Baidu's CFO stated that the company's profit from AI investments could eventually be comparable to its traditional search business profits, signaling a shift toward reducing reliance on search revenue and driving growth through AI.

BlackRock’s latest research revisits Bitcoin after a roughly 50% decline from its October 2025 peak and argues that the asset’s role in diversified portfolios remains intact. Using a rolling 10-year analysis through May 29, 2026, the firm found that a traditional 60/40 stock-and-bond portfolio delivered about 9.9% annualized returns with 10.1% annualized standard deviation. Adding a 1% Bitcoin allocation lifted annualized returns to about 10.9% with volatility at roughly 10.3%, while a 2% allocation pushed returns to around 11.8% with standard deviation at about 10.6%. The Sharpe ratio improved from 0.81 to 0.96, and maximum drawdown moved only slightly from -20.3% to -20.9%. The report also sits alongside BlackRock’s direct market experience. Its iShares Bitcoin Trust (IBIT), launched in January 2024, gathered more than $50 billion in assets in less than a year, which BlackRock has described as the largest exchange-traded product launch in history. According to ETF holdings data cited in the article, U.S. spot Bitcoin ETFs now hold about 1.25 million BTC, nearly 6% of Bitcoin’s fixed 21 million supply, with IBIT alone accounting for roughly 775,000 BTC. For corporate finance leaders, the article says the debate is shifting away from whether Bitcoin is simply too volatile to own and toward position sizing, risk contribution, liquidity, drawdown, governance, and capital allocation discipline.

Finmo Tech, a Singapore-based fintech company, took the top honor at the AFA Awards 2026 grand final held in Taipei on Sept. 1, winning the Grand Award for Cross-Border Fintech Unicorn Potential. The company’s main product is TreasuryOS, which it describes as a treasury operating system built to consolidate financial operations that multinational businesses often manage across separate banks, accounting tools and ERP systems. According to Finmo’s website and public information cited in the source report, the platform covers cross-border payments, multi-currency accounts, cash flow visibility, cash forecasting, FX hedging and working capital optimization, with a payment network spanning more than 180 countries. Finmo has also recently launched a conversational AI assistant called MO AI for CFOs and finance teams, designed to support real-time cash management, forecasting, compliance and reporting. Founded in 2021, Finmo obtained a Major Payment Institution license from the Monetary Authority of Singapore in 2023 and now holds related licenses in Singapore, Australia, the United States and New Zealand. The company has raised about $27 million to date, including an oversubscribed $18.5 million Series A completed in February 2025 and co-led by Quona Capital and PayPal Ventures, with Citi Ventures also participating.

A series of developments across enterprise AI, crypto lending, banking research and payment infrastructure is starting to sketch a shared direction for finance. On Aug. 25, Google Cloud rolled out Gemini Enterprise for Financial Services, pushing AI deeper into workflows such as KYC, credit analysis, portfolio monitoring, bond issuance and financial research. Two days later, MoonPay said it had connected PayBox to Kamino on Solana, letting users initiate lending and access yield through interfaces such as Claude and ChatGPT. Around the same time, the Federal Reserve Bank of Dallas published research on what happens to deposit stickiness if deposits can move in real time and AI can automatically compare yields for households and businesses. BIS also referenced Project Agorá, which recently completed a real-money test tied to tokenized deposits and central bank money. Taken together, these events suggest the key threshold for AI in finance is not simply whether models understand finance better. The bigger question is when they are allowed to make bounded financial decisions on a user’s behalf and actually move money. The article argues that the market is not jumping straight from manual finance to fully autonomous finance. Instead, it is building narrower forms of authorization, execution and auditability that let AI act within preset limits. That shift could eventually affect treasury operations, programmable money, deposit competition and the way banks defend customer relationships.

Cango Inc. reported unaudited second-quarter 2026 results for the period ended June 30, saying bitcoin mining remained its main source of revenue while the company continued to push an energy-and-AI compute platform strategy. Total revenue came in at $50.8 million, including $47.4 million from bitcoin mining and $3.4 million from other revenue. The company posted a net loss of $81.6 million, which it attributed mainly to non-cash impairment on mining machines and losses on disposals. Operationally, Cango said it produced 656 BTC in the quarter and had total operating hash rate of 27.58 EH/s as of June 30, including 19.84 EH/s self-operated and 7.74 EH/s leased. Average cash cost per bitcoin fell about 5% from the prior quarter to $73,313, according to the company, following fleet optimization and tighter execution. Cango also said it began selectively using hedging strategies to reduce the impact of bitcoin price swings. Management said the company is prioritizing unit economics over scale expansion in mining. CEO Paul Yu said the Georgia site completed its upgrade in early July and now supports up to 3 megawatts, with room for future expansion. Cango said cabinets have been installed, GPU hardware has been purchased and delivered in batches, and customer onboarding is underway at the Georgia site, with revenue expected to be recognized in the third quarter.

Bitcoin mining company Cango Inc. (NYSE: CANG) on Aug. 31 released its unaudited financial results for the second quarter ended June 30, 2026. The company posted total revenue of $50.8 million for the quarter, including $47.4 million from bitcoin mining. Net loss came in at $81.6 million, which the company said was mainly driven by non-cash impairment of mining machines and disposal losses. As of the end of the quarter, Cango held 1,056 BTC as digital asset reserves and reported $31.2 million in long-term debt. Operationally, its total mining hashrate reached 27.58 EH/s, including 19.84 EH/s self-operated and 7.74 EH/s hosted. Cango mined 656 BTC during the quarter, while the average cash cost per bitcoin fell about 5% from the prior quarter to $73,313. The company also said it has launched a hedging strategy. CEO Paul Yu said Cango has completed upgrades at its Georgia site, which can support up to 3 megawatts, and expects its planned bare-metal GPU hosting and colocation business to begin generating revenue in the third quarter.

Bitcoin mining company Cango Inc. (NYSE: CANG) released its unaudited financial results on Aug. 31 for the second quarter ended June 30, 2026, reporting total revenue of $50.8 million. Of that amount, $47.4 million came from bitcoin mining. The company posted a net loss of $81.6 million, which it said was mainly driven by non-cash impairment on mining machines and disposal losses. CFO Simon Tang separately said the net loss was largely tied to non-cash items. Cango said it held 1,056 BTC as digital asset reserves and carried $31.2 million in long-term debt. Its total operating hashrate reached 27.58 EH/s, including 19.84 EH/s from self-operated capacity and 7.74 EH/s from leased capacity. During the quarter, the company mined 656 BTC. Average cash cost per bitcoin fell about 5% from the prior quarter to $73,313. Cango also said it has started a hedging strategy. CEO Paul Yu said the company is focused on mining unit economics. He added that Cango has completed upgrades at its Georgia site, which can support up to 3 megawatts, and that containerized units have been installed while GPU hardware is arriving in batches. The company plans to launch bare-metal GPU hosting and hosting services, with revenue expected in the third quarter.

OpenAI has started offering a new pricing model to some large customers in recent months, according to The Information, shifting away from charging only for tokens, usage volume, or seats. Under the arrangement, customers pay after AI actually completes a task. One use case already in practice is customer service, where companies can be billed based on customer support tasks completed by AI rather than on standard consumption metrics. The company has not disclosed which customers are using the model, how much it charges, or how it defines a task as completed, and it declined to comment. The report says this is not a broadly available enterprise package. OpenAI’s publicly listed Enterprise pricing still mainly relies on quotas and token-based billing, suggesting the new structure is being negotiated privately with selected large clients. The move lines up with comments made earlier this year by OpenAI CFO Sarah Friar. In January, she said AI business models would move toward charging by outcomes. In July, OpenAI also argued that the more relevant measure is not simply the price per token, but the total cost required for AI to successfully finish a job. That framework now appears to be making its way into real commercial contracts.