{"id":11549,"date":"2026-01-03T03:32:25","date_gmt":"2026-01-02T19:32:25","guid":{"rendered":"https:\/\/ai-stack.ai\/?p=11549"},"modified":"2026-01-06T10:36:16","modified_gmt":"2026-01-06T02:36:16","slug":"nvidia-groq-meta-manus-2025","status":"publish","type":"post","link":"https:\/\/ai-stack.ai\/en\/nvidia-groq-meta-manus-2025","title":{"rendered":"Nvidia&#8217;s $20B Groq Deal &amp; Meta&#8217;s Manus Acquisition: The Two Blockbuster AI Deals of 2025 Reshaping the Chip War and AI Agent Landscape"},"content":{"rendered":"\n<p class=\"wp-block-paragraph\">At the end of 2025, the AI industry witnessed two landscape-changing mega-deals: <strong>Nvidia secured Groq&#8217;s LPU technology and core team for $20 billion<\/strong>, and <strong>Meta acquired general-purpose AI agent startup Manus for over $2 billion<\/strong>. These deals reveal three major trends: (1) Tech giants are using &#8220;non-acquisition&#8221; licensing structures to circumvent antitrust scrutiny; (2) The AI chip market is shifting from training to inference; (3) The AI industry is evolving from &#8220;conversation&#8221; to &#8220;action&#8221; in the AI Agent era.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Introduction: The Tip of the AI Iceberg<\/strong><\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The artificial intelligence (AI) industry is racing forward at an unprecedented pace, with new models and applications emerging constantly. However, what truly reshapes the industry landscape and determines future direction isn&#8217;t the glamorous product launches\u2014it&#8217;s the meticulously planned mega-deals happening behind the scenes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Recently, tech giants Nvidia and Meta completed two highly significant transactions. These aren&#8217;t merely business deals\u2014they&#8217;re carefully calculated chess moves. These acquisitions reveal the future direction of the AI war, the deep strategies giants employ to consolidate their dominance, and the industry secrets hidden within this high-speed race.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Nvidia&#8217;s $20 Billion &#8220;Non-Acquisition&#8221;: A Legal Loophole Designed to Absorb Competitors<\/strong><\/h2>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<span class=\"embed-youtube\" style=\"text-align:center; display: block;\"><iframe class=\"youtube-player\" width=\"1300\" height=\"732\" src=\"https:\/\/www.youtube.com\/embed\/MA54gl-2we8?version=3&#038;rel=1&#038;showsearch=0&#038;showinfo=1&#038;iv_load_policy=1&#038;fs=1&#038;hl=en-US&#038;autohide=2&#038;wmode=transparent\" allowfullscreen=\"true\" style=\"border:0;\" sandbox=\"allow-scripts allow-same-origin allow-popups allow-presentation allow-popups-to-escape-sandbox\"><\/iframe><\/span>\n<\/div><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Video Source: <em><a href=\"https:\/\/www.youtube.com\/@JulianGoldieSEO\" target=\"_blank\" rel=\"noreferrer noopener\">https:\/\/www.youtube.com\/@JulianGoldieSEO<\/a><\/em><\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>The Deal&#8217;s Core: A Licensing Agreement, Not an Acquisition<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">At the heart of this deal is<a href=\"https:\/\/www.cnbc.com\/2025\/12\/24\/nvidia-buying-ai-chip-startup-groq-for-about-20-billion-biggest-deal.html\" target=\"_blank\" rel=\"noopener\"> Nvidia paying $20 billion<\/a> to acquire key technology and core team members from AI chip startup Groq. This is not only Nvidia&#8217;s largest deal in its 32-year history but also the highest-value acquisition in AI history.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">However, the most intriguing aspect is that this isn&#8217;t a traditional &#8220;acquisition&#8221;\u2014it&#8217;s a &#8220;non-exclusive licensing agreement.&#8221; Nvidia obtained the technology and hired Groq&#8217;s founder and CEO Jonathan Ross, President Sunny Madra, and other core members.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Nvidia-Groq Deal Key Facts:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Deal Value: $20 billion<\/li>\n\n\n\n<li>Deal Type: Non-exclusive licensing agreement<\/li>\n\n\n\n<li>Key Talent: Jonathan Ross (Groq CEO, Google TPU co-inventor)<\/li>\n\n\n\n<li>Strategic Significance: Strengthening AI inference capabilities while avoiding antitrust scrutiny<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Why Choose a &#8220;Non-Acquisition&#8221; Structure?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">The motivation behind this move is crystal clear: <strong>avoiding strict antitrust scrutiny<\/strong>. Nvidia&#8217;s 2022 attempt to acquire chip designer ARM failed precisely due to regulatory opposition. Now, with Nvidia controlling over 90% of the AI chip market, any direct acquisition would be extremely sensitive.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Through this clever legal structure, Nvidia can absorb potential competitors while perfectly avoiding regulatory scrutiny. As<a href=\"https:\/\/www.cnbc.com\/2025\/12\/26\/nvidia-groq-deal-is-structured-to-keep-fiction-of-competition-alive.html\" target=\"_blank\" rel=\"noopener\"> Bernstein analyst noted<\/a>:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">&#8220;The deal structure keeps the fiction of competition alive.&#8221; \u2014 Bernstein Analyst Stacy Rasgon<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This statement cuts to the heart of the truth: maintaining the illusion of competition in legal form while actually completing further market consolidation.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>The Hidden Cost of &#8220;Reverse Acqui-hire&#8221;: When Startup Employees&#8217; Dreams Evaporate<\/strong><\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">This seemingly brilliant deal structure has a cruel flip side\u2014the impact on ordinary employees. This strategy, called &#8220;reverse acqui-hire,&#8221; is a method tech giants use to acquire top teams and their intellectual property while avoiding the legal and financial responsibilities of a complete acquisition.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>The Sacrifice of Employee Rights<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">In traditional acquisitions, when a company is bought, employees&#8217; stock options are typically vested, often bringing life-changing wealth to early employees\u2014the main reason they&#8217;re willing to accept lower salaries to join startups.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">However, in the Nvidia-Groq licensing agreement model, things are completely different. Since the company wasn&#8217;t legally &#8220;sold,&#8221; stock vesting clauses aren&#8217;t triggered. The result:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><strong>Founders and senior leadership<\/strong>: Recruited by Nvidia with premium compensation packages<\/li>\n\n\n\n<li><strong>Investors<\/strong>: Received handsome returns through licensing fees<\/li>\n\n\n\n<li><strong>Rank-and-file employees<\/strong>: Especially those holding significant unvested equity, may get nothing<\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Now a Standard Playbook for Giants<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">This isn&#8217;t an isolated case\u2014it has become the giants&#8217; standard playbook. Recently, Google&#8217;s acquisition of AI code startup Windsurf, Microsoft&#8217;s acquisition of Inflection AI, and Amazon&#8217;s acquisition of Adept all employed similar approaches, showing that giants are skillfully using this strategy to absorb talent and technology while externalizing costs to abandoned employees.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>The Real Trigger Behind the Deal: How Google&#8217;s Success Kept Nvidia Up at Night<\/strong><\/h2>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Google Gemini 3 and the TPU Milestone<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Nvidia&#8217;s elaborate maneuvering stems from a major breakthrough by competitor Google. In November 2025, Google released its powerful<a href=\"https:\/\/ai-stack.ai\/en\/gemini3\"> Gemini 3 model<\/a>, and the crucial point is: <strong>this top-tier model was trained entirely on Google&#8217;s own TPU (Tensor Processing Unit) chips, without using any Nvidia hardware<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This event proved to the world for the first time that building world-class AI capabilities doesn&#8217;t require Nvidia. When the news broke, Google&#8217;s stock rose while Nvidia&#8217;s fell.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For a deeper understanding of the differences between TPU and GPU, check out our detailed article:<a href=\"https:\/\/ai-stack.ai\/en\/google-tpu-vs-nvidia-gpu\"> Google TPU vs Nvidia GPU Complete Comparison<\/a>.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>The Disruptive Nature of Groq&#8217;s LPU Technology<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">What kept Nvidia even more worried was the disruptive nature of Groq&#8217;s technology itself. As we explored in our<a href=\"https:\/\/ai-stack.ai\/en\/asic-vs-gpu\"> ASIC vs GPU Deep Analysis<\/a>, different chip architectures have their own advantages:<\/p>\n\n\n\n<figure class=\"wp-block-table is-style-stripes\"><table class=\"has-fixed-layout\"><tbody><tr><td><strong>Chip Type<\/strong><\/td><td><strong>Original Design Purpose<\/strong><\/td><td><strong>AI Training<\/strong><\/td><td><strong>AI Inference<\/strong><\/td><\/tr><tr><td>Nvidia GPU<\/td><td>Gaming Graphics<\/td><td>\u2b50\u2b50\u2b50\u2b50\u2b50<\/td><td>\u2b50\u2b50\u2b50<\/td><\/tr><tr><td>Google TPU<\/td><td>AI Computing<\/td><td>\u2b50\u2b50\u2b50\u2b50<\/td><td>\u2b50\u2b50\u2b50\u2b50<\/td><\/tr><tr><td>Groq LPU<\/td><td>AI Inference<\/td><td>\u2b50\u2b50<\/td><td>\u2b50\u2b50\u2b50\u2b50\u2b50<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/techcrunch.com\/2025\/12\/24\/nvidia-acquires-ai-chip-challenger-groq-for-20b-report-says\/\" target=\"_blank\" rel=\"noopener\">According to TechCrunch<\/a>, Groq&#8217;s LPU performs inference tasks <strong>10x faster than Nvidia&#8217;s GPU with 10x better energy efficiency<\/strong>. This is exactly where the AI industry is heading\u2014from model training to large-scale inference deployment.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Nvidia&#8217;s &#8220;Defense Is Offense&#8221; Strategy<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Facing dual threats to its dominance, Nvidia&#8217;s response was textbook &#8220;defense is offense&#8221;: they directly hired the key figure behind Google&#8217;s TPU design\u2014Jonathan Ross, Groq&#8217;s founder.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Nvidia&#8217;s Three Strategic Objectives for Acquiring Groq:<\/strong><\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>Defending against Google TPU threat<\/strong>: Gemini 3 trained entirely on TPU proved you don&#8217;t need Nvidia to build top-tier AI<\/li>\n\n\n\n<li><strong>Strengthening inference capabilities<\/strong>: GPU excels at training but is relatively weaker at inference; LPU is the opposite<\/li>\n\n\n\n<li><strong>Eliminating potential competitors<\/strong>: Absorbing threats directly rather than waiting for rivals to grow strong<\/li>\n<\/ol>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Meta&#8217;s Giant Leap: The Critical Step from &#8220;Chat&#8221; to &#8220;Action&#8221;<\/strong><\/h2>\n\n\n\n<figure class=\"wp-block-embed is-type-video is-provider-youtube wp-block-embed-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio\"><div class=\"wp-block-embed__wrapper\">\n<span class=\"embed-youtube\" style=\"text-align:center; display: block;\"><iframe class=\"youtube-player\" width=\"1300\" height=\"732\" src=\"https:\/\/www.youtube.com\/embed\/NCzpFv9Byus?version=3&#038;rel=1&#038;showsearch=0&#038;showinfo=1&#038;iv_load_policy=1&#038;fs=1&#038;hl=en-US&#038;autohide=2&#038;wmode=transparent\" allowfullscreen=\"true\" style=\"border:0;\" sandbox=\"allow-scripts allow-same-origin allow-popups allow-presentation allow-popups-to-escape-sandbox\"><\/iframe><\/span>\n<\/div><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\"><em>Video Source: <a href=\"https:\/\/www.youtube.com\/watch?v=NCzpFv9Byus\" target=\"_blank\" rel=\"noopener\">https:\/\/www.youtube.com\/watch?v=NCzpFv9Byus<\/a><\/em><\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>What Is a General-Purpose AI Agent?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Turning to another giant, Meta&#8217;s<a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2025-12-29\/meta-acquires-startup-manus-to-bolster-ai-business\" target=\"_blank\" rel=\"noopener\"> acquisition of Manus<\/a> is equally significant. This acquisition gave Meta a key capability it had long lacked in the AI race.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/ai-stack.ai\/en\/manus-ai\">Manus<\/a> is fundamentally different from ChatGPT or Claude that we&#8217;re familiar with. The latter are chatbots that excel at conversation; Manus is a &#8220;general purpose AI agent&#8221; capable of autonomously executing complex tasks. <strong>It doesn&#8217;t just &#8220;talk&#8221;\u2014more crucially, it &#8220;does&#8221;<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/manus.im\/blog\/manus-joins-meta-for-next-era-of-innovation\" target=\"_blank\" rel=\"noopener\">According to Manus&#8217;s official announcement<\/a>, its AI agent can:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Connect to your Google Docs, automatically send emails, and schedule your calendar<\/li>\n\n\n\n<li>Fetch top five news stories daily and automatically generate a Google Doc summary<\/li>\n\n\n\n<li>Independently write code to build websites, analyze data, and even generate videos<\/li>\n\n\n\n<li>Screen resumes, plan travel itineraries, and analyze stock portfolios<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">For more on AI Agent development trends, see:<a href=\"https:\/\/ai-stack.ai\/en\/chatgpt-agents2025\"> ChatGPT Agents 2025 Complete Guide<\/a> and<a href=\"https:\/\/ai-stack.ai\/en\/mcp-ai-agents\"> MCP AI Agents Technical Analysis<\/a>.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Meta&#8217;s Strategic Motivation<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/www.cnbc.com\/2025\/12\/30\/meta-acquires-singapore-ai-agent-firm-manus-china-butterfly-effect-monicai.html\" target=\"_blank\" rel=\"noopener\">According to CNBC<\/a>, Meta&#8217;s motivation is crystal clear. Before this, Meta was clearly behind in the AI agent race compared to<a href=\"https:\/\/openai.com\/index\/introducing-operator\/\" target=\"_blank\" rel=\"noopener\"> OpenAI&#8217;s Operator<\/a> and Google&#8217;s AI agents. Its own Meta AI was basically just a chatbot with limited functionality.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Meta-Manus Deal Key Facts:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Deal Value: Over $2 billion<\/li>\n\n\n\n<li>Manus Founded: March 2025<\/li>\n\n\n\n<li>Data Processed: 147 trillion tokens<\/li>\n\n\n\n<li>Virtual Computers: 80 million<\/li>\n\n\n\n<li>Annualized Revenue: Over $125 million<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Acquiring Manus allows Meta to rapidly close this massive gap, leapfrogging directly to the next stage of AI development: <strong>evolving from &#8220;chat&#8221; to &#8220;do&#8221;<\/strong>.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>The Astonishing Speed of AI: From Launch to Giant Acquisition in Just Ten Months<\/strong><\/h2>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Manus&#8217;s Exponential Growth<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Manus&#8217;s growth story is the perfect embodiment of the AI industry&#8217;s astonishing development speed. Manus officially launched its first general AI agent product in March 2025, and by December of the same year, the company was acquired by Meta. <strong>From product launch to giant acquisition\u2014just under ten months<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In this short period, Manus achieved incredible results:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>Processed <strong>147 trillion tokens<\/strong><\/li>\n\n\n\n<li>Created <strong>80 million virtual computers<\/strong><\/li>\n\n\n\n<li>Served millions of users globally<\/li>\n\n\n\n<li>Reached <strong>$125 million annualized revenue<\/strong><\/li>\n<\/ul>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Comparison with Other Tech Companies<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">To put this speed in perspective:<\/p>\n\n\n\n<figure class=\"wp-block-table is-style-stripes\"><table class=\"has-fixed-layout\"><tbody><tr><td><strong>Company<\/strong><\/td><td><strong>Time to Reach Similar Market Position<\/strong><\/td><\/tr><tr><td>Manus<\/td><td>Less than 1 year<\/td><\/tr><tr><td>Slack<\/td><td>3 years<\/td><\/tr><tr><td>Shopify<\/td><td>7 years<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">This exponential growth speed clearly reveals the AI field&#8217;s uniqueness and brutality: <strong>the innovation window is extremely short\u2014miss it, and you might be left completely behind<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is exactly why enterprises need to<a href=\"https:\/\/ai-stack.ai\/en\/manage-gpu-effectively\"> effectively manage GPU resources<\/a> and choose the right<a href=\"https:\/\/ai-stack.ai\/en\/cloud-or-on-premises\"> cloud or on-premises deployment strategy<\/a> to keep pace with AI development.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Far-Reaching Industry Impact<\/strong><\/h2>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Changes in AI Chip Market Landscape<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">These two blockbuster deals reveal several core trends in the AI industry:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>1. Legal Structure Innovation to Consolidate Monopoly Position<\/strong> Tech giants are using clever legal structures to consolidate their near-monopoly positions. The &#8220;non-exclusive licensing agreement&#8221; model will likely be emulated by more companies, becoming the standard template for future AI industry consolidation.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>2. Shift in Focus from Training to Inference<\/strong> Nvidia&#8217;s acquisition of Groq clearly shows the AI industry&#8217;s center of gravity shifting from model training to large-scale inference deployment. This has significant implications for<a href=\"https:\/\/ai-stack.ai\/en\/what-is-ai-data-center\"> AI data center<\/a> planning.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>3. The AI Agent Era Has Officially Arrived<\/strong> Meta&#8217;s acquisition of Manus marks the AI industry&#8217;s rapid shift from &#8220;conversation&#8221; to &#8220;action.&#8221; Future AI won&#8217;t just answer questions\u2014it will be digital workers capable of autonomously completing complex tasks.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Opportunities and Challenges for Startups<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">Behind these glamorous deals may lie unfair &#8220;human costs&#8221; to ordinary employees. As giants continue consolidating power, the line between competition and innovation grows increasingly blurred.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>FAQ<\/strong><\/h2>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Q: Why didn&#8217;t Nvidia just acquire Groq directly?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A: Since Nvidia already controls over 90% of the AI chip market, any direct acquisition would face strict antitrust scrutiny. Through the &#8220;non-exclusive licensing agreement&#8221; structure, Nvidia can obtain technology and talent while avoiding regulatory scrutiny. According to analysts, this structure &#8220;keeps the fiction of competition alive.&#8221;<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Q: What&#8217;s the difference between Groq&#8217;s LPU and Nvidia&#8217;s GPU?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A: GPU was originally designed for gaming graphics and later applied to AI training, excelling in the training phase. LPU (Language Processing Unit) was designed from scratch specifically for AI inference (generating responses). According to reports, LPU performs inference tasks 10x faster than GPU with 10x better energy efficiency. For detailed comparison, see<a href=\"https:\/\/ai-stack.ai\/en\/asic-vs-gpu\"> ASIC vs GPU Complete Analysis<\/a>.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Q: Why did Meta acquire Manus?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A: Meta was behind in the AI agent race compared to OpenAI&#8217;s Operator and Google&#8217;s AI agents. Acquiring Manus allows Meta to quickly evolve from &#8220;chatbot&#8221; to &#8220;AI agent&#8221; capable of autonomously executing tasks, closing the gap with competitors. This deal gives Meta mature technology that has processed 147 trillion tokens and served millions of users.<\/p>\n\n\n\n<h3 class=\"wp-block-heading\"><strong>Q: What do these acquisitions mean for regular users?<\/strong><\/h3>\n\n\n\n<p class=\"wp-block-paragraph\">A: In the short term, this means more powerful AI services: After Nvidia integrates LPU technology, AI inference speed and efficiency will significantly improve; After Meta integrates Manus, AI assistants on Facebook, Instagram, and WhatsApp will be able to perform more complex tasks, not just answer questions. Long-term, this may accelerate AI Agent adoption, changing how we interact with technology.<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Conclusion: Reflections on the Giants&#8217; Chess Game<\/strong><\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">These two blockbuster deals reveal several core trends in the AI industry:<\/p>\n\n\n\n<ol class=\"wp-block-list\">\n<li><strong>Legal innovation serving market consolidation<\/strong>: Tech giants are using clever legal structures to consolidate their near-monopoly positions<\/li>\n\n\n\n<li><strong>AI shifting from &#8220;talking&#8221; to &#8220;doing&#8221;<\/strong>: The AI development focus is rapidly shifting from &#8220;conversation&#8221; to &#8220;action&#8221;<\/li>\n\n\n\n<li><strong>Who bears the cost of success?<\/strong>: Behind these glamorous deals may lie unfair &#8220;human costs&#8221; to ordinary employees<\/li>\n<\/ol>\n\n\n\n<p class=\"wp-block-paragraph\">As giants continue consolidating power, the line between competition and innovation grows increasingly blurred. What does this mean for the future of the entire AI ecosystem? Will the next world-changing innovation come from agile startups, or from within the giants&#8217; towering walls?<\/p>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<h2 class=\"wp-block-heading\"><strong>Further Reading<\/strong><\/h2>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/ai-stack.ai\/en\/google-tpu-vs-nvidia-gpu\">Google TPU vs Nvidia GPU: The AI Chip Dominance Battle<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/ai-stack.ai\/en\/gemini3\">Gemini 3 Complete Guide: Google&#8217;s New AI Milestone<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/ai-stack.ai\/en\/manus-ai\">Manus AI: The Rise of General-Purpose AI Agents<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/ai-stack.ai\/en\/chatgpt-agents2025\">ChatGPT Agents 2025: The AI Agent Era Arrives<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/ai-stack.ai\/en\/asic-vs-gpu\">ASIC vs GPU: Complete AI Chip Architecture Comparison<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/ai-stack.ai\/en\/manage-gpu-effectively\">How to Effectively Manage GPU Resources<\/a><\/li>\n<\/ul>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Sources:<\/strong><\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li><a href=\"https:\/\/www.cnbc.com\/2025\/12\/24\/nvidia-buying-ai-chip-startup-groq-for-about-20-billion-biggest-deal.html\" target=\"_blank\" rel=\"noopener\">CNBC: Nvidia buying AI chip startup Groq for about $20 billion<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.cnbc.com\/2025\/12\/26\/nvidia-groq-deal-is-structured-to-keep-fiction-of-competition-alive.html\" target=\"_blank\" rel=\"noopener\">CNBC: Nvidia-Groq deal keeps &#8216;fiction of competition alive&#8217;<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.bloomberg.com\/news\/articles\/2025-12-29\/meta-acquires-startup-manus-to-bolster-ai-business\" target=\"_blank\" rel=\"noopener\">Bloomberg: Meta to Buy Manus<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/www.cnbc.com\/2025\/12\/30\/meta-acquires-singapore-ai-agent-firm-manus-china-butterfly-effect-monicai.html\" target=\"_blank\" rel=\"noopener\">CNBC: Meta acquires Manus<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/manus.im\/blog\/manus-joins-meta-for-next-era-of-innovation\" target=\"_blank\" rel=\"noopener\">Manus Official Blog<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/techcrunch.com\/2025\/12\/24\/nvidia-acquires-ai-chip-challenger-groq-for-20b-report-says\/\" target=\"_blank\" rel=\"noopener\">TechCrunch: Nvidia acquires Groq<\/a><\/li>\n\n\n\n<li><a href=\"https:\/\/newsletter.semianalysis.com\/p\/tpuv7-google-takes-a-swing-at-the\" target=\"_blank\" rel=\"noopener\">SemiAnalysis: Google TPU Analysis<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>At the end of 2025, the AI industry witnessed two landscape-changing mega-deals: Nvidia secured Groq&#8217;s LPU technology and core team for $20 billion, and Meta acquired general-purpose AI agent startup Manus for over $2 billion. These deals reveal three major trends: (1) Tech giants are using &#8220;non-acquisition&#8221; licensing structures to circumvent antitrust scrutiny; (2) The AI chip market is shifting from training to inference; (3) The AI industry is evolving from &#8220;conversation&#8221; to &#8220;action&#8221; in the AI Agent era. Introduction: The Tip of the AI Iceberg The artificial intelligence (AI) industry is racing forward at an unprecedented pace, with new models&#8230;<\/p>\n","protected":false},"author":253372376,"featured_media":11550,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_crdt_document":"","jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[96987604,96987592],"tags":[96987776],"class_list":["post-11549","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news","category-featured-articles","tag-nvidia-en"],"blocksy_meta":[],"acf":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/ai-stack.ai\/wp-content\/uploads\/2026\/01\/%E6%A8%A1%E5%9E%8BA-31.jpg?fit=1920%2C1080&quality=100&ct=202603031250&ssl=1","jetpack_shortlink":"https:\/\/wp.me\/ph344V-30h","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/11549","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/users\/253372376"}],"replies":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/comments?post=11549"}],"version-history":[{"count":2,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/11549\/revisions"}],"predecessor-version":[{"id":11564,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/11549\/revisions\/11564"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media\/11550"}],"wp:attachment":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media?parent=11549"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/categories?post=11549"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/tags?post=11549"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}