, sounding confident without being correct). Sampura’s hybrid evaluation approach aims to make such gaming harder by leveraging different blind spots across human and automated assessments.
Founders previously worked at DeepMind on high-profile projects including AlphaFold and on AI safety research at Google. Sampura will recruit founding technical staff with AI research experience, building on related DeepMind initiatives from 2023 (SPAR and MARS fellowship programs).
For crypto traders, this is a governance-and-safety development with no direct token impact, but it highlights growing institutional focus on credible AI evaluation and oversight.
Neutral
AI oversightAI evaluationReward hackingDeepMindTech regulation
Richmond Fed President Thomas Barkin warned that rising US debt could eventually deter investors from buying US bonds. He described the current debt trajectory as a persistent “wind” that the Fed must navigate. The 30-year Treasury yield has surged to levels not seen since before the 2008 Global Financial Crisis, reflecting higher fiscal concerns and lingering inflation risk.
The US national debt has crossed $40 trillion (August 2026), with publicly held debt nearing 100% of GDP. The Congressional Budget Office projects sharply higher interest costs, creating a feedback loop: more debt raises interest payments, which then increases the need for additional borrowing.
To ease pressure at the long end of the curve, the Treasury has begun buybacks of longer-dated securities, seen by analysts as a temporary fix. Barkin’s concern is that weak demand for US bonds at these yield levels would force the market to price greater duration risk, potentially lifting yields across assets and weighing on equities and corporate credit.
Overall, Barkin argued monetary policy tools can’t solve a structural debt problem. The Fed can adjust rates and its balance sheet, but the market may increasingly demand higher yields, complicating inflation control and tightening financial conditions.
Bearish
US bonds30-year Treasury yieldsFederal Reservefiscal riskinflation outlook
Bitcoin’s Bull Score (CryptoQuant) jumped from 30 to 80 in one week, its highest reading since October 2025, pushing the market into an early bull-market zone. CryptoQuant says 8 of 10 indicators now signal bullish conditions, supported by rising spot and futures demand.
For traders, the confirmation trigger is price—not headlines. CryptoQuant requires a weekly close above the 365-day moving average, currently around $83,000. A quick move above $80,000 is not enough; the weekly close is the filter between “recovery” and a confirmed bull cycle.
Price context: BTC rallied more than 24% from below $64,000, briefly trading above $80,000, and was around $79,000 at the time of writing after giving back part of the move.
Where resistance may sit: LMAX Group strategist Joel Kruger flagged the May 2026 peak around $82,820 and the ~$83,000 area as the first key test. A sustained breakout above that level would support a shift toward a next leg that could challenge $100,000 and higher.
But risk of pullbacks is rising. CryptoQuant highlighted overheating signals: traders’ unrealized profit margin rose to 20.5% (highest since June 2025), and short-term holder “whales” booked about $1.2B in realized profits between Aug. 20–22. Exchange deposits also rose by roughly 53,000 BTC, which can increase potential sell-side supply.
Market support: U.S. spot Bitcoin ETFs recorded about $1.9B net inflows for the week ending Aug. 21, with continued demand reflected in daily inflow prints (e.g., Aug. 24).
Bottom line for the Bitcoin Bull Score: bullish momentum is building, but traders should watch for acceptance above the ~$83,000 weekly close to reduce the odds of a post-rally pullback.
Crypto-native exchange data from Phantom shows traders are overwhelmingly bullish on Nvidia into earnings. On Phantom, 88.7% of positions in the NVDA-USD perpetual market are long as the chipmaker’s earnings approach. This NVDA setup is widely traded on the platform as a synthetic equity contract that settles in USDC, with up to 10x leverage.
The bullish skew looks persistent. In late June, Phantom reported 89.3% long, so sentiment has shifted by less than 1 percentage point over roughly two months—suggesting traders on Phantom have stayed consistently positioned for upside.
The main trading implication is leverage risk. A 10x long means a 5% move in NVDA can translate into about a 50% gain or loss for the position. If earnings disappoint, liquidation cascades could accelerate downside as leveraged longs hit liquidation thresholds and trigger automated sell-offs. Perpetual futures also run continuously and may lack traditional exchange circuit breakers, raising the chance of sharp, after-hours volatility.
Bottom line for traders watching Phantom: the tape points to strong near-term long positioning in NVDA ahead of earnings, but downside scenarios could be amplified by 10x leverage and perpetual-futures liquidation dynamics.
Researchers from the University of Edinburgh and NVIDIA unveiled Dynamic Memory Sparsification (DMS), a KV cache compression method that reduces inference-time memory by 8x without degrading quality. DMS works by selectively keeping only the most useful tokens in the key-value (KV) cache, cutting KV cache size to one-eighth while enabling deeper reasoning within the same compute budget.
Results reported across major tests were strong: on AIME 24, DMS-compressed models scored about 12 points higher; on GPQA Diamond (graduate-level science), scores rose by more than 8 points; and on LiveCode Bench (practical coding), models gained around 10 points even while processing the same amount of KV cache data. Lead researcher Dr. Edoardo Ponti said models can “reason faster but with the same quality.”
The work was presented at NeurIPS and detailed in “Inference-Time Hyper-Scaling with KV Cache Compression.” Evaluations were conducted using Llama and Qwen models, pointing to cheaper, more efficient deployment of reasoning-capable LLMs on edge devices such as wearables, smart home hardware, and other resource-limited platforms.
For traders, the immediate link to crypto markets is indirect: the news signals potential cost-down and efficiency gains in AI infrastructure, but no direct token, protocol, or chain is affected.
Neutral
AI inference optimizationKV cache compressionLLM efficiencyNeurIPS researchEdge AI deployment
Ethereum (ETH) is extending a strong uptrend, up roughly 30% over about a week and printing a seven-month high above $2,500. The move is being driven by spot Ethereum ETF inflows, a large derivatives reset that forced major short liquidations, and macro tailwinds as long-end yields ease ahead of US Treasury policy actions.
As ETH trades around $2,478–$2,500+, momentum looks stretched. Daily RSI is about 79 (overbought), and liquidity/leverage cluster in the $2,500–$2,550 area, where trading flows can quickly flip from short-covering to profit-taking.
Key levels for ETH traders:
- Upside: liquidity and liquidation pockets at $2,540–$2,570. A sustained break above $2,550 increases odds of continuation toward $2,600, then $2,700 and possibly $3,000.
- Downside: if ETH rejects near $2,500, expect pullbacks toward $2,410–$2,450, with near-term support also seen around the 4-hour Supertrend near ~$2,344 and daily MA zones around ~$2,000–$2,080.
Bottom line: ETH remains bullish on trend, but the $2,500–$2,550 zone is the immediate battleground. Watch for acceptance (weekly strength above ~$2,550) versus rejection, which could reopen a liquidity-driven retracement.
Bitcoin open interest collapses to about 12% as crypto-margined Bitcoin futures lose dominance. Glassnode data shows crypto-collateralized BTC futures are now roughly 12% of total open interest across exchanges, down from nearly 100% in 2019–2020. The shift matters because crypto margin shrinks during drawdowns, increasing liquidation and margin-call risk.
In the latest move, Bitcoin rebounded from around $57,000 to a weekly close near $79,175 and was up about 1.88% on the day. But the 24-hour liquidations were aggressive: about $570.08 million total, with shorts hit harder than longs ($329.60M shorts vs $240.48M longs). CoinGlass cited a $103.54M BTC position on Bitget as the biggest single blowup.
The article frames these as related-but-separate forces. The liquidations look like a classic short squeeze, while the long-term structural trend is the steady replacement of crypto collateral with stablecoin (dollar-denominated) margin. That does not eliminate leverage risk: leverage “is leverage” regardless of whether margin is in BTC or stablecoins. Traders should therefore avoid assuming the squeeze is automatically over just because crypto-collateral share fell to 12%.
Federal Reserve Chairman Kevin Warsh will deliver his inaugural Jackson Hole keynote on August 28, 2026. The event is set for 10 a.m. ET and comes as fixed-income markets face elevated volatility, with persistent inflation concerns and shifting US fiscal expectations.
Warsh has signaled a deliberate change in communication style for the Federal Reserve: less forward guidance and fewer coded signals. Instead, markets should respond more “organically” to incoming economic data. This approach was already previewed in his June 2026 post-meeting press conference and subsequent congressional testimony.
For the Jackson Hole speech, Warsh is expected to focus on longer-run, big-picture drivers such as productivity trends, demographic shifts, and structural economic forces—rather than granular policy details. Traders are watching both the content and the reaction: if bond yields swing sharply even without explicit policy messaging, it could imply that removing forward guidance itself increases instability.
Bond-market nerves are also tied to broader macro uncertainty. Treasuries have been whipsawing amid inflation and fiscal-policy concerns. Warsh was confirmed by a 55–45 Senate vote after President Donald Trump nominated him earlier in 2026.
Key takeaway for traders: this Jackson Hole keynote is a real-time stress test for how the Federal Reserve’s reduced signaling affects rates volatility, risk pricing, and broader market liquidity.
Neutral
Federal ReserveJackson Holebond yieldsforward guidancemacro volatility
South Korea’s POSCO International (POSCO Holdings’ trading arm) completed an August 25 pilot to tokenize real trade receivables on the Avalanche blockchain. The receivables were purchased with actual capital by Olea, a trade-finance firm backed by Standard Chartered’s venture arm, SC Ventures.
The project used POSCO International America and Intain (a Layer-1 network built on Avalanche infrastructure). First, AI verified trade documents such as invoices and shipping records for consistency. After verification, the receivables were registered on Intain’s network to create an immutable ledger of ownership and transaction history. Olea then acted as both the financing platform and the capital provider for the tokenized trade receivables.
This is POSCO International’s second blockchain receivables pilot in a month. On July 27, a proof-of-concept with LG CNS tokenized receivables on the Injective blockchain, also using live trade data from POSCO’s overseas subsidiaries (not synthetic test data). POSCO International reported $22.2B in revenue and operates 80+ branches across industries including steel, energy, and battery materials.
The article links the use of tokenized trade receivables to the wider trade-finance gap (Asian Development Bank estimate: about $2.5T), arguing that AI can reduce document-verification friction while blockchain improves trust between counterparties. The teams also signaled interest in stablecoins for cross-border transactions and digital treasury tools.
US Treasury Secretary Scott Bessent expanded the US Treasury bond buyback plan to curb selling pressure in 10–30 year Treasuries. From Sept. 9 to Nov. 4, the single-operation ceiling for the US Treasury bond buyback is lifted from $2B to at least $4B, aiming to reduce available supply in longer-dated, less-liquid securities and improve liquidity.
Long-term Treasury yields hit the highest levels in nearly 20 years, keeping fiscal/debt-service costs elevated. After the announcement, yields briefly fell but then reversed, with traders questioning whether this US Treasury bond buyback is mainly “price management” rather than true liquidity support. Critics also note the scale may still be small versus the ~$32T Treasury market.
Funding context: the US Treasury’s General Account at the Fed holds around $940–$950B, potentially allowing flexibility beyond the published $4B cap while keeping auction schedules unchanged.
Crypto-trader takeaway: persistent higher long-end yields can lift refinancing costs and weigh on risk assets. If the US Treasury bond buyback fails to stabilize yields, expect renewed volatility; if it overshoots the cap meaningfully, short-term relief is possible but may be fragile.
Bearish
US Treasury bond buybacklong-term yieldsliquidity supportfiscal impactrisk assets
The Crypto Fear & Greed Index has flipped to “extreme greed” at 81 for the first time since late 2024. CoinMarketCap’s gauge jumped from 41 to 81 within a week (36 a month ago) and was still at 81 on Aug. 25, crossing the 80 threshold.
The move tracks a Bitcoin-led rally. BTC gained about 24% in seven days, briefly pushed above $80,000, and kept its dominance near 59.7% as total crypto market cap rose to roughly $2.67T (+23.8% on the week).
Key catalyst: a leveraged short squeeze. Coverage citing CoinGlass data estimated about $2.7B in bearish crypto liquidations over 24 hours, with more than $1B of BTC shorts closed in roughly one hour (around 92% of liquidations).
Macro/flow support also played a role. U.S. Treasury plans to expand 10–30-year bond buybacks (effective Sept. 9) were linked to a weaker dollar and more risk-on positioning toward BTC. Spot Bitcoin ETFs added demand, with reports citing about $1.9B inflows over five trading days ending Aug. 21 (daily net inflows around $517M on Aug. 19 and $606M on Aug. 20). Spot Ether ETFs added about $221M.
Alternative.me’s sentiment metric stayed in “greed” (methodology difference), but CoinMarketCap warns that extreme greed can signal an overheated market and higher pullback risk if ETF/inflows slow.
Bullish
Crypto Fear & Greed IndexExtreme GreedBitcoin RallyShort LiquidationsSpot Bitcoin ETFs
InvoXYZ, a social copy-trading app built on Hyperliquid, has overtaken Trust Wallet to claim second place in Hyperliquid builder code volume rankings tracked by HyperTracker. Over the past 30 days, Invo recorded $1.49B trading volume, up 28%, and reached 40,801 unique traders—more than double the previous #2 front end.
Invo’s growth is tied to two factors: copy-trading on Hyperliquid’s perpetual markets and aggressive user acquisition via TikTok. The app lets users mirror top-performing traders and supports 170+ trading pairs. HyperTracker credited Invo’s “user-led growth initiatives,” noting its emphasis on social features rather than traditional wallet integrations.
Hyperliquid builder codes are referral tags embedded in the protocol. When a front end routes trades, it stamps each transaction with its builder code. HyperTracker aggregates this data to show which interfaces drive real economic activity, not just downloads. Invo’s builder code is INVO, and its $1.49B figure reflects perpetual futures volume routed through the platform.
Risks remain: copy-trading concentrates leverage exposure, so follower drawdowns can compound quickly if the mirrored trader underperforms—especially with younger users potentially underestimating perpetual liquidation mechanics.
For traders, the key takeaway is the shift in Hyperliquid builder code volume toward social/trading front ends. This can increase attention and liquidity on Hyperliquid in the near term, while copy-trading-driven churn may add short-term volatility to follower segments.
Neutral
HyperliquidBuilder Code VolumeSocial Copy TradingPerpetual FuturesDeFi User Growth
RockawayX has begun seeking $150 million for a new liquid opportunities fund after acquiring crypto hedge fund Relayer Capital, according to Forbes.
The reported vehicle will target “undervalued tokens and crypto-related equities,” aiming for exposure to assets that can be traded more easily than RockawayX’s existing private venture holdings. Relayer founder Austin Barack is expected to manage the new fund.
Forbes cited sources saying Relayer returned about 70% in 2026, with performance linked to positions in Hyperliquid (HYPE) and Venice AI (VVV). Hyperliquid’s HYPE token reportedly gained 219% year-to-date at the time of the report, while Venice AI’s VVV rose 1,006%. RockawayX and Relayer have not published audited results or full trade/position details, and the article notes key fundraising terms—timetable, minimums, fees, and eligible jurisdictions for the new fund—remain undisclosed.
RockawayX says it oversees about $2 billion across investment, liquidity, and blockchain infrastructure divisions, including market-neutral strategies and validator/liquidity operations. The firm is also expanding beyond pure venture investing with additional credit and liquidity approaches.
For traders, this is a read-through on potential future demand for higher-liquidity crypto assets—especially tokens aligned with the reported Relayer track record—though the timing and exact allocation cannot be confirmed yet.
US domestic banks’ commercial and industrial (C&I) loans rose sharply, up about $247B year over year to roughly $2.93T—the highest level of business lending since June 2020. Federal Reserve H.8 data shows C&I loans were flat for around 18 months from mid-2023, but the pace accelerated in 2026. In the first half of 2026, year-to-date additions reached about $185B, with February alone adding $50.43B.
By mid-May, outstanding C&I loans hit ~$2.89T, then climbed past $2.93T by mid-August (still about $180B below the May 2020 peak of $3.07T). Large banks reported gains: Wells Fargo +8.3% QoQ in Q1 2026, PNC +6.4% over the same period, and JPMorgan Chase also posted strong sequential increases.
A key dynamic is a rotation away from private credit and back toward traditional bank financing. As banks compete more on pricing, private credit’s cost/terms advantage has weakened for corporate treasurers. The February surge also hints at new capex plans—equipment, facilities, and strategic investments—plus ongoing working-capital needs amid inflation, partially offset by softer labor-cost pressure.
For crypto traders, this is a macro liquidity and credit-cycle signal: improving credit availability can support risk appetite, but it may also coincide with higher rates/real-economy capex that affects broader capital flows.
Neutral
US bankingC&I loansprivate creditmacro liquidityrate environment
The Crypto Fear and Greed Index, maintained by Alternative.me, has jumped from 41 (one week ago) to 74 today—its first “extreme greed” reading since late 2024. In February 2026, the index hit a record low of 5, then staged one of the fastest recoveries in its history.
As the composite score rises, some platforms report readings above 80. CoinStats’ alternative version is pegged near 83. Historically, sustained “Crypto Fear and Greed Index” levels above 75 have often come before local tops and subsequent corrections.
The index is built from five components: volatility (25%), market momentum & volume (25%), social media sentiment (15%), Bitcoin dominance (10%), and Google Trends search activity (10%). The current surge is attributed mainly to diminishing volatility, strengthening momentum, and higher social engagement—while the previously included survey component remains paused.
Key trading takeaway: watch volatility and momentum closely. If volatility starts expanding while momentum stalls, the Crypto Fear and Greed Index can reverse quickly—similar to past episodes where a rapid 30+ point move over weeks was followed by heightened volatility and pullbacks.
Bearish
Crypto Fear and Greed IndexMarket SentimentExtreme GreedVolatilityBitcoin Dominance
US Treasury bond buybacks are being doubled as US national debt crosses $40 trillion. On Aug. 19, Treasury Secretary Scott Bessent said the Treasury will raise liquidity-support buybacks of longer-dated securities from $2B to at least $4B per operation, effective Sep. 9 through Nov. 4.
The goal is to absorb less actively traded long bonds and ease pressure on yields. Before the announcement, the 30-year Treasury yield rose to about 5.33%—the highest level since 2007. Markets initially dipped after the news, but yields later “un-dipped,” implying the intervention may be limited.
Analysts likened the US Treasury bond buybacks to a “band-aid.” The underlying issue is a persistent federal budget deficit of roughly $2T per year (about 6.4% of GDP). Because the government keeps issuing Treasuries to fund the gap, supply continues to climb and upward yield pressure persists.
Net interest payments on the debt are projected to exceed $1T (some estimates up to $1.2T), potentially rivaling or exceeding major federal spending lines. The article also notes shifting demand: foreign central banks have been reducing Treasuries holdings, leaving more of the burden to domestic investors who typically require higher yields for long-duration risk.
Bottom line for traders: stronger US Treasury bond buybacks may reduce volatility, but they do not change the deficit-driven yield trend. Fiscal consolidation would matter most; recent tax policy reportedly extends deficit pressure.
Bearish
US TreasuryBond yieldsFederal deficitRate-sensitive marketsLiquidity/market plumbing
Goldman Sachs backed Coinbase and Robinhood, citing Bitcoin’s breakout above $80,000 and a potential recovery in crypto activity in H2 2026. In its Americas Brokerage and Crypto Industry report, the bank said crypto trading volume fell 30% in July and another 21% in August, with activity still far below recent peaks. However, total crypto market cap rebounded about 21% to around $2.8T, and Goldman expects volumes to improve if valuation holds.
Regulation also remains central. Goldman highlighted that uncertain rules are the biggest barrier for institutional investors (35% cited unclear regulation; 32% cited regulatory clarity as a catalyst). It pointed to the SEC’s proposed “Regulation Crypto Assets” framework (with exemptions for qualifying startups raising up to $5M over four years, and other pathways up to $75M over 12 months), alongside broader market-structure momentum such as the US CLARITY Act.
On equities, Goldman maintained buy ratings: Coinbase price target raised to $196 (from $173) and Robinhood set at $124. The thesis includes products beyond spot trading—tokenized stocks, prediction markets, and derivatives—to support revenue when spot volume is weak.
Goldman also disclosed $86.5M exposure across five spot XRP ETFs via its Q2 filings, after having reported no XRP ETF positions in Q1. Bitcoin’s move (up roughly 26% on the week) helped lift crypto-linked stocks as trading volume rose sharply in the latest 24 hours.
Solana (SOL) is seeing record network usage alongside a sharp price rally. Onchain data cited by The Kobeissi Letter shows Solana processed 4.2 billion transactions in July, a monthly record and up 13.5% vs. June. Weekly non-vote activity also hit a new high, while DefiLlama reported about 109 million transactions and 2.7 million active addresses in a 24-hour window.
Trading and DeFi activity increased with the market move. DefiLlama data showed Solana-based DEX volume reached about $20.14 billion over seven days (up 103% week-over-week), with daily volume near $3.02 billion. Solana stablecoin supply rose to roughly $15.94 billion in the same dataset, with USDC at 44.95% of the stablecoin market.
Real-world assets (RWA) on Solana also expanded. RWA.xyz estimated distributed RWA value at about $3.97 billion as of Aug. 25 (up 11.81% over 30 days), with 1,243 tracked assets and around $125.9 million in represented asset value. Broader tokenized government-debt themes remain significant, though providers’ methodologies differ.
Price action: SOL gained about 40% in eight days and traded above $100 for the first time since February, briefly topping $102.88. Analysts flag nearby resistance around $100 and mixed momentum signals after the run.
Macro catalyst: The US Treasury announced it would at least double long-dated liquidity-support buybacks (from $2B to at least $4B per operation from Sept. 9), coinciding with a strong one-day crypto rally. However, resistance levels and potential momentum cooling could affect follow-through after the breakout.
Mistral (open-weights AI lab) and HUMAIN (owned by Saudi Arabia’s Public Investment Fund) announced plans to build “sovereign AI” infrastructure in Saudi Arabia and across the region. The tie-up is valued in the “hundreds of millions of euros,” targeting data and model control for regulated use cases.
The project spans in-region inference, open models, and frontier Arabic-language models. Early deliverables focus on cybersecurity and voice. Mistral said it will be able to run workloads on HUMAIN’s Saudi data-center capacity, reducing reliance on European and U.S. compute.
The firms also plan a joint go-to-market push in Saudi Arabia’s regulated industries. HUMAIN brings a broader buildout: an NVIDIA partnership for up to 600,000 of the company’s latest systems over three years (including GB300 platforms), an xAI data-center agreement around a 500 MW-plus facility, and an AWS deal to manage up to 150,000 GPUs in a dedicated Riyadh AI Zone. HUMAIN, AMD, and Cisco also formed a joint venture aiming for up to 1 GW of AI infrastructure by 2030.
For traders, the headline is about sovereign AI capacity and deployment rather than direct token implications; it mainly signals long-horizon capital flows into AI compute and regional tech partnerships.
Neutral
Sovereign AISaudi ArabiaAI InfrastructureOpen ModelsCompute Partnerships
US bank profits rose in the second quarter as lending expanded, according to the FDIC. Net income at 4,238 insured lenders climbed 12% to $90.1 billion, despite a rise in unrealized losses.
Earnings were supported by a 1-basis-point increase in the net interest margin to 3.32%. The FDIC’s quarterly assessment suggests banks continued to expand lending while keeping profitability intact.
At the same time, the FDIC noted that unrealized losses increased during the period. That mix—higher US bank profits alongside worsening mark-to-market pressures—signals banks are still funding growth, but balance-sheet risks remain elevated.
Key figures: 4,238 insured lenders; net income $90.1 billion (+12% QoQ); net interest margin 3.32% (+1 bps); unrealized losses increased.
Source reported by Bloomberg; article dated Aug. 25, 2026.
Four of the Federal Reserve’s 12 regional banks urged a July 2026 Fed rate hike via an increase in the discount rate, citing persistent inflation concerns. The Fed Board still voted 9-3 to keep the federal funds target range unchanged at 3.50%–3.75%, underscoring an internal hawkish-vs-dovish split.
Recent inflation data remain sticky. June PCE inflation is estimated at 3.7%, with core PCE at 3.3%. The July CPI rose 3.4% year over year, well above the Fed’s 2% goal. Despite the Fed rate hike call from some regional banks, the central bank chose to hold rates steady while assessing broader economic conditions.
Markets are treating the push as a potential signal for future tightening. Traders increased speculation around a Fed rate hike by the September 2026 meeting, with the article citing a current probability of 33.5% (YES).
Key watch items for traders include upcoming core CPI and PCE releases, plus comments from Fed Chair Jerome Powell. Any shift in market pricing and Fed communications ahead of September could raise or fade expectations for further tightening.
Canada has announced counter-tariffs on more than 700 US-made products, targeting C$27.6 billion in trade value. The measures take effect on September 8 and impose tariff rates of 15%, 25%, or 50% depending on the category.
The action is a response to the US introducing 50% tariffs on about $20 billion worth of Canadian exports under Section 338 of the Tariff Act of 1930 (effective August 22). Three days later, Ottawa retaliated with its own counter-tariffs.
Key impact areas include steel and aluminum, dairy, electronics, and appliances. Steel and aluminum face the highest burden: Canada doubled its existing 25% tariffs on these sectors to 50%, matching the US rate imposed on Canadian goods. The broader triggered US duties covered a wide range of Canadian exports such as dairy, clothing, and building materials.
To cushion the shock, Canada also unveiled a C$7.5 billion support package for small and medium-sized enterprises and workers in industries most exposed to the disruption. Prime Minister Mark Carney and Finance Minister François-Philippe Champagne called the counter-tariffs a defensive necessity, while acknowledging likely downsides for Canadian consumers, including higher prices and reduced choice for US-made products.
Both countries cited Section 338 authority for retaliation and said negotiations are possible to reduce economic fallout amid rising protectionism.
Neutral
trade warcounter-tariffssteel and aluminumtariff rateseconomic support package
Mantle Vault has moved its stablecoin yield product from the Bybit distribution model to native DeFi for Mantle users. Deposits in USDC and USDT are routed into a non-custodial Mantle Vault setup using Fluxion infrastructure, with strategy design by CIAN and yield connectivity provided via Grove’s integration.
The new DeFi Mantle Vault structure targets variable returns by providing exposure to sUSDS (Sky Protocol’s yield-bearing USDS variant) and uses no leverage, aiming to reduce liquidation risk. Mantle also connects payouts to Sky’s savings rate via governance, meaning APY can change over time. At launch, Mantle materials target up to ~6.5% APY.
In parallel, Mantle is launching a token incentive program distributing 5.14M GROVE on top of underlying strategy yield. For traders, this is a DeFi adoption and incentive-driven liquidity signal around Mantle Vault, with potential short-term lift to on-chain stablecoin demand and Mantle ecosystem activity.
Key risks remain smart-contract risk, stablecoin price/liquidity risk, and governance-driven changes to Sky savings rates. For US users, access may depend on Fluxion’s terms and local stablecoin-yield regulations.
Bitcoin price pulled back toward $79,250 after stalling above $81,200. The move came as the daily RSI jumped to 82.44 (deep overbought), while shorter-term momentum began to cool.
On Aug. 25, Bitcoin price traded near $79,250 after briefly reaching about $81,250, retreating roughly 2.5% from the intraday high. The rally followed a ~25% advance from the $63,000–$65,000 area, lifting BTC through $70,000, $72,000 and $78,000 before meeting heavier selling above $80,000.
Catalysts behind the advance included a weaker US dollar, concerns about currency debasement, and strong US spot Bitcoin ETF demand. SoSoValue data showed $337.56M in net inflows on Aug. 24, with BlackRock’s IBIT leading ($208.9M) and Fidelity’s FBTC ($104.6M). However, the next completed ETF flow reading may clarify whether dips are being used for entries or exposure is being reduced.
Technical/positioning cues point to near-term volatility rather than a clear trend break. BTC remains above major daily moving averages (20-day ~ $68.3K and 200-day ~ $69.2K). The 4-hour ADX is easing and Bull/Bear Power has fallen from peak levels, aligning with a higher chance of consolidation.
Liquidation data highlights key liquidity zones: $78,000 (most notable), then $77,200–$77,500 on a stronger break. Upside liquidity clusters sit around $79,700–$80,100, then $80,500, $81,000–$81,700, and $82,200–$82,500. Traders are watching whether Bitcoin price can reclaim and hold $79,200–$80,000; losing $78,000 would increase downside odds toward $76,500–$77,000, and potentially $72,000–$74,000.
Metric, an aggregator-turned-exchange, has launched its own DEX for tokenized stocks, ETFs, commodities, and real-world assets. The key change is pricing: instead of deriving asset value from liquidity pool ratios (typical AMM design), Metric will use Chainlink Data Streams to pull real-time, verifiable reference prices.
Metric says it routed more than $5B in trades across nine chains before stepping into this new phase. In the quarter leading up to July 2026, it processed over $5B in volume, with about $2B in the final month. For liquidity providers, the platform’s pitch focuses on capital efficiency, targeting ~450x monthly capital turnover on active pools.
Why it matters for traders: thin liquidity is common in tokenized equities and commodities, and pool-derived pricing can become stale, widen spreads, and create arbitrage risk. Chainlink Data Streams use a pull-based model where pricing is computed off-chain and delivered on-demand with cryptographic proof. The article also notes the feed’s always-on nature could enable near 24/7 markets for assets that normally move only during traditional exchange hours.
Metric frames external pricing as the missing infrastructure layer for RWA trading—reducing adverse selection for LPs and aligning spot-market mechanics with what oracle-based pricing has already enabled in derivatives on protocols like GMX and dYdX.
The Federal Reserve kept the primary credit/discount rate at 3.75%, but minutes revealed a growing rift over inflation. The Board voted unanimously to hold the discount rate after discount window meetings on July 20 and July 29. However, the parallel FOMC decision on the federal funds target range (3.5%–3.75%) was 9–3, with Beth M. Hammack, Neel Kashkari, and Lorie K. Logan backing a 25-basis-point hike.
Inflation remains above the Fed’s 2% goal. The minutes noted that “several” participants wanted a more restrictive stance, and Chair Kevin Warsh used a “hedged-hawkish” tone, saying tightening may be needed if inflation does not moderate. Interest on reserves was kept at 3.65%. The maintained discount rate takes effect July 30, 2026, and the next FOMC meeting is scheduled for September 15–16.
Markets will focus on whether hawks gain support. If even one additional member dissents next time, the case for a September rate increase becomes more likely. The next seven weeks include at least one CPI release and a jobs report, adding catalysts before the September vote.
Overall, the Fed’s unanimous discount rate hold paired with a contested funds-rate decision signals caution, but the direction of travel depends on incoming inflation and labor data—key drivers for crypto liquidity and risk appetite.
Bearish
Federal Reservediscount rateinflation hawksFOMC minutesrates outlook
A federal jury convicted Las Vegas businessman Brent C. Kovar in a major crypto fraud case tied to his Profit Connect operation. Prosecutors said he raised about $24 million from at least 400 investors from late 2017 to July 2021 by promising fixed annual returns of 15%–30% and a 100% money-back guarantee.
Kovar marketed an “AI + supercomputer” setup for crypto mining and transaction verification and claimed investors’ money was backed by hundreds of millions in crypto reserves. The government said these claims were false: Profit Connect was not profitable and had no real reserves, so the promised returns could not be paid. Instead, investor funds were allegedly used to run the business, cover payroll and personal expenses (including a house), and repay earlier investors while presenting the activity as mining/verification proceeds.
The jury found Kovar guilty on 11 counts of wire fraud, 2 counts of mail fraud, and 2 counts of money laundering. Sentencing is scheduled for Nov. 30, 2026, with a statutory maximum of 280 years in prison.
Separately, a federal jury also convicted Japheth Dillman for wire fraud and conspiracy involving Block Bits Capital. He allegedly misled 20+ investors about automated trading profits using a tool called “Autotrader,” while knowing the algorithm wasn’t working.
For crypto traders, this crypto fraud verdict reinforces enforcement risk against “high-tech/mining” themed schemes and can briefly raise caution around returns marketing and yield-like products.