AI Compute
Data centers · chips · modelsEvery story across China’s AI compute stack — infrastructure, supply chain, foundation models, and the companies behind them.
Report: ByteDance Plans to Raise AI Infrastructure Spending by 25%
ByteDance is preparing to spend more than 200 billion yuan, sharply lifting its AI budget as it doubles down on domestic chips and global infrastructure. Its spending still trails that of U.S. giants, but China’s AI sector is breaking through under a distinctive collaborative ecosystem, and its pace of catch-up is increasingly worrying experts in Silicon Valley.
ZTE Published a Paper Mapping the Frontier Directions for AI
Synced Editorial Team
ByteDance’s AI Ledger: Doubao Makes Less Than 1 Million Yuan a Day, While Seedance Runs at 70% Gross Margin
How do you get users to pay for services again in a market trained on free products?
Why Ascend 950PR Could Be the Core Investment Theme in the Second Half of 2026
Huawei’s Ascend 950PR, slated for release in the second half of 2026, marks a critical shift for domestic AI computing power from “usable” to “good enough to deploy at scale.
The World’s Largest Data Center Falls Through
This article is from the WeChat public account Fenghuang Finance, by Fengbaoyan.
ByteDance Lights Up Its Own ChatGPT Moment During Chinese New Year
The real AI offensive is a hard decision to turn itself into a technology company.
How Did AI Become a Business About Steel, Power and Land?
This article is from the WeChat public account Wang Zhiyuan, by Wang Zhiyuan. Original title: “AI Faith Has Turned Into FOMO, and Capital’s Patience Is Now on the Clock.” Header image: AI-generated.
Tokens Go Global, Selling China’s Electricity to the World
This article is from the WeChat public account Deep Tide TechFlow, written by Black Lobster. The header image is from Visual China Group.
Moonshot AI Denies Kimi K3 Is a Distilled Copy as Viral Model Demand Surges
Moonshot AI says Kimi K3 is based on original architectural work, not distillation, after the model’s strong benchmark results and rapid user demand drew praise from Elon Musk and scrutiny from rivals. The company also signaled that full weights for its open-source path are planned for release before July 27, 2026.
DeepSeek’s One Line Sent Domestic Chip Stocks Soaring. What Exactly Is UE8M0 FP8?
After DeepSeek V3.1 was released, one official comment set China’s AI community buzzing:
Chinese GPU Players Convene an Open-Source Gathering, Bringing in Core Developers From SGLang and Beyond
There were no executives from major tech companies on stage, but the room was packed with familiar faces from the open-source community.
After Years Away, AI Chips Are Again the Star of a Huawei Launch
Huawei’s chip cadence has moved onto a new track across the board.
Zero GPUs, World’s No. 1 Supercomputer: Shenzhen Manufacturing Rewrites the Rules
The World’s Most Powerful Supercomputer Has a New Owner
40,000 Agents Working Together, Multi-Model Teams: Agent Infrastructure Starts to Be Rebuilt
AI agents now have dedicated computing power.
Chinese 1,000-GPU Platform for 100B-Scale Model Training Goes Live, With 91% Linear Speedup From 70B to 130B Models
China’s first large-scale computing power cluster built on domestic full-function GPUs has officially gone live.
DeepSeek Starts Hiring Civil Engineers for Its Own GW-Scale Data Centers
You will have the opportunity to take part in planning and building infrastructure from the MW (megawatt) to GW (gigawatt) scale.
Who Can Solve AI’s Computing Power Hunger and Energy Trap? Two Post-95 Founders Are Building a New Paradigm With Phase-Change Optical Computing
Synced Editorial Team
How Can SRAM Help Solve the GPU Shortage as Inference Demand Surges?
This article is drawn from PRO member newsletter content. Follow “Synced PRO Membership” at the end of the article for more in-depth topic briefings.
ByteDance’s AI Push Is Racing Through Anxiety
If Doubao keeps up the user growth it has seen over the past 18 months, computing power costs will put ByteDance under enormous pressure.
Agents Need Scale and Brains: Inspur Information Packs 40,000 Agents Into One Rack While Letting Large Models Team Up on Answers
The core job of AI infrastructure has shifted from supporting large-model inference to supporting the scaled operation of massive numbers of agents and the continuous production of high-quality tokens.
Converging Intelligence and Compute: A Deep Integration Architecture for AI and Crypto
Artificial intelligence is at a critical point in its shift from monopoly by centralized giants toward decentralized, transparent “open intelligence.” The crypto industry, meanwhile, urgently needs AI to address complex on-chain interactions, fragile security, and limited application utility after years of infrastructure buildout.
Doubao’s 500-yuan-a-month Pro Plan Is Officially Here. Is It Expensive?
Can an Agent-Driven Office Workflow Really Replace Employees?
Decentralized AI in 2026: Five Core AI Crypto Infrastructure Projects
This article systematically reviews the five core projects leading decentralized AI in 2026, based on real-world adoption.
Cambridge AI Supercomputer Crashes in 37.7°C Heat, Halting 350 Research Projects
Recent reports about Europeans rushing to buy Chinese air conditioners have been everywhere. But here is another possibility:
ByteDance Goes All In on B2B AI, With Doubao’s Paid Tier Only the First Step
ByteDance’s AI Pivot
Breaking Down Decentralized AI: How Incentive Mechanisms Coordinate Machine Intelligence
Imagine this scenario: thousands of AI models are generating results across a network at the same time, but no single company is there to tell you which outputs are “better.” Who judges them? Who allocates the rewards? And who is responsible when something goes wrong? This is the core challenge facing decentralized AI. Unlike centralized AI, it does not rely on an authority to endorse outcomes. Instead, it uses incentive mechanisms and evaluation rules to let machine intelligence operate on its own in an open environment. Intelligence is no longer an asset that is “owned...
China’s First 100,000-Card Cluster Is Complete: Domestic Computing Power Enters the 100,000-Card Era
China’s AI computing power has hit the accelerator and entered the 100,000-card era.
ZTE Teams Up With Domestic Computing Power Players to Win WAIC’s SAIL Star Award
This year’s WAIC SAIL Star award once again went to a supernode, and once again to ZTE.
The Specter of Computing Power: AI Hegemony, Physical Limits, and the Inevitability of Decentralization
Computing power is showing an unprecedented tendency toward gatekept concentration.
The First Year of Capitalization for Chinese AI Computing Power Chips: Vimicro Takes the Lead, Enflame Clears Review, and xFusion Is Accepted
Macro Introduction: Three Forces Shaping China’s AI Computing Power Chip Industry
In 2026, AI Investing Will Need to Beat Expectations
In 2026, AI Investing Will Need to Beat Expectations. Technology became the main theme in A-shares in 2025, with the AI sector standing out and AI computing power among the top gainers. AI computing power should remain a relatively certain theme in 2026, with related companies potentially delivering earnings growth and subsegments such as liquid cooling creating investment opportunities.
OpenAI and Anthropic Went Head to Head Last Night; Expert Says Agents Will Bloom Across Industries in 2026
If one word captures the AI industry in 2025, it is “threshold.
NVIDIA Reports Fiscal Q1 2026 Results, With H20 Restrictions Driving $8 Billion Hit
NVIDIA has released its results for the first quarter of fiscal 2026, showing steady performance. But the company expects roughly $8 billion in lost H20 revenue due to export control restrictions, a hit that may show up in the next quarterly report.
A Look at the 2026 Top 30 AI Chip Companies: These Six Are Already in the IPO Sprint
A Look at the 2026 Top 30 AI Chip Companies: These Six Are Already in the IPO Sprint
In 2026, AI Investing Will Need to Beat Expectations
Produced by | Miaotou App Author | Dong Bizheng Editor | Guan Xuejing Lead image | AI-generated Looking back at 2025, technology became the main theme in A-shares. Within tech, the standout was AI. AI computing power names such as “Yi Zhong Tian” (Eoptolink, Zhongji Innolight, TFC Communication) and “Ji Lian Hai” (Cambricon, Foxconn Industrial Internet, Hygon Information) carried much of the market’s technology trade. AI applications and AI edge devices, by contrast, lagged AI computing power. Wind code Securities short name Gain/loss since 2025 (%) 8841678.WI AI Computing Power Index 46.67% 8
2026 China AI Chip Industry Watch: A Latest Analysis of AI Computing Power Chip Companies Planning Listings
In 2026, the "Xingyuan Agent," built on "Starlight Smart No. 5," was officially released. Based on a self-developed multi-core heterogeneous XPU processor architecture, the agent offers full domestic controllability, high security, and broad compatibility. It can work with mainstream open-source large models, support standalone operation or cluster expansion, and quickly build industry-specific AI data center systems. ...
ByteDance Raises 2026 Capex to More Than $30 Billion, Deepening AI Push and Shifting Toward Domestic Chips
C114 News: ByteDance has raised its 2026 capital expenditure plan by 25% to more than $30 billion, driven by AI investment and rising memory costs. The company plans to spend about $14 billion on NVIDIA GPUs and is increasing its budget for domestic chip purchases.
2026 Semiconductor M&A Wave: 250 Deals Reshape the Supply Chain
2026 Semiconductor M&A Wave: 250 Deals Reshape the Supply Chain as Domestic Substitution Enters an Accelerated Consolidation Phase, production lines, supply chain, M&A wave, semiconductor devices
China's Optical Chips Enter a Golden Window to Move Upmarket
China’s optical chip sector has been gaining momentum recently, with supportive signals from policy, industry and capital markets. In June 2026, the Ministry of Industry and Information Technology formally issued the Implementation Opinions on Innovative Development of “AI + Information and Communications,” calling for stronger R&D in core technologies including high-end optoelectronic chips, CPO and all-optical switching. Overseas giants such as NVIDIA and Meta have committed billions of dollars to secure optical interconnect capacity. Falling prices and rising shipments of 800G optical modules, together with pilot production of 1.6T modules, are helping meet the bandwidth demands of AI computing power networks...
2026 China AI Chip Inference Acceleration Test: An Analysis of Inference-Specific, Full-Stack Self-Developed, and General-Purpose Technology Paths
Introduction