February 13, according to information on the relevant page, Bitcoin Improvement Proposal 360 (BIP-360) has been added to the official codebase. The proposal introduces the Pay-to-Merkle-Root (P2MR) output type, aimed at enhancing Bitcoin's defense against quantum computing threats through a soft fork.

February 13th, according to Coinbob Popular Address Monitoring, on-chain data shows that the largest ETH long whale (0xa5b…) has reduced its position twice recently: on February 9th at $2035 and around $1960 on February 11th, selling a total of 7000 ETH (about $14 million) at a loss of $596,000. In the past 6 hours, the whale has gradually re-entered the market with an average price of $1919, engaging in swing trading.The whale currently holds a 15x leveraged ETH long position, with a position size of 60,000 ETH, equivalent to about $116 million. The average entry price has decreased from $2059 to $2048, resulting in an unrealized loss of $6.27 million (-80%). The liquidation price is $1328, with $27 million in remaining collateral in the account.Previously, on February 9th, the whale closed out its ETH long position held for about two months, resulting in a loss of approximately $9.63 million, before re-entering a similar position size of around 60,000 ETH.

February 13th, the U.S. will release the January Non-Farm Payrolls at 21:30 tonight (UTC+8), with a previous value of 2.70% and an expected value of 2.50%. It will also announce the January CPI M/M, with a previous value of 0.3% and an expected value of 0.3%.Goldman Sachs analyst Kay Haigh stated yesterday that they still believe the Fed has room for two more rate cuts this year. However, if Friday's CPI release unexpectedly shows an increase, it may tilt the Fed towards a more hawkish stance.According to PolyBeats monitoring, the probability on Polymarket for the January Non-Farm Payrolls to be less than or equal to 2.8% is 96%, with only a 2% probability for 2.9% and 3.0%.

February 13th, according to HTX market data, Ethereum briefly entered the $1800 range this morning, and in the future may usher in a "lost decade" due to yields lower than the cash rate.As early as January 2018, the price of Ethereum had touched above $1400. Adjusted for U.S. CPI inflation compounding, that same $1400 in February 2026 would be equivalent to about $1806. If investors do not stake Ethereum or hold it naked, they may have underperformed the U.S. dollar cash over the past 8 years.

February 13th, According to Coinbase's 8-K report submitted to the U.S. Securities and Exchange Commission (SEC), in the fourth quarter of last year, Coinbase increased its Bitcoin holdings in its proprietary cryptocurrency investment portfolio by $39 million through weekly purchases. As of December 31, 2025, the fair market values of cryptocurrency assets held for proprietary investments and cryptocurrency assets held as collateral were $2 billion and $823 million, respectively.Coinbase's corporate Bitcoin holdings are currently about 14,548 BTC this month, valued at around $960 million.

February 13th, according to the Ark Invest Tracker, Cathie Wood's Ark Invest increased its holdings yesterday as follows: 174,677 shares of Robinhood stock (approximately $12.42 million); 212,314 shares of BitMine stock (approximately $4.19 million); 74,323 shares of Bullish stock (approximately $2.36 million).

February 13th, according to Bitget market data, Hong Kong stock Smart Vision (HK2513) saw a rapid rise of over 14.5% at the opening, with the highest stock price reaching 462 Hong Kong dollars, hitting a new high since its IPO, with a cumulative increase of over 120% this week.Yesterday, Smart Vision's AI programming subscription package, GLM Coding Plan, went online and sold out immediately, a rare occurrence in the Chinese domestic AI programming model paid subscription package in the industry. The GLM-5 programming and Agent capabilities are approaching the level of top international closed-source models, while being open-source and dominating in terms of pricing. Previously, it has been sold out multiple times due to a surge in users, leading to concurrent tightness in computing power.The price of Chinese domestic large-model tokens is only equivalent to a fraction of the U.S. price. On OpenRouter's global large-model API aggregation platform, OpenClaw, the recent Kimi K2.5 quickly rose to the top in just one week, overshadowing large-model giants such as Gemini and Claude. Domestic large models are using low prices as a catalyst, coupled with high performance to drive user retention, bringing a double blow of price and performance to the AI market.

February 13th, according to on-chain analyst Ai Auntie (@ai_9684xtpa) monitoring, accompanying ETH breaking below $2000 again, the "cumulative long 105,000 ETH" whale once again increased its position by 3,000 ETH in a long order, with an average holding price of $1919. This whale holds a total of 105,000 ETH across two addresses, worth $204 million, with an unrealized loss of $10.229 million.Address 0xa5B…01D41: Holding 60,000 ETH ($116 million), entry price $2048.9, unrealized loss $6.36 million;Address 0x6C8…D84F6: Holding 45,000 ETH ($87.34 million), entry price $2029.38, unrealized loss $3.869 million.

February 13: A JPMorgan analyst stated that, based on a volatility-adjusted comparison with gold, the long-term price target for Bitcoin is $266,000. The U.S. winter storm and the exodus of high-cost miners from the market are the main reasons for the drop in hash rate, but signs of a hash rate rebound are now being observed. Despite short-term fluctuations, the JPMorgan team remains optimistic about the crypto market in 2026, expecting institutional investors to drive market growth.

February 13th, Today OpenAI released the GPT-5.3-Codex-Spark research preview, a slimmed-down version of GPT-5.3-Codex and OpenAI's first model specifically designed for real-time coding. It supports targeted editing, logic reshaping, or interface optimization, with immediate feedback. Codex-Spark is optimized to run on ultra-low-latency hardware, achieving nearly instant responses—delivering over 1000 tokens per second while still handling real coding tasks effectively.