Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
The company behind the “very dark AI company” description is Scale AI. In June 2025, Meta made a reported $14.3 billion investment for a 49% non-voting minority stake—not a full acquisition—and Scale co-founder Alexandr Wang left his CEO role to join Meta’s AI effort. The label is editorial, not an official designation: it points mainly to Scale’s defense work and criticism of the human labor behind AI training data.
What Meta invested in
Scale AI is an AI data, evaluation, and infrastructure-services company, not a consumer chatbot maker. It prepares and labels data and helps customers evaluate and test machine-learning systems. Those services support the development and refinement of AI models. Scale was founded by Alexandr Wang and Lucy Guo. Scale AI describes its business, while The Associated Press and TechCrunch covered the company and transaction.
How large was Meta’s investment?
Scale announced a “significant new investment” on June 12, 2025, and said the transaction valued the company at more than $29 billion. Major outlets reported that Meta invested approximately $14.3 billion for about 49% of Scale. Meta’s later SEC filing confirmed that its interest was a non-voting minority stake. The reported dollar amount and percentage are not figures stated in Scale’s announcement; they should therefore be understood as reported deal terms, not as details Scale published in that announcement. Scale’s announcement, Bloomberg’s deal report, TechCrunch’s report, and Meta’s SEC filing describe different parts of the transaction.
That distinction matters: Meta did not buy all of Scale AI, and a large economic stake is not the same thing as voting control or ownership of every customer relationship or dataset. Scale said it would remain independent and that the investment would not integrate its operations into Meta. Its assurances about protecting customer data are the company’s own statements, not independent verification. Scale’s customer-trust statement explains its position.
#1 Best Overall
- System Compatibility Note: 2-slot card, 271x112x39mm, single 8-pin power, 200W TDP. Verify chassis clearance and PSU capacity before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- 24GB GDDR6 on 192-Bit Bus: Massive 24GB memory with 456 GB/s bandwidth – ideal for LLMs, AI inference, 3D rendering, and generative design.
- Intel Xe2-HPG Architecture: Built on Intel's next-gen architecture with 20 Xe cores and 160 XMX engines for AI acceleration (197 INT8 TOPS).
- PCIe 5.0 Support: PCI Express 5.0 x16 interface for maximum bandwidth with the latest workstation platforms.
Why Alexandr Wang moved to Meta
Scale said Wang would leave his CEO role to work on Meta’s AI efforts while remaining on Scale’s board. Scale named its chief strategy officer, Jason Droege, interim CEO. Meta CEO Mark Zuckerberg later identified Wang as leading Meta’s superintelligence effort, with Nat Friedman leading AI products and applied research and Shengjia Zhao serving as chief scientist. Scale’s announcement and Meta’s statement on its Q2 2025 results set out those roles.
Meta’s apparent rationale combines recruitment, commercial cooperation, and access to expertise in data generation and model evaluation. Scale said the agreement would substantially expand its commercial relationship with Meta. Reporting also highlighted Wang’s government relationships as a potential asset as Meta pursued government-facing AI work. The public information supports a mix of strategic aims; it does not establish that Meta invested solely to obtain Scale’s data. Axios’s analysis discusses the deal’s strategic and customer implications.
Why Scale’s military work attracts criticism
Scale has publicly described defense contracts for AI-enabled data, planning, and decision-support work. One example is Thunderforge, a program awarded through the Defense Innovation Unit and intended to support military decision-making and operations. Scale has also announced a $500 million Pentagon AI partnership expansion centered on Scale Donovan and related defense capabilities. The company’s descriptions and Wang’s congressional testimony provide context for its government work; see also Scale’s announcements on Thunderforge and the Pentagon partnership expansion.
Free tools Windows power users keep installed
One-click scans. No signup required.
Rank #2
- System Compatibility Note: This 2-slot card measures 271 x 112 x 39 mm and requires a single 12V-2x6-pin power connector. Please verify chassis and PSU compatibility before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- Professional Intel Arc Pro B70 GPU: Built on the Intel Xe2-HPG architecture, it features 32 Xe cores and 256 XMX engines, designed to accelerate AI, rendering, and complex visualization workloads.
- Massive 32GB GDDR6 VRAM: Equipped with 32GB of high-speed GDDR6 memory on a 256-bit bus, running at 19 Gbps, which allows for handling large AI models and complex datasets locally.
- High-Performance Engine Clock: Delivers an engine clock of 2540 MHz, providing the compute power needed for demanding professional applications and AI inference.
These facts do not establish that Scale itself operates autonomous weapons or makes final decisions about targeting. Three activities often blurred together in criticism are different:
- Data labeling and evaluation: people or systems organize, annotate, rate, or test information used in AI development.
- AI-enabled decision support: software helps users analyze information or plan operations; people remain responsible for how they act on that support.
- Autonomous weapons control: a system selects or engages targets without the same level of direct human decision-making. The cited descriptions of Scale’s contracts do not, on their own, establish that Scale controls such systems.
That distinction does not settle the ethical debate. Critics can reasonably question how data and decision-support tools may contribute to military operations, even where a company is not itself making a lethal decision. An earlier congressional hearing record also addresses Scale’s defense-related work: the House hearing material.
The labor concerns behind the “dark” description
Human workers have been central to the data-labeling and evaluation work used to develop AI systems. Tasks can include classifying images, rating text, and assessing model responses. That work can be invisible to people who use AI products, even though it can shape how models behave and what they learn.
Rank #3
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
A California complaint against Scale alleges labor-law violations involving workers doing generative-AI data-labeling work. A complaint documents allegations, not a final ruling that they are true. The filing is available here. Without treating the claims as established findings, the controversy raises concrete questions about pay, worker protections, location, exposure to sensitive material, and what safeguards apply when contractors handle customer information.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Futurism used the “dark” framing to connect the investment with defense work and labor criticism. That is opinionated coverage, not a formal label or proof of illegality. Futurism’s article presents that critical framing; readers should distinguish its characterization from what contracts, company statements, and legal filings establish.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What could go wrong for Meta and Scale?
The investment gives Meta a significant economic interest without making Scale a wholly owned subsidiary. That structure can create strategic upside and unresolved tensions at the same time.
Rank #4
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
- Customer trust: Scale serves multiple AI companies, including businesses competing with Meta. Those customers may question whether the relationship affects confidentiality or neutrality, even though a 49% stake does not by itself prove Meta has access to their data.
- Governance: A non-voting minority stake is not equivalent to control. The distinction can complicate accountability when Meta has a large financial interest and Wang holds roles at both companies.
- Competition scrutiny: Regulators or competitors could examine the combination of a major investment and the recruitment of Scale’s founder. That is a potential scrutiny issue, not evidence that regulators have found the deal unlawful.
- Execution and price risk: Better data and evaluation capabilities do not guarantee a breakthrough model. The reported price is substantial for a business whose services rely in part on human labor.
- Reputational and ethical exposure: Defense applications and labor practices may attract public and regulatory criticism regardless of whether the company remains formally independent.
The tension is not simply whether Meta “owns the data.” That has not been established. The sharper question is whether Scale can continue serving a broad customer base while a major competitor becomes its largest reported investor and recruits its founder.
What “very dark” means—and what it does not
In this headline, “dark” is shorthand for controversies around military AI and the human work behind AI data services. It is not a formal designation of Scale as criminal, illegal, or a weapons manufacturer. The most precise description is that Scale is an AI data and evaluation provider with defense contracts, while a lawsuit has alleged labor-law violations and the company says it remains independent and protects customer data.
Do these 3 things before closing this tab:
1Repair Windows errors before they cause bigger problems2Fix the driver behind crashes, sound loss and screen glitches3Clear out junk files and repair common Windows errorsQuick 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.
