See the full story in the original article.
Read the original on MarketWatch ↗
Keywords
METAartificial intelligenceAI models
Open a keyword to see other news on the same topic.
Related tickers and likely impact
META
▲ Likely positiveMeta Platforms
Relevance: medium
| Market | US stocks |
|---|---|
| Sector | Communication Services |
| Industry | Interactive Media & Services |
| Market cap | $1.9T |
| Close on 10/09/2026 | 718.67 |
| Change that session | -0.31% |
Impact is scored by AI from the headline and may be wrong. It is not investment advice.
Related news
- Meta finally pushes into frontier AI with its new Muse Wonder AI model
Meta has launched Muse Wonder, a frontier AI model designed for personal and professional AI agents, comprising the full Muse Wonderful model and a faster version. It will first roll out to some subscribers and eventually power Muse, Meta's AI agent. CEO Mark Zuckerberg said in a post on X that Meta has caught up with OpenAI and Anthropic, marking its return to the frontier after more than a year on the sidelines, following the scrapped Llama 4 Behemoth and a $14 billion acquihire of Scale AI co-founder Alexandr Wang, now Chief AI Officer. The discussion notes that these claims rest on benchmarks and that real-world developer feedback will be the true test.
- Multiverse Computing Launches Comet 60B in collaboration with NVIDIA: Multimodal Capabilities at Half the Parameters
Multiverse Computing launched Comet 60B, an open-source multimodal model built by compressing and optimizing NVIDIA's Nemotron 3.5 Super VL, which has 120 billion total parameters. Comet 60B halves total parameters to 60 billion while keeping 12 billion active parameters and support for text, image and video, so it fits on less hardware. The company says it stays within 5% of the original's accuracy and improves on some benchmarks, and it scores above the previous-generation Nemotron 3 Super on all eight benchmarks shown. The model was built with the CompactifAI framework through tensor-network profiling, compression, healing and GRPO-style reinforcement learning, and ships in BF16, FP8 and NVFP4 formats. Multiverse says it is aimed at developers building agentic AI workloads.
- Nvidia debuts three new open AI models
Nvidia (NVDA) is expanding its open AI model lineup with three new Nemotron models: Nemotron 3.5 Super VL, Nemotron 3.5 Bolt 74B-A9B 24GB, and Nemotron Bolt E4B 4GB, built to run agents from data centers to desktops. Super VL targets tasks lasting days or weeks and can be post-trained on enterprise data, Bolt 74B handles general agentic tasks like browsing and app use, and Bolt E4B is aimed at physical and conversational agents. The company argues open-weight models are cheaper and more private, since companies can run them on their own hardware, which could drive more chip sales over time. China currently leads the open-model space with models like Kimi K3 and Qwen, and CEO Jensen Huang has urged US firms to build more open models to boost competition.
- Tesla Roadster hype, the Meta-Apple AI deal, and how AI spending is saving the economy
A Market Hang panel discusses the excitement around the revamped Tesla (TSLA) Roadster event and a new AI partnership between Meta (META) and Apple (AAPL) to produce an open-source AI. The panelists also debate whether the economy would be on the verge of recession amid high inflation if not for AI spending. The panel includes Ross Gerber, Christine Short, Drew Pettit, and Shay Boloor.
- Could a Meta-Apple AI team-up restart the iPhone upgrade cycle?
Hosts said Apple's Home view announcement signaled that Apple is trying to expand into services because people are upgrading their phones less often, since three-year-old phones still work well. They argued that if Meta and Apple bring AI agents that run locally on the iPhone, this could start a new upgrade cycle, since it requires newer chips and more memory. One host said he had been buying Apple over recent weeks despite a valuation higher than he prefers, with the stock trading at about 40 times earnings, because he believes security will be a key factor in AI. They also expected models from OpenAI, Meta or Gemini to run natively on phones and laptops going forward.
- Apple and Meta team up to make Muse AI model run on Mac computers
Meta and Apple announced a collaboration that lets developers easily run Meta's open-weight model Muse Spark Open on Apple's high-end desktop Macs. According to Meta, Muse Spark Open is the most capable open-weight model released under a license allowing limited commercial use; it has 816 billion parameters and uses a mixture-of-experts architecture. Users can pair two to four Mac Studio computers to run the model locally rather than in the cloud, though a Mac Studio with the M5 Ultra chip and 256GB of memory costs at least $10,799, and a 512GB version is due later this month. The announcement, following months of collaboration, signals a shift in Silicon Valley alliances after years of tension, and may also show Apple's growing strength in AI chips against Nvidia.
Related markets
Latest news
- Cardano holders are being asked to bet $5 million on copper hitting a new record
Toto Finance is asking Cardano's treasury for 22 million ADA, about $5 million, to buy and tokenize copper, with voting closing Nov. 15. About two-thirds of the budget would buy copper stored in a COMEX-approved warehouse for 18 to 24 months before sale, while the rest covers storage, insurance, audits and Toto's work. If the full budget is spent and ADA's dollar price stays unchanged, copper would need to rise about 49.4% for the treasury to recover its investment, and there is no guarantee the treasury gets its original ADA back. Toto could receive up to 3.36 million ADA, roughly 15% of the request, and would keep its commercial rights to its tokenized-copper business. The proposal would use just over half of Cardano's remaining withdrawal allowance, which would not be replenished under existing rules.
- Bitcoin Core changes security policy as AI makes bugs easier to find
Bitcoin Core is changing its security disclosure policy because large language models have made some flaws easy to rediscover, so long embargoes offer little protection. Under the new policy, vulnerabilities deemed easily discoverable will be fixed as soon as possible across all maintained major versions, with security advisories published alongside the updates. Other flaws will follow severity-based schedules: low-severity bugs will be disclosed two weeks after the fix is released, while medium and high-severity bugs will be disclosed two weeks after the last affected release reaches End of Life, typically a year after the fix. The change follows a volunteer Bitcoin Red Team report of nearly 5,000 potential issues across 390 projects in August, though only about a fifth had been reproduced at that point.
- Asia FX steadies as oil surge weighs on risk appetite, yen holds gains
Asian currencies and the U.S. dollar were largely subdued on Friday as investors weighed rising energy costs and geopolitical tensions, while the yen held its gains after a safe-haven rally tied to concerns over global growth and tech valuations. The U.S. Dollar Index was flat at 101.60 by 03:33 GMT, near a three-week low and on track for a second straight weekly loss. USD/JPY edged up 0.1% to 156.90 after a nearly 1% drop in the previous session, as the yen reached its highest level in more than a month. Brent crude near $114 a barrel, after a 4% jump on Thursday following attacks in Saudi Arabia, raises inflation risks for energy importers such as Japan, India and South Korea. The Indian rupee extended its losses, with USD/INR up 0.3% to 95.109, while the yuan hovered near four-year lows; the Fed is expected to hold rates at 3.75%-4.00% on October 28, with ECB and BoJ decisions due on October 29 and 30.
- Securities firm with the heaviest loss in the sector in Q3 2026 revealed
APG Securities (APG) posted a pre-tax loss of nearly 108 billion VND in Q3 2026, compared with a profit of nearly 20 billion VND a year earlier, the largest loss among securities firms that have reported Q3 results. Revenue fell 21% to nearly 69 billion VND and gains from FVTPL financial assets dropped 93%, while operating expenses rose to nearly 124 billion VND, mainly due to a 119 billion VND loss on FVTPL assets. Financial expenses rose nearly 33% to 43.5 billion VND, mostly interest costs. For the first nine months, APG recorded a pre-tax loss of nearly 105 billion VND, against a profit of over 28 billion VND a year earlier, putting its 2026 pre-tax profit target of 20 billion VND under pressure.
- IonQ vs. Rigetti Computing: Which Quantum Computing Stock Is the Better Buy?
The article compares quantum computing companies IonQ (ion-based) and Rigetti Computing (superconducting). Rigetti reported Q2 revenue of $5.1 million, up from $1.8 million in 2025, but an operating loss of $28.1 million, and held $541.3 million in cash and investments; it also received a $100 million award from the U.S. Department of Commerce. IonQ posted Q2 revenue of $80.1 million and forecasts 2026 sales of $280–290 million versus $130 million in 2025, though its Q2 operating loss widened to $337.2 million from $160.6 million, and the Elisity deal involves stock valued at up to $1.3 billion. The author judges IonQ the better investment given its consistent revenue growth, says they bought IonQ shares, and notes Rigetti's price-to-sales ratio is more than six times IonQ's.
- Hoa Binh Construction dismisses a Deputy General Director
Hoa Binh Construction (HBC) dismissed Nguyen Kinh Luan from his position as Deputy General Director effective 1 November 2026, for personal reasons, and he will remain a Board member. Earlier, on 15 September 2026, Mr. Luan resigned from all his positions at the company and related entities. Separately, on 18 September 2026, the company distributed more than 51.4 million shares to 99 creditors to swap debt of nearly VND 514.2 billion, equal to 100% of the shares issued. The swapped shares are restricted from transfer for one year and are expected to be delivered within 30 days after the issuance closes.
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.