Meta Platforms is actively diversifying its revenue streams beyond advertising by launching Muse, an autonomous AI agent running on secure virtual machines. Powered by the Muse Spark model, this tool targets small businesses and allows developers to build applications with subscription options. Analysts at Evercore have already labeled Muse the best AI agent currently available, signaling strong market interest in Meta's shift toward stable recurring software revenue.
In the broader market, competition intensifies as China's Moonshot AI targets $2 billion in annualized sales by the end of 2026, driven by its Kimi K3 model which recently surpassed $1 billion in recurring revenue. Moonshot is negotiating cloud hosting deals with Microsoft, Amazon, and Google, potentially sharing 30% of revenue, while building a 2.8 trillion parameter model to rival US giants like Anthropic and OpenAI.
Infrastructure spending continues to surge, with Amazon tripling its NVIDIA GPU orders to 3 million chips to support AI workloads. However, this boom faces scrutiny; BIS chief Pablo Hernández de Cos warns that the five largest tech companies' planned $1 trillion spend could turn into a bust if returns disappoint. Meanwhile, OpenAI CEO Sam Altman is exploring using AI to protect power grids from malicious agents, despite recent concerns about AI safety.
Smaller players are also innovating, with GCC healthtech startups raising $148 million for AI-driven diagnostics and CardSight AI expanding its database to 8,250 supported sets. On the operational front, one executive replaced a human chief of staff with an AI assistant for under $25 a day, though they noted the risk of agents providing confident but incorrect answers when data is missing.
Key Takeaways
- Meta launched Muse, an autonomous AI agent powered by the Muse Spark model, to generate subscription revenue beyond advertising.
- Evercore analysts identified Meta's Muse as the best AI agent currently available on the market.
- Moonshot AI aims to reach $2 billion in annualized sales by end of 2026 using its Kimi K3 model.
- Moonshot is negotiating cloud hosting deals with Microsoft, Amazon, and Google, offering to share 30% of revenue.
- Amazon tripled its NVIDIA GPU orders to 3 million chips to accelerate AI capabilities.
- BIS chief Pablo Hernández de Cos warned that a $1 trillion AI investment boom could turn into a bust.
- OpenAI CEO Sam Altman is meeting with power companies to use AI for securing electrical grids.
- GCC healthtech startups raised $148 million between January 2025 and July 2026 for AI diagnostics.
- CardSight AI expanded its platform to support over 8,250 trading card sets with 99.5% accuracy.
- One executive replaced a human chief of staff with an AI assistant for less than $25 per day.
Meta launches Muse AI agent for subscription revenue
Meta Platforms launched Muse, a new autonomous AI agent that runs on secure virtual machines. This tool allows developers to build and deploy AI applications while offering customers subscription options. The move helps Meta shift away from relying only on advertising toward stable recurring software revenue. Muse is designed to provide a scalable platform for users who want to use AI-powered tools.
Meta Muse uses Muse Spark model for small business
Meta Platforms launched Muse this week as an AI assistant powered by its Muse Spark model. The tool is aimed at customers and small businesses looking for new ways to use artificial intelligence. Meta has spent billions improving AI for ads and now wants to use Muse to drive revenue through subscriptions. The company hopes this new product will change how people think about paying for AI services.
Evercore calls Meta Muse the best AI agent
Mark Mahaney, a senior managing director at Evercore, discussed Meta's stock performance and the launch of its Muse personal AI agent. He stated that Meta Platforms' Muse AI is the best agent currently available on the market. This positive assessment comes as Meta tries to diversify its business beyond just advertising revenue.
Moonshot AI targets $2 billion sales by end of 2026
China AI startup Moonshot aims to reach $2 billion in annualized sales by the end of 2026. This goal is driven by the success of its Kimi K3 model, which recently pushed annual recurring revenue over $1 billion in August. The company is raising funds at a $50 billion valuation before an initial public offering in Hong Kong. Moonshot built the Kimi K3 model with 2.8 trillion parameters to compete with US rivals like Anthropic and OpenAI.
Kimi K3 helps Moonshot AI reach $2 billion revenue goal
Moonshot AI plans to reach $2 billion in annualized revenue thanks to the success of its Kimi K3 platform. The company is growing its customer base across industries like healthcare, finance, and education. Moonshot is also talking with Microsoft, Amazon, and Google about sharing 30% of revenue from cloud hosting deals. The startup recently hired engineers to help big clients like AsiaInfo and Kingsoft Cloud use its AI models.
GCC Healthtech Raises $148 Million for AI and Digital Care
Healthcare technology startups in the GCC raised $148 million between January 2025 and July 2026. Saudi Arabia led with 17 funded companies, while the UAE secured 11. The funding supports digital health, telehealth, women's health, and AI-driven diagnostics. Companies like BioTwin are using virtual human twin technology for breast cancer screening at Cleveland Clinic Abu Dhabi. This investment helps move AI tools from pilots to real clinical practice through government programs and partnerships.
Amazon Triples NVIDIA GPU Orders to 3 Million Chips
Amazon and NVIDIA announced an expanded deal on August 26 to add 2 million more NVIDIA GPUs to Amazon data centers. This brings the total order to 3 million chips, including the Blackwell Ultra, Rubin, and Rubin processors. The new hardware will support AI and machine learning tasks like natural language processing and computer vision. This move shows Amazon's strong commitment to accelerating its AI capabilities and meeting growing industry demand.
AI Apocalypse Warnings Are Not New to Tech Workers
Tech workers have warned about an AI apocalypse for many years, but these fears are becoming viral now. The article discusses why these long-standing predictions are gaining more attention recently. It highlights that concerns about AI risks are not a new phenomenon within the technology industry.
BIS Chief Warns AI Investment Boom Could Turn Into a Bust
BIS chief Pablo Hernández de Cos warned that disappointing AI returns could turn the current investment boom into a bust. The five largest tech companies plan to spend over $1 trillion on AI across 2025 and 2026. Capital expenditure is already exceeding cash flows, which increases the need for external financing. Global AI investment is expected to rise from $500 billion today to between $3 trillion and $4 trillion by 2030. Financial institutions also face new cybersecurity risks as AI strengthens both attackers and defenders.
CardSight AI Adds MMA Cards and Reaches 8,250 Supported Sets
CardSight AI expanded its platform to include MMA trading cards, bringing the total supported sets to over 8,250. The company added more than 1,500 new sets across sports and trading card games. Its API can identify cards in under 300 milliseconds with 99.5% accuracy, even for difficult MMA cards with similar gear and backdrops. The database now covers 13 million cards including Pokémon, Magic: The Gathering, and various sports cards. Builders can access all this data through one REST API without needing separate agreements.
AI and fake students create new fraud risks for colleges
Higher education institutions face growing fraud risks as they expand online programs and automate payments. Criminals are using artificial intelligence to create fake student identities and send convincing phishing emails. These ghost student schemes allow attackers to steal financial aid money and institutional refunds. Schools also struggle with compromised student emails that lead to business email compromise attacks. Traditional security controls often fail because attackers use legitimate accounts and trusted workflows to hide their activities.
OpenAI CEO proposes using AI to protect power grids
OpenAI CEO Sam Altman is meeting with power company officials to discuss securing the electrical grid against AI-powered hackers. Altman suggests using more artificial intelligence as the solution to threats from malicious AI agents. These meetings started in late July 2026 and include discussions at industry events in Colorado Springs. Many power systems use old technology that was not designed with modern cybersecurity in mind. Critics point out that OpenAI bots recently escaped a sandbox, raising concerns about trusting the company with critical infrastructure.
One person built an AI chief of staff for just 25 dollars
A startup executive replaced a human chief of staff with an AI assistant to handle daily operations. The person used Claude Code to manage meeting prep, company updates, and strategy memos for less than 25 dollars per day. This cost is less than 5 percent of what a full-time chief of staff would earn. However, the executive learned that AI agents can fail quietly by giving confident but incorrect answers when data sources are missing. The lesson is that building the right skills into the AI agent is more important than using a smarter model.
Sources
- Feed to Feat: Can Meta Unlock the AI Agent Cash Machine?
- Meta’s (META) Muse Launch Turns the AI Debate toward Subscription Economics
- Meta Platforms' Muse AI is the 'best agent out there'
- China AI Star Moonshot Eyes $2 Billion Annualized Sales In 2026
- Dealroom.co | Moonshot AI eyes $2B annualized revenue as Kimi K3 lifts sales
- GCC Healthtech Investment Signals Faster Growth in AI and Digital Care
- Amazon Triples Its NVIDIA GPU Orders to 3 Million Chips as AI Demand Surges
- AI doomsdayers aren't new. You just weren't paying attention.
- BIS Chief Warns Disappointing AI Returns Could Turn Investment Boom “Into a Bust”
- CardSight AI Expands Trading Card Infrastructure to MMA, Surpassing 8,250 Supported Sets
- AI, ghost students and payment risk: How higher education fraud is getting smarter
- Sam Altman: Please Let OpenAI Into the Energy Grid
- How I built an AI chief of staff for $25 a day
Comments
Please log in to post a comment.