{"id":13804,"date":"2026-07-24T16:00:00","date_gmt":"2026-07-24T08:00:00","guid":{"rendered":"https:\/\/ai-stack.ai\/?p=13804"},"modified":"2026-07-23T16:08:19","modified_gmt":"2026-07-23T08:08:19","slug":"waic-2026-ai-agents-robots-chips","status":"publish","type":"post","link":"https:\/\/ai-stack.ai\/en\/waic-2026-ai-agents-robots-chips","title":{"rendered":"WAIC 2026 Deep Dive: AI&#8217;s Pivot from &#8220;Chatting&#8221; to &#8220;Doing&#8221; \u2014 Robots Go to Work, Chip Wars Heat Up"},"content":{"rendered":"<style>table{border-collapse:collapse;width:100%;margin:1em 0}th,td{border:1px solid #ddd;padding:8px 12px;text-align:left}th{background-color:#f5f5f5;font-weight:bold}tr:nth-child(even){background-color:#fafafa}<\/style>\n<h1 id=\"waic-2026-deep-dive-ais-pivot-from-chatting-to-doing-robots-go-to-work-chip-wars-heat-up\">WAIC 2026 Deep Dive: AI\u2019s Pivot from \u201cChatting\u201d to \u201cDoing\u201d \u2014 Robots Go to Work, Chip Wars Heat Up<\/h1>\n<p><strong>July 20, Shanghai<\/strong> \u2014 When 300 world-premiere products debut simultaneously, 200 robots autonomously handle factory-floor tasks on the expo floor, and AMD chooses the same day to launch its first rack-scale AI system directly challenging NVIDIA, the 2026 World Artificial Intelligence Conference (WAIC) is no longer just an annual tech gathering. It marks the moment the AI industry pivoted from a model-parameter arms race to a deployment-at-scale race.<\/p>\n<p>Themed \u201cIntelligent Partners, Co-creating the Future,\u201d WAIC 2026 set new records: exhibition space surpassed <strong>100,000 square meters<\/strong> for the first time, drawing <strong>1,100+ enterprises<\/strong>, <strong>3,000+ products<\/strong>, and <strong>300+ world premieres<\/strong> (<a href=\"https:\/\/www.news.cn\/tech\/20260719\/e6d801e02f1341309006fd6177da8bcb\/c.html\" target=\"_blank\" rel=\"noopener\">via Xinhua<\/a>). Fourteen concurrent forums hosted 9 Turing Award and Nobel Prize laureates alongside 1,400+ global guests, generating <strong>\u00a5162 billion in cooperation intent<\/strong> (approx. $22B USD).<\/p>\n<p>Three unmistakable signals cut through the noise: AI Agents are moving from concept to product, embodied intelligence is crossing from lab demos to factory floors, and global AI governance is shifting from unilateral controls to multilateral institutions. Here\u2019s the full breakdown across six dimensions.<\/p>\n<hr \/>\n<h2 id=\"record-scale-1100-exhibitors-300-premieres-ais-deployment-reckoning\">1. Record Scale: 1,100 Exhibitors, 300+ Premieres \u2014 AI\u2019s \u201cDeployment Reckoning\u201d<\/h2>\n<p>WAIC 2026\u2019s sheer numbers tell a story of industrial maturation. Compared to 2025, exhibitor count grew nearly <strong>40%<\/strong>, floor space expanded roughly <strong>90%<\/strong>, and world-premiere products doubled. But the structural shift matters more:<\/p>\n<p><strong>63.8% of exhibiting LLM companies now identify as \u201cAI Agent \/ application companies,\u201d<\/strong> up from under 30% in 2025 (<a href=\"https:\/\/www.tmtpost.com\/8072380.html\" target=\"_blank\" rel=\"noopener\">via TMTPost<\/a>). The hall layout reinforced this: H1 was all about models and agents, H2 housed computing infrastructure, and H3 was the embodied intelligence arena \u2014 a physical mapping of the AI value chain.<\/p>\n<p>The no-shows were equally telling: <strong>DeepSeek<\/strong> has never exhibited, <strong>ByteDance<\/strong> skipped its second consecutive year, and <strong>Zhipu AI<\/strong> was absent for the first time in three years due to a pre-IPO quiet period. The center of gravity in China\u2019s AI industry is shifting from a handful of star players to a broader ecosystem.<\/p>\n<p>\ud83d\udd17 For deeper context on AI\u2019s transition from model research to commercial deployment, see our earlier analysis: <a href=\"https:\/\/ai-stack.ai\/en\/chatgpt-agents2025\">ChatGPT Agents 2025: The Era of Autonomous AI Tasks<\/a>.<\/p>\n<hr \/>\n<h2 id=\"the-agent-tipping-point-from-chat-windows-to-autonomous-execution\">2. The Agent Tipping Point: From Chat Windows to Autonomous Execution<\/h2>\n<p>If WAIC 2025\u2019s keyword was \u201clarge models,\u201d WAIC 2026\u2019s was unquestionably \u201cAI Agents.\u201d Every headline-grabbing product on the floor shared one capability: autonomous, multi-step task execution across applications.<\/p>\n<p><strong>StepFun\u2019s STEPX Neo<\/strong> \u2014 a smartphone with no app icons, just a dialog box \u2014 was named one of the conference\u2019s ten \u201cHall Treasures.\u201d Tell it \u201cbook me a hotel near Jing\u2019an for tomorrow, reserve a nearby Italian restaurant, and call me a car,\u201d and the phone autonomously price-compares hotels, makes reservations, and hails a ride across multiple apps. It\u2019s the world\u2019s first phone to pass China\u2019s L3 certification for AI terminal intelligence.<\/p>\n<p><strong>Baidu\u2019s \u201cDuMate\u201d<\/strong> took the software-only route as the sole general-purpose AI Agent among the Hall Treasures, running seamlessly across phone, PC, and in-vehicle systems \u2014 betting on ecosystem coverage over dedicated hardware.<\/p>\n<p><strong>Tencent<\/strong> debuted its <strong>WorkBuddy<\/strong> standalone app alongside AI-infused smart glasses, signaling that Agent products are moving from developer tools to consumer endpoints.<\/p>\n<p>IDC projects the global AI Agent market to grow at a <strong>139% CAGR from 2025 to 2030<\/strong> \u2014 transforming from an emerging category into a multi-hundred-billion-dollar market within five years (<a href=\"https:\/\/news.cgtn.com\/news\/2026-07-17\/Beyond-bigger-models-What-WAIC-2026-reveals-about-AI-s-next-chapter-1OQOdVTqqsg\/p.html\" target=\"_blank\" rel=\"noopener\">via CGTN<\/a>).<\/p>\n<hr \/>\n<h2 id=\"robots-that-work-from-stage-props-to-production-lines\">3. Robots That Work: From Stage Props to Production Lines<\/h2>\n<p>If AI Agents were WAIC 2026\u2019s software headline, embodied intelligence robots were its hardware showstopper. Hall H3 packed <strong>200+ embodied intelligence companies<\/strong> (up from ~80 in 2025), showcasing <strong>208 terminal products and 300+ physical robots<\/strong>. They weren\u2019t dancing \u2014 they were assembling, sorting pharmaceuticals, directing traffic, and loading production lines.<\/p>\n<p><strong>The production numbers tell the real story:<\/strong><\/p>\n<table>\n<colgroup>\n<col style=\"width: 22%\" \/>\n<col style=\"width: 52%\" \/>\n<col style=\"width: 25%\" \/>\n<\/colgroup>\n<thead>\n<tr>\n<th>Company<\/th>\n<th>Production Milestone<\/th>\n<th>Timeline<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>AgiBot<\/strong><\/td>\n<td>15,000+ units shipped; targeting 40\u201350K for full-year 2026<\/td>\n<td>June 2026<\/td>\n<\/tr>\n<tr>\n<td><strong>UBTech<\/strong><\/td>\n<td>13,361 orders for U1 full-size humanoid series<\/td>\n<td>H1 2026<\/td>\n<\/tr>\n<tr>\n<td><strong>XPENG<\/strong><\/td>\n<td>Targeting 1,000 IRON humanoids\/month; in-store sales guides by Q1 2027<\/td>\n<td>End 2026<\/td>\n<\/tr>\n<tr>\n<td><strong>Beijing Innovation Center<\/strong><\/td>\n<td>\u201cTiangong 3.0\u201d mass production and delivery<\/td>\n<td>H2 2026<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>China shipped <strong>14,400 humanoid robots<\/strong> in 2025, representing <strong>84.7% of the global market<\/strong>. First-half 2026 production already exceeded <strong>40,000 units<\/strong>, with full-year projections above 100,000 (<a href=\"http:\/\/ex.chinadaily.com.cn\/exchange\/partners\/70\/rss\/channel\/www\/columns\/y38633\/stories\/WS6a4da9d1a310986e2b464131.html\" target=\"_blank\" rel=\"noopener\">via China Daily<\/a>).<\/p>\n<p><strong>The critical shift: the bottleneck is no longer \u201ccan we build them\u201d \u2014 it\u2019s \u201cwhere do they pay off?\u201d<\/strong> AgiBot deployed 8 robots for 24\/7 operation at a Longcheer Technology tablet assembly line in Nanchang, handling material loading and transfer with <strong>1mm precision and 100% quality-inspection accuracy<\/strong>. Embodied intelligence has crossed the chasm from proof-of-concept to production deployment. The next challenge is business-model: who pays for maintenance, upgrades, and continuous operation?<\/p>\n<p>\ud83d\udd17 For more on AI hardware evolution, see: <a href=\"https:\/\/ai-stack.ai\/en\/gpu-npu-tpu-lpu\">GPU, NPU, TPU, LPU: A Complete Guide to AI Chip Architectures<\/a>.<\/p>\n<hr \/>\n<h2 id=\"the-chip-arms-race-amd-helios-google-frozen-v2-and-nvidias-two-front-response\">4. The Chip Arms Race: AMD Helios, Google Frozen v2, and NVIDIA\u2019s Two-Front Response<\/h2>\n<p>Three major hardware stories broke around WAIC 2026, showing that AI chip competition has escalated from single-card benchmarks to system-level warfare.<\/p>\n<h3 id=\"amd-helios-a-direct-challenge-to-nvidias-rack-dominance\">AMD Helios: A Direct Challenge to NVIDIA\u2019s Rack Dominance<\/h3>\n<p>On July 20, AMD unveiled <strong>Helios<\/strong>, its first rack-scale AI system. The 72-GPU, fully liquid-cooled rack integrates MI455X GPUs (CDNA 5), 6th-Gen EPYC Venice CPUs (Zen 6, TSMC 2nm), and Pensando networking silicon \u2014 delivering <strong>31 TB of HBM4 memory<\/strong>, <strong>2.9 Exaflops FP4 inference<\/strong>, and drawing roughly 225\u2013245 kW per rack at an estimated $5.0\u20135.5 million per unit (<a href=\"https:\/\/wccftech.com\/amd-helios-ai-rack-mi455x-6th-gen-epyc-challenging-nvidia\/\" target=\"_blank\" rel=\"noopener\">via Wccftech<\/a>).<\/p>\n<p><strong>Microsoft<\/strong> has signed on as a launch customer, deploying Helios on Azure for frontier model inference with three new VM series. Other disclosed adopters include <strong>Meta<\/strong> (long-term agreements up to 6 GW), <strong>OpenAI<\/strong>, <strong>Oracle<\/strong>, and <strong>Tata Consultancy<\/strong>. AMD projects data-center AI revenue in the hundreds of billions beginning 2027.<\/p>\n<h3 id=\"google-frozen-v2-etching-gemini-directly-into-silicon\">Google Frozen v2: Etching Gemini Directly into Silicon<\/h3>\n<p>The Information reported on July 20 that Google is developing <strong>\u201cFrozen v2\u201d<\/strong> \u2014 an inference chip that hardcodes Gemini\u2019s neural architecture (attention mechanisms, layer structures, residual connections) directly into physical silicon circuits, eliminating the runtime data-movement and architecture-computation overhead that general-purpose chips incur. The chip reportedly achieves <strong>6\u201310\u00d7 better tokens-per-watt efficiency<\/strong> versus Google\u2019s latest Trillium TPUs, with deployment targeted for <strong>2028<\/strong> (<a href=\"https:\/\/www.cnbc.com\/2026\/07\/20\/alphabet-googl-stock-ai-chip-report.html\" target=\"_blank\" rel=\"noopener\">via CNBC<\/a>).<\/p>\n<p>Unlike the original \u201cFrozen\u201d concept (which hardcoded model <em>weights<\/em>, locking the chip to a single static model version), Frozen v2 uses a <strong>\u201cflexible hardcoding\u201d<\/strong> approach: the architectural skeleton is fixed in silicon, but model weights remain updatable \u2014 extending the chip\u2019s useful life as long as future Gemini iterations share the same base architecture. Alphabet shares rose ~3% on the news.<\/p>\n<h3 id=\"nvidia-agent-toolkit-and-a-20b-groq-hedge\">NVIDIA: Agent Toolkit and a $20B Groq Hedge<\/h3>\n<p>NVIDIA launched its <strong>Agent Toolkit<\/strong>, enabling local AI agent deployment on DGX Station (GB300 Blackwell Ultra) in roughly <strong>30 minutes<\/strong>, integrated with Omniverse for robotics and digital-twin simulation. Meanwhile, NVIDIA\u2019s <strong>$20 billion technology licensing deal with Groq<\/strong> signals a hedge in the inference market, where custom silicon threatens its GPU dominance.<\/p>\n<p>\ud83d\udd17 For the competitive landscape: <a href=\"https:\/\/ai-stack.ai\/en\/google-tpu-vs-nvidia-gpu\">Google TPU vs NVIDIA GPU: The AI Accelerator Showdown<\/a> and <a href=\"https:\/\/ai-stack.ai\/en\/gpu-npu-tpu-lpu\">GPU, NPU, TPU, LPU: Four AI Chip Architectures Compared<\/a>.<\/p>\n<hr \/>\n<h2 id=\"global-governance-goes-institutional-29-nations-sign-waico\">5. Global Governance Goes Institutional: 29 Nations Sign WAICO<\/h2>\n<p>The most consequential political signal at WAIC 2026 was the formal signing of the <strong>World AI Cooperation Organization (WAICO)<\/strong> agreement by <strong>29 nations<\/strong> \u2014 including Kazakhstan, Laos, Pakistan, Russia, and Indonesia \u2014 marking the shift from great-power-dominated AI governance to a multilateral framework.<\/p>\n<p>UN Secretary-General <strong>Ant\u00f3nio Guterres<\/strong> opened the conference by declaring that AI \u201ccannot be governed by a handful of countries or companies,\u201d announcing forthcoming proposals for a <strong>global AI fund<\/strong>. China committed to providing <strong>5,000 AI training slots<\/strong> to developing nations over five years, establishing international AI application cooperation centers across ASEAN, AU, Arab League, CELAC, SCO, and BRICS regions, and deploying the \u201cMazu\u201d meteorological AI early-warning system to 30 countries.<\/p>\n<p>Three key documents were released: the <em>AI Cooperation Development Action Plan<\/em>, the <em>International AI Ethics Governance Action Plan<\/em>, and \u2014 most notably \u2014 the <em>Agent Interconnection &amp; Interoperability Global Cooperation Initiative<\/em>, which directly tackles one of the Agent era\u2019s thorniest technical questions: how do agents from different vendors securely communicate and collaborate?<\/p>\n<p>\ud83d\udd17 For the geopolitical dimension, see: <a href=\"https:\/\/ai-stack.ai\/en\/gpt-5-6-us-government-ai-gatekeeping\">GPT-5.6 and the US Government\u2019s AI Gatekeeping Strategy<\/a>.<\/p>\n<hr \/>\n<h2 id=\"the-hidden-bottlenecks-energy-and-capital\">6. The Hidden Bottlenecks: Energy and Capital<\/h2>\n<p>Behind the gleaming products, two constraints are accelerating into view.<\/p>\n<p><strong>Energy:<\/strong> Taiwan is considering new regulations requiring large power users (&gt;5 MW) to self-generate and store energy. <strong>TSMC<\/strong> \u2014 which already consumes ~9% of Taiwan\u2019s electricity and could hit 23.7% by 2030 \u2014 sits directly in the crosshairs. The tension between AI\u2019s energy appetite and regulatory pressure is becoming a global semiconductor challenge (<a href=\"https:\/\/ai-stack.ai\/en\/ai-energy-2025\">related: TSMC\u2019s position mirrors broader data-center energy constraints<\/a>).<\/p>\n<p><strong>Capital:<\/strong> <strong>BlackRock<\/strong> plans to issue <strong>over $12 billion in bonds<\/strong> \u2014 the largest single-datacenter bond issuance in history \u2014 to finance <strong>Meta\u2019s 1-gigawatt AI data center<\/strong> in El Paso, Texas (target operation: 2028), underwritten by J.P. Morgan and Morgan Stanley. AI infrastructure financing is shifting from corporate capex to leveraged capital-market instruments.<\/p>\n<p>\ud83d\udd17 Full guide: <a href=\"https:\/\/ai-stack.ai\/en\/what-is-ai-data-center\">What Is an AI Data Center? From Traditional Server Rooms to Intelligent Compute Factories<\/a>.<\/p>\n<hr \/>\n<h2 id=\"conclusion-three-structural-shifts-defining-ai-in-2026\">7. Conclusion: Three Structural Shifts Defining AI in 2026<\/h2>\n<p>WAIC 2026 wasn\u2019t just a conference \u2014 it was the industry\u2019s coming-of-age moment. Three structural shifts are now unmistakable:<\/p>\n<p><strong>1. The measure of value has shifted from \u201cmodel capability\u201d to \u201ctask completion rate.\u201d<\/strong> Enterprise buyers aren\u2019t benchmarking parameter counts anymore; they\u2019re measuring headcount saved and throughput gained. This is a fundamentally different market.<\/p>\n<p><strong>2. Robots have graduated from showpieces to production tools.<\/strong> A year ago, robots on the expo floor were pouring water and dancing. At WAIC 2026, they were running 24\/7 on real assembly lines. When AgiBot scales from 5,000 to 15,000 units in six months, and XPENG targets 1,000 units\/month \u2014 the manufacturing cost curve is about to bend hard.<\/p>\n<p><strong>3. AI infrastructure has become a national-security and capital-markets problem.<\/strong> A $12 billion bond, 29 nations signing a multilateral AI treaty, a single company consuming 9% of a country\u2019s power \u2014 these are not tech-industry metrics. They are macroeconomic and geopolitical signals.<\/p>\n<p><strong>For enterprise decision-makers:<\/strong> If your AI strategy still revolves around \u201cwhich LLM to use,\u201d you are already a generation behind. The real questions for H2 2026 are: Are you ready to rebuild workflows around Agents? Can your production lines accommodate robot colleagues? Does your energy budget account for the inference compute surge ahead?<\/p>\n<pre class=\"wp-block-verse has-text-align-center\"><div class=\"wp-block-button\"><\/div>Stay ahead. Join 7,000+ subscribers for curated global trends!<br><div class=\"wp-block-button\"><a class=\"wp-block-button__link has-text-align-center wp-element-button\" href=\"https:\/\/www.infinitix.ai\/en\/subscription\/\" target=\"_blank\" rel=\"noopener\">  Subscribe for Free  <\/a ><\/div><\/pre>\n<hr \/>\n","protected":false},"excerpt":{"rendered":"<p>WAIC 2026 set records with 1,100+ exhibitors, 300+ world premieres, and 100,000 sqm of exhibition space. This deep dive covers the AI industry&#8217;s historic pivot from model competition to real-world deployment \u2014 from AI Agents reaching productization and humanoid robots entering mass production, to the escalating chip arms race between AMD, Google, and NVIDIA.<\/p>\n","protected":false},"author":253372376,"featured_media":13818,"comment_status":"closed","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_feature_clip_id":0,"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_post_was_ever_published":false},"categories":[96987604,96987592,96987604,96987592],"tags":[96987968,96987968,96988740],"class_list":["post-13804","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news","category-featured-articles","tag-ai-agent-2","tag-humanoid-robots"],"blocksy_meta":[],"acf":[],"jetpack_featured_media_url":"https:\/\/i0.wp.com\/ai-stack.ai\/wp-content\/uploads\/2026\/07\/en-3.jpg?fit=1920%2C1080&quality=100&ct=202603031250&ssl=1","jetpack_shortlink":"https:\/\/wp.me\/ph344V-3AE","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13804","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=13804"}],"version-history":[{"count":2,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13804\/revisions"}],"predecessor-version":[{"id":13811,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/posts\/13804\/revisions\/13811"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media\/13818"}],"wp:attachment":[{"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/media?parent=13804"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/categories?post=13804"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ai-stack.ai\/en\/wp-json\/wp\/v2\/tags?post=13804"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}