December closed out 2025 with a burst of high‑impact activity in the artificial intelligence sector, marked by major acquisitions and strategic investments from Nvidia, Meta, SoftBank, and IBM. These moves underscore intensifying competition across AI hardware, agentic automation, cloud infrastructure, and enterprise data systems. For antitrust observers and practitioners, these developments serve as another indicator of accelerating vertical integration and consolidation across the AI supply chain — from compute to models to deployment‑layer tools.
These deals are included in the latest update to the Mogin Law AI Deal Table.
Nvidia Acquires Major Groq Assets to Consolidate Its Inference Dominance
Nvidia made the largest AI‑related acquisition of the month by purchasing substantial assets from Groq — an AI‑chip startup known for its specialized Language Processing Unit (LPU) designed for high‑speed, energy‑efficient inference. The approximately $20 billion transaction includes Groq’s core LPU technology and the hiring of its CEO and senior engineering leadership, further entrenching Nvidia’s position atop the AI‑compute stack. Read more.
This move builds on a year in which Nvidia aggressively expanded its hold on both training and inference silicon. With Groq’s inference‑optimized architecture absorbed into Nvidia’s portfolio, the company strengthens control over key inputs essential to model deployment. As regulators globally evaluate the competitive impacts of compute concentration, Nvidia’s continued acquisition of both human capital and proprietary chip technologies may attract renewed scrutiny.
Groq is a privately held AI‑chip company founded in 2016 by former Google engineer Jonathan Ross and other ex‑Alphabet engineers, according to reporting from Reuters via U.S. News and background details from Wikipedia. The company reported $3.2 million in revenue in 2023 per its public financials on Wikipedia’s Groq page, and it is expected to generate about $500 million in 2025, driven largely by a major Saudi infrastructure commitment, as reported by Reuters. Ownership is distributed among founders and a broad mix of institutional investors — including Disruptive, BlackRock, Neuberger Berman, Samsung, and Cisco — detailed in coverage by U.S. News/Reuters and Tech Startups. The company continues operating independently under newly shifted leadership, as reported by CNBC.
Meta announced its acquisition of Manus, a viral agentic AI system that drew widespread attention earlier in 2025. The deal, revealed during the year‑end holiday period, fits into Meta’s broader strategy of enhancing its ecosystem of autonomous digital agents and interactive AI products. Read more.
Agentic AI — systems capable of independently carrying out multi‑step tasks — is poised to become a competitive battleground as platforms race to integrate assistants into messaging, commerce, productivity, and immersive environments. Meta’s acquisition signals its intent to build a deeper bench of agent‑based tools, potentially expanding the company’s reach within both consumer and enterprise AI workflows. As Meta layers increasingly capable agents into its platform, overlapping markets with competitors like OpenAI, Google, and Anthropic may raise familiar questions around bundling, self‑preferencing, and data access.
Manus is a Chinese‑founded AI startup that launched in March 2025 and relocated its headquarters to Singapore later that year, according to Business Insider. Built by the AI product studio Butterfly Effect, Manus markets what it calls the world’s first “general” AI agent, capable of autonomously executing tasks such as market research, coding, slide creation, and data analysis. The company grew rapidly, surpassing $100 million in annual recurring revenue with a $125 million revenue run rate just eight months after launch, as reported by CNBC. Manus was co‑founded by Xiao Hong and Ji Yichao, both of whom previously built AI and productivity products in China, per Business Insider. The company raised $75 million in funding led by Benchmark and was backed by Tencent, ZhenFund, and HSG before agreeing to be acquired by Meta for more than $2 billion, a deal reported by The Wall Street Journal and summarized by The Detroit News. Meta says Manus will continue operating as a standalone subscription product while its agentic technology is integrated into Meta’s broader AI systems.
SoftBank Moves to Acquire DigitalBridge, Expanding AI Compute Infrastructure
SoftBank announced it would acquire DigitalBridge, a major digital‑infrastructure investor, in a transaction estimated at $4 billion. This acquisition reflects SoftBank’s growing interest in AI‑enabling infrastructure, particularly data‑center capacity and edge deployments. Read more.
As generative AI models continue to scale, availability of compute — especially low‑latency, high‑throughput infrastructure — has become one of the most constrained resources in the global technology economy. DigitalBridge’s footprint gives SoftBank a larger platform in the physical‑infrastructure layer that underpins AI workloads. Because compute has become a supply‑chain chokepoint with significant competitive implications, acquisitions of large infrastructure providers may attract regulatory interest depending on market overlaps and regional saturation.
DigitalBridge is a global digital infrastructure investment firm headquartered in Boca Raton, Florida. The company was founded in 1991 as Colony Capital before rebranding to DigitalBridge in 2021, marking its strategic shift to digital infrastructure investments such as data centers, fiber networks, and edge deployments. In 2024, DigitalBridge reported annual revenue of approximately $607 million, though its trailing twelve-month revenue for 2025 fell to about $112 million amid portfolio adjustments, according to Stock Analysis. Ownership is dominated by institutional investors, with Vanguard Group, Wafra, and BlackRock among the largest stakeholders, and insiders such as founder Thomas Barrack holding notable positions (See WallStreetZen). As of mid-2025, DigitalBridge managed over $106 billion in assets under management, reinforcing its role as a key player in the global digital infrastructure ecosystem (More at Wikipedia).
IBM Acquires Confluent to Strengthen Enterprise Data Streaming and AI Automation
IBM rounded out the month with an $11 billion acquisition of Confluent, a leading provider of real‑time data‑streaming technology. The company intends to integrate Confluent’s capabilities to support more robust enterprise data pipelines, accelerate automated workflows, and enhance AI‑driven decision‑making across industries. Read more.
As enterprises modernize their internal data systems to power AI applications, control over streaming‑data infrastructure has become a strategic priority. IBM’s move strengthens its already significant position in enterprise software and cloud services, pushing the company deeper into industries where real‑time analytics and automation are quickly becoming essential. Depending on the degree of vertical integration created across IBM’s data, cloud, and AI platforms, antitrust authorities may examine whether the combination could disadvantage third‑party competitors reliant on similar data‑flow tools.
Confluent is an enterprise data‑streaming company founded by the original creators of Apache Kafka, specializing in real‑time data pipelines that connect, process, and govern information across cloud and on‑premise systems. According to IBM’s December 8, 2025 press release, IBM agreed to acquire Confluent for $11 billion, offering $31 per share in a move aimed at building an end‑to‑end “smart data platform” to support generative and agentic AI. Confluent’s platform—used to prepare clean, connected, real‑time data for AI systems—underpins a growing portion of enterprise automation, analytics, and hybrid‑cloud architectures, as noted by TechCrunch. Co‑founded by Jay Kreps, Neha Narkhede, and Jun Rao, the company will integrate its cloud and self‑managed streaming technologies into IBM’s broader data and AI ecosystem once the acquisition closes, a move expected to accelerate IBM’s competitiveness in enterprise AI and real‑time data management, according to MarketChameleon.
Conclusion
These December transactions highlight the rapid evolution of the AI ecosystem and the aggressive strategic moves by major firms seeking control over essential inputs — compute hardware, data infrastructure, and intelligent agentic systems. For legal professionals focused on antitrust, merger review, and competitive‑market structure, the deals offer early indicators of where consolidation pressures may intensify in 2026.
