cost

MiniMax announced that M3.1 Flash has officially been released, and the earlier version, Space Bunny, which topped global rankings and processed over 200T tokens, is its predecessor.M3.1 Flash achieved scores of 62% on Automation-Bench, 30% on ALE, 74% on DeepSWE, and 91% on BabyVision in tasks such as Cowork, Coding, and Visual Reasoning.

The cryptocurrency trading platform Bitfinex stated on X that even though the ETF is in a sell state, new BTC buyers have still accumulated about 570,000 BTC since the end of July.Bitfinex indicated that the average cost for these buyers is approximately $75,000, and this cost has been maintained during each pullback since August.

Ethereum co-founder Vitalik Buterin shared that the cost of Bitcoin mining is roughly half derived from capital expenditures on computing power equipment and the other half from electricity consumption; in contrast, the cost of AI inference is almost entirely dominated by capital expenditures on computing power. This difference reflects the distinct cost structures of the two computing businesses. Bitcoin miners need to continuously pay for electricity and determine their profit margins based on mining machine efficiency, coin prices, and network difficulty changes; AI inference, on the other hand, relies more on investments in expensive hardware like GPUs, as well as equipment utilization and computing efficiency.The industrial boundaries of both are also changing. As the demand for AI computing power grows, some Bitcoin mining companies are beginning to shift their existing electricity resources and data center facilities towards AI and high-performance computing businesses, hoping to transition from pure mining to computing infrastructure services. Some mining companies are leveraging previously allocated electricity capacity for mining to enter the AI data center market. However, transformation does not mean that cost pressures disappear: mining operations are more directly exposed to fluctuations in electricity prices and Bitcoin prices, while AI computing businesses must bear risks related to GPU procurement, depreciation, and underutilization. Both types of businesses compete for electricity and computing resources, but ultimately, profitability depends on how these resources are converted into revenue.

The third-quarter report released by the stablecoin payment network Borderless.xyz shows that the median cost of sending a $200 remittance through stablecoin funding channels is 1%, or $2; whereas, according to World Bank data, the median cost of traditional remittance channels is 6.36%, or $12.72. In comparison, the cost of stablecoin remittances is approximately 84% lower. This comparison uses Borderless's pricing for the third quarter of 2026 and World Bank data for the same period in 2025.In emerging markets, the cost advantage is also evident. The cost of stablecoin remittances in Sub-Saharan Africa is 2.9%, while traditional transfers are 8.46%; in Latin America, the figures are 1.25% and 5.64%, respectively. In terms of corporate payments, a $10,000 stablecoin transfer to the Eurozone can cost as low as $26 when choosing the lowest-cost service provider, with the median service provider costing $52; traditional international bank wire transfers for the same amount require $135 to $150, including transfer fees and a 1% foreign exchange spread.Regarding settlement speed, the report states that stablecoin payments can reach the recipient's local bank account within minutes. The median settlement time for Brazil's PIX network in the third quarter is 2 minutes, Mexico's SPEI is 3 minutes, and Colombia's PSE is 2 minutes; traditional SWIFT transfers typically settle the next day, with the slowest 10% taking 3 business days. The success rate for receiving funds varies significantly: Mexico's SPEI is 98.6%, Brazil's PIX is 97.8%, Colombia's PSE is 79%, and Costa Rica's SINPE is only 73%. Standard Chartered Bank recently pointed out that although the total market capitalization of stablecoins has grown relatively slowly this year, transaction volumes are still rising, and it is expected that this market will reach $2 trillion by 2030, two years later than previously predicted.

CryptoQuant analyst MorenoDV_ published an analysis indicating that the realized market cap of new whales in Bitcoin is $78.61 billion, while old whales stand at $159.15 billion. New whales account for 33.06%, compared to about 18% earlier in 2026.Since September, the realized market cap of new whales has continued to grow, while old whales have remained relatively stable, with the growth primarily coming from new whales. The realized market cap is calculated based on the price at the last on-chain transfer of the tokens and does not represent the proportion of BTC supply.In terms of cost, the overall cost basis of new whales has decreased from a previous high of around $110,000 to about $68,000 at the end of August, before rebounding to around $77,000. The current BTC price is $82,134, which is approximately 6.7% higher than the cost basis of new whales.The analysis suggests that the data shows large holders buying during the downturn and continuing to accumulate as prices rebound. However, the turnover of the holder group and transfers between addresses can also inflate the realized market cap, so the growth in realized market cap cannot be interpreted as an equivalent influx of new funds. If BTC continues to fall below around $77,000, new whales will incur losses again, increasing the likelihood of selling during a rebound. The inflow to exchanges and realized losses can be used to determine whether this is short-term pressure or if large holders are starting to sell off again.

On October 16, Circle announced that in the first half of 2026, the on-chain settlement amount of USDC exceeded $36 trillion, and Circle minted and redeemed over $315 billion of USDC during the same period, averaging about $1.7 billion per day. The company stated that as of June, USDC accounted for approximately 77% of the global stablecoin trading volume, an increase from 60% a year ago.More attention from the market is focused on the income distribution of USDC. Circle disclosed that the company generated approximately $1.4 billion in revenue in the first half of the year, of which about $819 million was allocated to USDC distribution partners and transaction costs, accounting for about 59% of the revenue; another approximately $473 million was invested in licensing, banking partnerships, settlement liquidity, compliance systems, product development, and multi-chain deployment, accounting for about 34%. According to the announcement, for every $1 of revenue Circle earns, it ultimately retains only about 7 cents in net profit.This announcement may be in response to market concerns regarding "the monopolization of reserve income by stablecoin issuers." Circle stated that partners can share corresponding reserve income by holding and expanding their USDC balances on their platforms; currently, more than 150 trading platforms, wallets, banks, fintech companies, payment platforms, and enterprises have formed economic interests tied to the growth of USDC. Circle also mentioned that Binance signed a five-year commercial agreement with it in September and invested $100 million in Circle.On the product side, Circle announced the expansion of its free redemption mechanism: all Circle customers can still redeem USDC for free through traditional banking channels, with the funds typically arriving within two business days; customers who make payments through the Circle Payments Network or use eligible products such as Arc, Circle Mint, and Circle Wallets will be able to enjoy free, near-real-time redemption capabilities.

NVIDIA CEO Jensen Huang stated that COREWEAVE's strength is increasing every day, and its capital costs are actually decreasing.

Stripe announced a collaboration with Plasma to develop a multi-currency stablecoin card, aimed at allowing individuals and businesses to spend the currencies they already hold in their accounts, reducing unnecessary currency conversion and foreign exchange costs. Plasma will serve as the design partner, integrating with Stripe's stablecoin card infrastructure, enabling customer account balances to be used for payments through the multi-currency card.The expanded infrastructure will support US dollars, Brazilian reais, euros, Mexican pesos, and Colombian pesos; stablecoins include USDC, OUSD, and qualified stablecoins that meet the requirements of the GENIUS Act. The Stripe stablecoin card service currently covers over 30 countries, with plans to expand to over 100 countries by 2027.For example, users holding euros will be able to spend directly in euros in the future, without needing to convert to US dollars first; businesses can issue currency cards that match their spending needs. Stripe also plans to support cardholders in spending from multiple account balances and automatically selecting the appropriate currency at the time of purchase.

The Japanese Bitcoin financial company Metaplanet announced that the company's board of directors has decided to implement a share buyback through the Tokyo Stock Exchange ToSTNeT-3 off-market trading mechanism. According to the announcement, Metaplanet plans to use the closing price of 261 yen per share on October 13, 2026, as the buyback price, and will execute buy orders through ToSTNeT-3 at 8:45 AM on October 14. The number of common shares intended for repurchase is 96,934,800 shares, accounting for approximately 7.1% of the total issued common shares after deducting treasury shares as of September 30, 2026, with an expected total buyback amount of approximately 25.2999828 billion yen.The company stated that the number of shares to be repurchased will not be adjusted, and the final transaction results will be announced after the trading ends at 8:45 AM on October 14. In addition, Metaplanet had previously approved a larger-scale share buyback plan, intending to repurchase up to 200 million shares through market trading on the Tokyo Stock Exchange from October 14, 2026, to October 13, 2027, accounting for approximately 14.7% of the issued shares, with a total buyback amount cap of 60 billion yen.

The artificial intelligence company Zhipu released a legal-specific model GLM-Legal, which is further trained based on GLM-5.3-Flash. In the Harvey LAB evaluation covering 24 categories of legal services, the task pass rate increased from 6.7% to 19.2%, nearly tripling the original version, surpassing models like Kimi K3 and Claude Fable 5, and ranking seventh on the Vals AI leaderboard.When reviewing software service contracts, GLM-Legal suggested shortening the termination notice period from 60 days to 30 days and required service providers to offer up to 90 days of handover support. In another software project dispute involving 40 documents, it found that the company reported to the client that the project was 60% complete, while internal records indicated only about 41%. It also identified emails from the CEO requesting to conceal risks and continue charging, and analyzed potential legal liabilities in conjunction with contract terms.According to independent evaluations by Vals AI, the cost for GLM-Legal to complete a test is $0.15, while the closely performing Gemini 4 Argon costs $10.55, a cost difference of about 70 times. GLM-Legal took an average of 72 minutes, while the latter took 44 minutes. Zhipu stated that the model was trained specifically on legal data and conducted reinforcement learning in tens of thousands of simulated legal environments. Currently, GLM-Legal requires an API trial application.

xAI launched the X subscription plan, integrating Grok, Grok Bot, Cursor, and X Premium benefits globally, with subscription fees ranging from 8 to 200 USD per month.

Ministers from OPEC and non-OPEC oil-producing countries met in Riyadh, Saudi Arabia, reaffirming their commitment to maintaining stability in the oil market through a joint mechanism. OPEC+ stated that disruptions in energy transportation have caused shipping and insurance costs to exceed one-third of the value of commodities, hindering the transportation of crude oil and refined products, and increasing inflationary pressures on the global economy. Countries will continue to monitor the situation and cooperate with international organizations to improve shipping conditions.

According to the Miner Weekly briefing published by BlocksBridge Consulting, Bitcoin mining companies expanding into the fields of artificial intelligence and high-performance computing are facing significantly different borrowing costs, with lenders increasingly scrutinizing large data center projects. The briefing points out that securing large AI clients is only part of the challenge; mining companies also need to raise billions of dollars to purchase computing power equipment, build data centers, and ensure the power infrastructure required for contract fulfillment. BlocksBridge indicates that large lending institutions have become more selective about data center loans, with decisions increasingly depending on client qualifications, contract terms, and project completion paths.The transition from Bitcoin mining to AI infrastructure by IREN reflects the differences in financing costs: the company secured $3.65 billion in GPU financing in June at a blended interest rate of 6%; in August, it completed another $2.4 billion financing with a fixed interest rate of 9%. The two agreements correspond to different projects and are not directly comparable, but they show that even within the same company, borrowing costs can vary significantly. Financing arrangements may also incur costs beyond interest; in another transaction involving Cipher Digital, Google guaranteed a $1.4 billion lease obligation in exchange for warrants representing a 5.4% equity stake in Cipher, with exercising the warrants potentially diluting existing shareholder equity.The rise in financing demand highlights the overall challenges mining companies face in entering the AI market. Data from VanEck shows that during the peak of the Bitcoin bull market in October last year, mining company debt surged by approximately 500% compared to the second quarter of 2024.

According to a report by Nomura Securities, AI company Kimi under the brand Moonlight's Dark Side is estimated to achieve an annual recurring revenue of $1.2 billion in September. Bloomberg previously reported that Kimi's annual recurring revenue in August had surpassed $1 billion, with a year-end target of $2 billion.The K2.8 Preview, launched on September 11, is estimated to contribute about 40%--45% of the annual recurring revenue, exceeding the flagship model K3's 30%--35%. The K2.8 Preview is priced lower and primarily targets programming and everyday agent tasks. Experts estimate that K3 has a gross margin of about 35%--40%, while lower-priced models like K2.8 Preview have about 20%--25%. The API is Kimi's largest source of revenue, with overseas markets contributing more than half.

Fitch Ratings pointed out that global borrowing costs are rising, and bond yields are climbing.