Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Meta’s superintelligence effort began with a reported $14.3 billion investment for a 49% stake in Scale AI and the recruitment of Scale founder Alexandr Wang. Wang became a senior leader in Meta’s new AI organization, later named Meta Superintelligence Labs. By April 2026, the lab had announced its first model, Muse Spark. That is evidence of a working product effort—not proof that Meta has achieved artificial superintelligence.
What Meta’s Scale AI deal involved
In June 2025, Meta agreed to invest approximately $14.3 billion for a reported 49% stake in Scale AI. It was a major investment, but not a full acquisition: Scale remained a separate company. At the same time, founder and CEO Alexandr Wang left the CEO role to take a senior position at Meta and lead its new superintelligence effort. Scale promoted strategy chief Jason Droege to CEO. AP reported the size and stake; CNBC reported Wang’s move and Droege’s appointment.
Reports described Meta’s stake as non-voting or otherwise structured to avoid ordinary corporate control, though accounts differed on the precise governance arrangements. The safest description is a large minority investment, not that Meta bought Scale or took full control of it.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Repair Windows errors before they cause bigger problemsFix Now →The transaction linked three goals: investing in an AI-data company, recruiting Wang and potentially other talent, and improving Meta’s own AI execution. Those aims overlap, but they are not interchangeable. An investment in Scale did not itself establish that Meta received Scale customers’ confidential data.
#1 Best Overall
- CARDBOARD MONKENAUT — Get our best Gorilla Tag bundle yet with this Amazon exclusive deal. Purchase Meta Quest 3S to get exclusive items, including the Gorilla Space Program Suit and Helmet, plus 2,000 SHINY ROCKS.
- NO WIRES, MORE FUN — Break free from cords. Game, play and explore immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Take gaming to a new level and blend virtual objects with your physical space to experience two worlds at once in your VR headset.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up. *Based on the graphic performance of the Qualcomm Snapdragon XR2 Gen 2 platform vs the Meta Quest 2 platform.
Why Meta made the bet
Meta had spent heavily on AI models and infrastructure, including its Llama family, but the reception of Llama 4 was widely reported as weaker than the company hoped. Zuckerberg was reported to be dissatisfied with Meta’s competitive position and personally involved in recruiting prominent AI talent. That account comes from reporting based on people familiar with the matter, not a formal company admission. NBC’s report on the deal placed it in the context of Meta’s push to catch rivals including OpenAI, Google, Anthropic and xAI.
Meta’s potential advantage was broader than a research team alone. It had large-scale computing infrastructure and could distribute AI features through Facebook, Instagram, WhatsApp, Messenger, its Meta AI app and wearable devices. Its compute overview describes the infrastructure behind that effort. Distribution can put a model in front of many users, although reach is not the same as model quality or user trust.
Why Alexandr Wang?
Wang was not recruited simply as a conventional frontier-model scientist. He built Scale AI into a significant provider of data preparation, annotation, evaluation and related services for AI developers. These services matter because model development depends on more than algorithms and computing power: teams also need data that is useful for training and methods for evaluating model behavior.
Wang brought experience in building a company, recruiting, working with enterprise customers and executing at scale. Reuters characterized the appointment as Zuckerberg betting on Wang as an execution-focused leader. Reuters coverage provides that framing. The distinction matters: running a fast-growing AI-data business is relevant experience, but it is not the same as leading fundamental research into systems that exceed human capabilities.
Rank #2
- CARDBOARD MONKENAUT — Get our best Gorilla Tag bundle yet with this Amazon exclusive deal. Purchase Meta Quest 3 to get exclusive items, including the Gorilla Space Program Suit and Helmet, plus 2,000 SHINY ROCKS.
- NEARLY 30% LEAP IN RESOLUTION — Experience every thrill in breathtaking detail with sharp graphics and stunning 4K+ Infinite Display.
- NO WIRES, MORE FUN — Break free from cords. Game, play and explore in immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Blend virtual objects with your physical space and experience two worlds at once in your VR headset.
What Meta Superintelligence Labs is
Meta Superintelligence Labs is best understood as a company-wide AI organization, not one small, self-contained research lab. Early reporting associated Wang with the overall effort and former GitHub CEO Nat Friedman with products and applied research. Meta’s investor materials also described Wang and Friedman’s roles. Later reporting said Meta named Shengjia Zhao chief scientist in July 2025. Reuters reporting carried by CNA covered the early leadership structure, while Reuters coverage of Zhao’s appointment documented another leadership addition.
The organization’s work spans frontier models, applied research and products, infrastructure, and longer-term research. It recruited researchers from competing AI companies and operated alongside Meta’s existing FAIR and Llama groups. Meta’s AI organization changed during 2025 and 2026, so reported internal teams and responsibilities should be treated as a snapshot rather than a permanent org chart.
The reorganization also had a human and operational cost. In October 2025, AP reported that Meta cut about 600 employees in its AI organization while continuing to hire for the superintelligence group. That points to restructuring as well as expansion: building a new elite effort can create friction with teams already doing AI work. AP’s report on the cuts covered that contrast.
Free tools Windows power users keep installed
One-click scans. No signup required.
What “superintelligence” means—and does not mean
In technical and policy discussions, artificial superintelligence usually means a hypothetical AI system substantially better than humans across most or all important cognitive tasks. Meta’s phrase personal superintelligence is a product vision: a personalized assistant that understands context and helps users, including by taking actions. That vision is different from demonstrating broad intellectual superiority.
Rank #3
- NO WIRES, MORE FUN — Break free from cords. Game, play, exercise and explore immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the SnapdragonTM XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Take gaming to a new level and blend virtual objects with your physical space to experience two worlds at once.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up.
- 33% MORE MEMORY — Elevate your play with 8GB of RAM. Upgraded memory delivers a next-level experience fueled by sharper graphics and more responsive performance.
Agentic AI is a more immediate product category: systems that can plan, use tools, access applications and attempt multistep tasks. A model that can carry out some such tasks may be useful without being reliable at every step, let alone superhuman across domains. Meta’s use of “superintelligence” describes an ambition and organizational strategy; it should not be read as a scientific finding that the company has achieved it.
Muse Spark: the lab’s first announced model
On April 8, 2026, Meta announced Muse Spark as the first model from Meta Superintelligence Labs and the start of a new Muse model series. Meta described it as relatively small and fast, with multimodal interaction and reasoning capabilities in areas including science, mathematics and health. These are company descriptions, not independent conclusions about how well the model performs against competitors. Meta’s launch announcement sets out its claims and framing.
Meta later said Muse Spark 1.1 could plan and take actions through connected applications, including connecting to email and calendars, creating slides, conducting research and handling tasks for a user. These, too, are product claims; performance and availability can vary. Meta’s July announcement describes the agent-like functions.
Recommended Free Tools
Meta said Muse Spark was being integrated into the Meta AI app and website, WhatsApp, Instagram, Facebook, Messenger and AI glasses. Rollout varied by product, location, account and date, so an announced capability should not be assumed available to every user. Meta’s glasses announcement describes its wearable direction.
Rank #4
- NEARLY 30% LEAP IN RESOLUTION — Experience every thrill in breathtaking detail with sharp graphics and stunning 4K Infinite Display.
- NO WIRES, MORE FUN — Break free from cords. Play, explore and exercise in immersive worlds — untethered and without limits.
- 2X GRAPHICAL PROCESSING POWER — Enjoy lightning-fast load times and next-gen graphics for smooth gaming powered by the Snapdragon XR2 Gen 2 processor.
- EXPERIENCE VIRTUAL REALITY — Blend virtual objects with your physical space and experience two worlds at once.
- 2+ HOURS OF BATTERY LIFE — Charge less, play longer and stay in the action with an improved battery that keeps up.
The launch establishes that the new organization produced a public model and was connecting it to consumer products. It does not, by itself, show that Muse Spark leads competitors on independent benchmarks, works reliably for every task, or represents a complete replacement for Llama. Meta’s broader model portfolio and the relationship between Muse and Llama remain evolving.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.The unresolved Scale AI questions
Scale’s business made it strategically useful to Meta, but the combination of a near-majority stake and Wang’s move raised questions about independence and conflicts. Scale served multiple AI developers; customers and competitors could reasonably ask whether Meta’s investment gave it preferential access to expertise or services, or affected Scale’s ability to serve rivals on neutral terms.
Those questions are not proof of improper data access or a legal violation. Public-interest groups urged the Federal Trade Commission to investigate the arrangement as a possible “de facto vertical acquisition.” That was an advocacy request, not a finding by the FTC. The groups’ letter to the FTC sets out their concerns. The distinction between a minority investment and a full acquisition matters, but it does not settle every question about influence, customer confidentiality or competition.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Wang was reported to remain connected to Scale through its board or another continuing relationship; details of the governance and conflict arrangements should be attributed to the reporting available rather than treated as settled. Droege’s appointment as CEO was one visible step in keeping Scale operating separately.
Can the strategy work?
Meta’s case for the strategy: Wang could help recruit and organize talent; Scale’s data and evaluation experience is relevant to model development; Meta can invest in compute; and its consumer platforms offer unusually broad distribution. Meta’s relatively open-model history may also give it a different route to adoption from rivals that emphasize closed services.
The risks: Wang’s operational strengths do not guarantee research breakthroughs. A new, high-profile group can compete for resources and status with existing teams. Very large recruiting packages—reported for some recruits, but not an audited account of Meta’s overall compensation—can raise costs without guaranteeing better models. Model development also requires substantial infrastructure spending; that is separate from the $14.3 billion Scale investment and from individual recruitment packages, and the figures should not be casually added together.
Agentic products introduce practical risks beyond benchmark scores. A system may hallucinate, misunderstand a request, or take an unsafe action through a connected app. Giving it access to email, calendars, messages, photos or social activity raises questions about permissions, privacy and accountability. Reliability can vary by language, region, device and product surface, and product versions may use different models, routing and safety layers. Users should review what an assistant can access and confirm consequential actions rather than assuming a polished demonstration guarantees dependable behavior.
The Tool Desk
Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →What the bet has—and has not—proved
Meta’s 2025 move was more than a cheque to an AI supplier: it tied a major minority investment to a leadership hire and a broad organizational reset. By August 2026, Meta Superintelligence Labs was an operating effort with the Muse model family and integrations into Meta products, rather than just a proposed lab. But shipped products and ambitious branding are milestones, not proof of artificial superintelligence. The harder test is whether Meta can turn its talent, data expertise, infrastructure and distribution into models that are reliably useful—and maintain trust while they act on users’ behalf.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
