Meta's war mode, and its first AI model in a year:
Meta has spent the last year buying its way into the AI agent race, acquiring Manus for over $2 billion and investing $14.3 billion in Scale AI. Now it has launched a proprietary model called Muse Spark, closing the door on the open-source playbook that made Llama famous.
Breaking down the buzz:
The logic behind the acquisitions is easier to follow than the headlines suggest. Manus, a Singapore-based AI agent startup that reached $100 million in annualized revenue in just eight months, brought Meta a team capable of building agents that execute complex tasks including research, coding and data analysis. Scale AI, where Meta invested $14.3 billion in June 2025 in a deal that brought founder Alexandr Wang onto Meta's AI leadership team, runs the data labeling pipelines that make AI models smarter in the first place. Layering Manus's agent execution capability on top of Scale's training data infrastructure and Llama's model base produces something that looks, at least on paper, like a vertically integrated AI stack.
The latest product from that stack is Muse Spark, a proprietary closed model and a departure from the open-source identity that made Llama a developer favourite, driven partly by the fact that Chinese models from Alibaba and DeepSeek began outpacing Llama 4 on general knowledge and coding benchmarks by early 2026. Wang confirmed the team rebuilt Meta's AI stack from scratch over nine months and said future versions will be open-sourced.
Why this matters:
Meta's AI pivot is happening at serious cost, with capital expenditure reaching $70 to $72 billion in 2025 and total expenses projected to grow faster still in 2026, so that spend needs to translate into revenue. The Manus acquisition is its best attempt to make it do so. With 200 million WhatsApp Business users on the platform, bundling AI agent capabilities into a subscription product gives Meta a direct monetisation path beyond advertising, which is both cyclical and increasingly contested.
The bull case is that Muse Spark earns developer credibility and Manus agents embed into WhatsApp Business workflows at scale, giving Meta a fundamentally different revenue profile by year-end. The bear case is that $135 billion in projected capital expenditure produces a model that still trails OpenAI and Anthropic, and an agent product that loses its identity inside a 75,000-person organisation.
Please copy my portfolio for steady and consistent growth, $META (Meta Platforms Inc)... Show More
Not investment advice. The author may have financial interests in the mentioned instruments.
My friend I copy your trades I wish you success and I hope you are always in the green... Show More
.
We use our own and third-party cookies on our websites to enhance your experience, analyze traffic, and for security and marketing. For more info, see our Cookie Policy or go to Manage Settings.
Privacy Preference Center
When you visit any website, it may store or retrieve information on your browser, mostly in the form of cookies. This information might be about you, your preferences or your device and is mostly used to make the site work as you expect it to. The information does not usually directly identify you, but it can give you a more personalized web experience. Because we respect your right to privacy, you can choose not to allow some types of cookies. Click on the different category headings to find out more and change our default settings. However, blocking some types of cookies may impact your experience of the site and the services we are able to offer.
More information
Manage Consent Preferences
Strictly Necessary Cookies
Always Active
These cookies are necessary for the website to function and cannot be switched off in our systems. They are usually only set in response to actions made by you which amount to a request for services, such as setting your privacy preferences, logging in or filling in forms. You can set your browser to block or alert you about these cookies, but some parts of the site will not then work. These cookies do not store any personally identifiable information.
Performance Cookies
These cookies allow us to count visits and traffic sources so we can measure and improve the performance of our site. They help us to know which pages are the most and least popular and see how visitors move around the site. All information these cookies collect is aggregated and therefore anonymous. If you do not allow these cookies we will not know when you have visited our site, and will not be able to monitor its performance.
Functional Cookies
These cookies enable the website to provide enhanced functionality and personalisation. They may be set by us or by third party providers whose services we have added to our pages. If you do not allow these cookies then some or all of these services may not function properly.
Targeting Cookies
These cookies may be set through our site by our advertising partners. They may be used by those companies to build a profile of your interests and show you relevant adverts on other sites. They do not store directly personal information, but are based on uniquely identifying your browser and internet device. If you do not allow these cookies, you will experience less targeted advertising.
Cookie List
label
ConsentLeg.Interest
label
label
label
Tue May 12 2026 18:49:08 GMT+0000 (Coordinated Universal Time)
