Mistral Large 4, a 1-trillion-parameter model announced in public preview on October 6, had not published its weights or commercial license as of October 23, Day 22 of the 31-day window the company set for an end-of-month release. Hugging Face model cards still show 'Coming soon' for both fields. Industry reports suggest Mistral may move from the permissive Apache 2.0 license used for Mistral Large 3 to a custom framework, joining the 'Club of Four' of StepFun, Meta, Moonshot AI and DeepSeek, which recently set a $20 million monthly revenue threshold for commercial use. Under this model, a model can be downloaded but deployment limits apply once a licensee reaches significant scale, so a business with volatile revenue could exceed the threshold after one strong quarter. The article argues that this turns model selection into a procurement risk, and that the license terms will matter as much as the benchmarks.
Summary written by Upside AI from the original article. Read the source for full details.
Read the original on Yahoo Finance ↗
Keywords
artificial intelligenceDeepSeek
Open a keyword to see other news on the same topic.
Related tickers and likely impact
META
• NeutralMeta Platforms
Relevance: low
| Market | US stocks |
|---|---|
| Sector | Communication Services |
| Industry | Interactive Media & Services |
| Market cap | $1.9T |
| Close on 10/22/2026 | 648.91 |
| Change that session | -6.97% |
Impact is scored by AI from the headline and may be wrong. It is not investment advice.
Related news
- The AI gap: How human review is starting to fall behind
In The Loop host Ejaaz Ahamadeen discusses a growing gap in how people understand AI and LLM capabilities. He says around 6 billion people, roughly 75% of the population, have barely used LLMs, while 1 to 2 billion people (around 20%) are casual, infrequent users of free chatbot tiers who treat them as a better search engine. He cites a figure of about 5,000 people with access to internal, unreleased models and says their cognitive dissonance is huge. The host argues that AI intelligence is now accelerating beyond human comprehension, so people are falling behind in understanding the solutions AI produces, both from alignment and solution perspectives.
- DeepSeek Completes Nearly $15 Billion Funding Round, Largest Ever in China
Chinese AI startup DeepSeek is finalizing a funding round of about 100 billion yuan (US$14.98 billion) from existing and new investors, including Tencent and CATL, according to Reuters citing sources. The round values the company at 500 billion yuan (about $74.9 billion) before the new capital, and its size is double the initial target of 50 billion yuan. Other participants include Midea, Meituan, Geely and GigaDevice, with more than 15 investors in total. The terms and investor list are not final and the deal is expected to close within days, while DeepSeek, CATL and Tencent have not commented. The move follows DeepSeek's hiring of CITIC Securities in September to prepare a potential IPO on China's STAR Market.
- DeepSeek's Latest Update Scores Lower on Accounting Work as Open Models Trail on Consistency, DualEntry Benchmark Finds.
DualEntry's Accounting AI Benchmark (DualBench100) finds DeepSeek V4.1 Flash 1017 correctly completed 82.2% of tasks, down from 84.2% for the earlier V4.1 Flash, tying for tenth of 19 models. Open models are close on accuracy, with GLM-5.3 Flash leading them at 88.1%, while Gemini 4 Argon tops the chart at 91.1%. The bigger gap is consistency: Gemini 4 Argon passed 86.1% of tasks in all three runs, versus 76.2% for GLM-5.3 Flash. The benchmark covers 101 synthetic accounting cases, with each model run three times and an 80% pass mark. DualEntry founder Benedict Dohmen said reliability, not accuracy, is now the key issue for agents working in the general ledger.
- U.S. bans DeepSeek from government agency systems amid AI rift with China
The U.S. Department of Homeland Security made public an order requiring federal civilian agencies to remove DeepSeek-related technology from their IT systems and barring them from buying DeepSeek products and services, including third-party products built on its technology. The order was originally issued on July 24 and covers agencies such as the Treasury, State and Justice departments but excludes the Department of Defense. DHS warned that DeepSeek's models pose multiple threats, including possible transfer of sensitive government data to China and critical security flaws in its iOS app. Agencies may seek exceptions, including compliance delays, by submitting written justifications within 60 days of the order. The ban follows earlier restrictions, including Congress's limits in December 2025 and the Defense Information Systems Agency's block in January 2025.
- US bans DeepSeek from civilian agency IT systems
The US Department of Homeland Security has ordered the removal of DeepSeek products from federal civilian agency IT systems and barred agencies from new purchases, with the order also covering DeepSeek affiliates and third-party products built on DeepSeek. DHS warned that use of DeepSeek by employees and contractors could expose sensitive information to China, including details about US citizens and federal programs, and cited security flaws in DeepSeek's iOS app. The order took effect immediately upon publication and expands earlier bans at the Commerce Department and parts of the defense, security and intelligence community, amid US-China competition in AI. Australia had already banned DeepSeek products from government systems and devices in February 2025.
- US Orders Civilian Agencies to Rip DeepSeek Out of Government Systems
The U.S. Department of Homeland Security on Monday ordered civilian federal agencies to remove DeepSeek products and third-party services built on the model from their IT systems, and to stop buying them. The notice warns that using DeepSeek may transfer sensitive data to China and could produce intelligence on U.S. persons and agencies. The Commerce Department, the Navy, and the states of New York, Tennessee and Virginia had already restricted DeepSeek. The FCC votes on October 29 on barring Chinese labs from testing electronics for U.S. use.
Related markets
Latest news
- Bitcoin's volatility is well above euro-dollar gyrations, though it has calmed to rival Nvidia's
CoinDesk compares the 90-day realized volatility of bitcoin with other assets, a measure of how much prices moved, based on the standard deviation of daily returns and annualized. As of Thursday, bitcoin stood at about 38%, versus about 23.5% for gold, about 11% for the S&P 500, and about 37% for Nvidia, while the EUR/USD pair was the steadiest at 4.6%. Bitcoin's 90-day volatility has averaged about 58% since January 2020, peaking at 134% in June 2020, and its calmest reading of 27.5% came in September 2025 when prices rose above $100,000. The article concludes that bitcoin has settled down and now moves like a big tech stock, but it has yet to rival a national currency, since most fiat currency pairs have realized volatility below 10%.
- Honeywell (HON) Unveils Forge Cognition To Bring AI Into Facility Operations
Honeywell International has launched Forge Cognition, an AI and simulation platform for facility operations, developed with NVIDIA and Dell Technologies. The platform applies AI models and physics-based simulation to asset reliability, energy use and day-to-day plant performance. The article says it extends Honeywell's Forge software stack across more than 300,000 installed sites where the company already manages controls, safety and building systems. NVIDIA contributes AI models such as Nemotron, Omniverse libraries and GPUs, while Dell brings its AI Factory stack for scaling from the data center to the edge. The launch fits Honeywell's push for a higher mix of software and services to support its long-term margin and growth goals.
- 3 Investment Grade Bond Ideas Riding Big Tech AI Capex Plans
Big Tech capex plans for AI and chips have unsettled credit markets, and this article highlights three US investment-grade issuers tied to them. Lattice Semiconductor, with Core Lattice sales of about US$651 million and a market value near US$15.3b, supplies low-power FPGAs used in data centers and networking gear. Datadog generated about US$4.0b from IT infrastructure products, has a market value near US$92.6b, and counts over 750 AI customers. Vertiv Holdings Co, with an equity value near US$81.7b, supplies power and cooling hardware for AI data centers, reported Q4 2025 organic orders up 252% year over year, and holds a $15 billion backlog equal to roughly 1.5 years of trailing revenue. The article notes that its wider screen surfaced 30 more US investment-grade issuers with similar capex-driven stories.
- Nebius Has Cratered Over The Last 3 Months: This Wall Street Ratings Firm Says 2.5X Returns Will Materialize Soon
A Gartner forecast that over 80% of neocloud providers will cease independent operations by 2030 triggered a sharp sell-off in Nebius Group (NBIS), which now trades at $159.17, about 47% below its 52-week high of $299.86. The consensus target from 21 analysts is $283.60, while Northland Securities holds the Street-high target of $410.00. Last quarter revenue reached $582.3M, beating the $574.65M consensus, though the operating loss was $175.9M and three customers made up 24%, 21% and 14% of revenue. Management reaffirmed full-year revenue guidance of $3 billion to $3.4 billion and roughly 40% adjusted EBITDA margin, and CEO Arkady Volozh said Nebius could sell its entire 2027 capacity on current terms. The article flags dilution risks from $20.8B of convertible debt, $12.1B in upcoming lease obligations and ongoing share sales.
- Microsoft Has a $1 Trillion Opportunity Hiding in Plain Sight
24/7 Wall St. sets a 12-month price target of $647.87 for Microsoft, implying 21.8% upside from $531.90, and rates the stock a buy. The thesis centers on contracted demand, with commercial remaining performance obligations up 84% to $678 billion and Azure exceeding $100 billion in annual revenue after growing 43%. Microsoft 365 Copilot has passed 30 million paid seats, and the bull scenario reaches $752.32. The main risk flagged is capex, which rose 79.62% to $115.95 billion in fiscal 2026 while free cash flow fell 6.46%, with a worst-case scenario at $547.61. The article notes Microsoft expects to stay free-cash-flow positive in fiscal 2027, unlike Alphabet (-$5.86 billion) and Oracle (-$5.40 billion).
- The 10-Year Yield Just Fell Below 5% as Investors Fled AI Stocks
The 10-year Treasury yield fell 19.3 basis points to 4.953% on October 22, 2026, as investors fled AI stocks for safer assets. Meta Platforms fell 7% after its results, the Philadelphia semiconductor index dropped 6.7%, and the Nasdaq Composite lost 2.6%. CME FedWatch odds of a December rate hike eased to about 78%, and Freddie Mac's 30-year mortgage average fell to 7.32%, though mortgage rates remain above 7%. The article says this fear-driven bond rally could reverse quickly, with the FOMC meeting on October 27 and 28 and PCE inflation data on October 29 as key tests.
Get the briefing on Telegram or Zalo
Join the Upside Telegram channel or Zalo group for the Vietnam stocks and gold briefing at 8:00 and 19:00 (Vietnam time) on trading days, plus a weekly wrap on Saturday morning.