On the first day of WAIC2026, you’ll understand why tickets are hard to come by after visiting these booths

'Hey, do you have a ticket?' In recent days, holding a ticket to the World Artificial Intelligence Conference (WAIC) has become the hottest social currency in the tech, investment, and trend circles. This year, the total exhibition area of WAIC has exceeded 100,000 square meters for the first time, with over 1,100 companies showcasing more than 3,000 cutting-edge technological achievements, including over 300 AI products making their global debut.

On the first day of WAIC, the curtain rises. The audience, eager for excitement, floods into the exhibition hall, and indeed, it is a spectacular event—worthy of its reputation—from the 'treasure of the museum' to newly launched products, from embodied intelligence to cross-modal authentication, every booth is refreshing people's understanding of the boundaries of AI implementation.

Leap Star, the booth you want to spend money at the most

Leap Star has brought the large model native intelligent agent mobile phone STEPX Neo to the site, and the booth is surrounded by a crowd eager to experience this terminal, regarded as the next-generation 'gateway', even wanting to buy one to take home. Previously, STEPX Neo has passed the first batch of national standards for 'Intelligent Terminal Intelligence Level' L3 level testing, obtaining national benchmark certification for AI terminal intelligence levels. At the same time, with its leading technological strength, it has stood out from numerous cutting-edge AI innovations, winning the WAIC 2026 'Treasure of the Museum' honor.

At the site, STEPX Neo demonstrated new experiences of intelligent agent interaction in multiple scenarios. In travel scenarios, users only need to inquire about travel plans, and Leap Amoo can automatically plan flights and hotels through Trip.com, avoiding schedule conflicts; in local life services, Amoo can also call Alipay to jump to complete subsequent operations in the form of information cards. The on-site demonstrations covering travel, life, and other scenarios intuitively showcase the efficiency and convenience of human-machine collaboration.

As an AI innovative terminal with milestone significance in the industry, STEPX Neo is equipped with the world's first intelligent agent native operating system Step AOS (Agentic-native OS), and is built with Leap's next-generation personal intelligent agent Leap Amoo, truly bringing the capabilities of large model native intelligent agents to personal terminals, achieving deep integration and symbiosis of models, software, and hardware.

Moreover, during WAIC, Leap's booth had another surprise: six robots worked continuously for 15 hours on-site, collaboratively assembling a Great Wall made of 80,000 blocks. Based on the Force Lingji embodied general foundation model and Leap's foundational large model, each robot can autonomously perceive, make independent decisions, and dynamically divide labor, completing precision operations comparable to skilled workers at a sub-millimeter scale—this is no longer a 'demonstration', but a public roadshow of industrial-grade capabilities.

Mosi Intelligence, the unforgettable booth

Speaking into a microphone, choosing a topic, and waiting for 30 seconds, a smooth and flowing voiceover or AI podcast will resonate in your ears—so natural that it’s hard to tell it was produced by AI.

Mosi Intelligence, an innovative company focusing on next-generation contextual interaction intelligence, made its debut at WAIC, showcasing the latest advancements in real-time video understanding, complex audio transcription, voice and audio-video generation, and real-time interaction. Before the conference opened, Mosi Intelligence and the OpenMOSS team released several heavyweight models and products: the real-time visual understanding model MOSS-VL-Realtime, the one-stop AIGC creation platform Mossland, and the public beta of the MOSS API open platform, as well as the open-source multi-speaker long audio transcription model MOSS-Transcribe-Diarize-0.9B.

Yes, it’s that MOSS—the open-source large model 'pioneer' that closely followed the launch of ChatGPT. Mosi Intelligence was founded by Professor Qiu Xipeng from Fudan University and a full-time mentor at Shanghai Chuangzhi Academy, and its project 'Speech Generation Foundation Model MOSS-TTS Based on Discrete Tokens' successfully made it to the TOP30 list of the SAIL Award (Excellence in AI Leadership) at this conference.

In terms of booth interactivity, Mosi Intelligence is almost unmatched. The Mossland creation phone booth and VBTI voice personality test set up on-site became hotspots for visitors. The phone booth uses Mosi's self-developed MOSS-TTS Family voice synthesis model: visitors can express an idea like making a phone call, which can be used to voice over film clips, generate AI podcasts, or receive a 'voice freshness invoice' that records the current sound; speaking into the microphone generates a 'voice personality card' that presents speech rate, rhythm, tone, and emotion—allowing everyone to hear the 'digital version of themselves'.

Hehe Information, the most low-key practical booth

When passing by Hehe Information's booth, be sure to step inside to 'check in'. As the company that launched the globally phenomenal product 'Scan All-in-One', in the AI era, it is more needed than you might imagine—and can help you more.

Images, videos, and text have become the main carriers of current AI-generated fake content. To help users cope with the increasingly rampant and frequent forgery phenomena, Hehe Information has introduced AI cross-modal authentication technology and the 'Scan All-in-One AI Authentication Mini Program', upgrading the scope of forgery identification from compliance risk control professional scenarios to social software chats, graphic information, e-commerce transactions, short video sharing, academic research, and other general public scenarios. For example, when you receive a payment screenshot but find it hard to tell its authenticity, you only need to scan it with this mini program to determine its authenticity instantly.

And the most eye-catching feature at the booth is the 'AI Real Scene Restorer'—bringing ancient national treasures back to their former glory. This function integrates two core technologies: anti-reflection and intelligent high-definition, relying on self-developed AI scanning algorithms to accurately remove reflective bright spots, enhance and restore blurred textures, and increase detail clarity pixel by pixel. It can process images to eliminate reflection and blurriness caused by extreme scenarios with one click, and even restore the colors of faded ancient paintings to their original state. With this technological breakthrough, Hehe Information has been selected as one of the top 30 growing cultural enterprises in the country.

In fact, the practicality of anti-reflection technology far exceeds expectations: any photos or paintings taken through glass, plastic, or frames can be rejuvenated with one click. On-site staff revealed, 'We are actively seeking IP collaborations, hoping to promote AI technology to a broader audience, allowing more people to feel the warmth of technology.'

Qingtian Rental, the most ubiquitous booth

Without a fixed booth, yet active in every corner of WAIC—Qingtian Rental, as a global AI-driven RaaS (Robot as a Service) application and service platform, made its debut at WAIC with 60 robots including the Expedition A3 and Lingxi X2, simultaneously deployed at three locations: the Pudong Expo Exhibition Hall, the Expo Center, and the Zhangjiang Science Hall. This is also the first time in the industry that full-size humanoid robots have been deployed on a large scale in public areas at a major exhibition.

This group of special 'robot workers' not only supports seamless bilingual communication but also transforms into 'workplace special forces' based on the characteristics of the exhibition hall. Among them, the Expedition A3 serves as a 'guide', responsible for core forum explanations, VIP receptions, group control performances, and official news broadcasts; while the agile Lingxi X2 becomes an 'intelligent guide', providing precise navigation services for attendees.

On the first floor of the Expo Exhibition Hall, an overseas exhibitor was asking Lingxi X2 a question, and its smooth switching between Chinese and English drew the attention of surrounding spectators who stopped to take photos. Excited audience members exclaimed, 'I used to think robots were just exhibits, but this time they directly became 'colleagues', helping me explain and navigate in public areas. I really felt that 'intelligent partners' are right beside me!'

With the opening of the pop-up store in the 1F atrium of the Expo World, Qingtian Rental's new IP 'Star Qingtian (Thinking)' also made its debut. With a blue and white color scheme, star-shaped headgear, and displaying both Qingtian Rental and Booming logos on its chest, Star Qingtian is positioned as 'an AI observer from beyond the world'—a true party guest that walks into the crowd. It responds with lightweight movements and short phrases, taking on roles such as check-in reception, photo interaction, and outdoor patrol, allowing the 'human touch' of technology to quietly spread inside and outside the venue.

Tezign Technology, the most 'abstract' avant-garde booth

Enterprise-level Agentic AI company Tezign Technology showcased its theme 'Digital Forest' at this conference, presenting real cases from various industries such as consumer product development, fast-moving consumer goods growth, technology going global, and user research, focusing on the large-scale deployment of its self-developed enterprise intelligent agent GEA (Generative Enterprise Agent) in core business scenarios, attracting numerous corporate executives, industry partners, and audiences for in-depth exchanges on-site.

The booth consists of multiple independent content nodes, each corresponding to a real landing case in an industry scenario. The metal structure, screen system, and natural materials build an open distributed space, with the theme of 'Digital Forest' running throughout—AI practice is like independently growing trees, interconnected under the same framework, collectively forming a sustainable operational enterprise intelligent ecosystem.

Tezign staff introduced that GEA currently covers four product lines: insight research, content growth, design creation, and product innovation, all directed towards the same goal: transforming enterprise AI from 'one-time delivery projects' into 'sustainable operational capability systems'. So far, Tezign has served over 180 enterprise clients, including more than 60 Fortune Global 500 and China 500 companies, with over 1 million professional users on the platform, deployed and operating in over 50 countries and regions worldwide, with an average monthly token deployment exceeding 10 billion.

(Source: Shangguan News Author: Wenhui Daily, Shen Qiaosha)