China Mobile Launches Token-Included Phone Plan

marsbitPubblicato 2026-05-19Pubblicato ultima volta 2026-05-19

Introduzione

China's three major telecom operators—China Telecom, China Mobile, and China Unicom—have simultaneously launched nationwide "Token packages" for consumers. Starting from as low as 9.9 RMB per month for 10 million tokens, these plans allow direct payment via phone bills, positioning AI usage tokens as a new type of basic telecom commodity akin to traditional data or call packages. Each operator has adopted slightly different strategies: China Telecom offers tiered personal and enterprise plans bundled with bandwidth and security services; China Mobile integrates tokens with cloud computing packages and partners with Tencent; while China Unicom provides free trial tokens and team plans. Despite minor price variations, their core approach is unified: selling computational tokens directly to their massive combined user base of over 1.7 billion subscribers. The significance lies not in groundbreaking pricing but in the unparalleled distribution channel. By embedding token sales into existing billing systems, operators are demystifying AI access, transforming tokens from a niche developer resource into a mainstream utility. This collective move by state-backed carriers signals a strategic industry shift from "traffic management" to "computing power management," potentially reshaping the AI API market landscape and making token purchases as simple as topping up phone credit.

Breaking News! China Telecom, China Mobile, and China Unicom All Enter the Fray

From May 16th to 17th, the three major telecom operators successively launched nationwide Token plans within two days.

Starting as low as 9.9 yuan/month for 10 million Tokens.

Direct payment via phone bill is supported.

AI has evolved from a tech circle concern into a line item on your phone bill.

What does this mean—Tokens are becoming the new generation of "data packages."

Different Selling Methods, Strikingly Similar Direction

First, look at China Telecom's offering.

Individual plan: Minimum 9.9 yuan/month, includes 10 million Tokens.

Higher tiers include 29.9 yuan and 49.9 yuan, corresponding to 40 million and 80 million Tokens respectively.

For developers and enterprises: Basic version at 39.9 yuan for 15 million Tokens, Professional version at 159.9 yuan, and Flagship version at 299.9 yuan.

China Telecom's selling point is a bundled package of "Tokens + Connectivity + Security," offering broadband uplink speed boost and security protection along with Tokens.

Next, China Mobile.

Shanghai Mobile, in collaboration with Tencent, launched an AI-native workbench, with 1 yuan buying 400k Tokens. Beijing Mobile was even earlier—minimum 5.99 yuan for a computing resource sub-package, and 24.99 yuan for 10 million Tokens.

The most notable move is that China Mobile integrated Tokens into its cloud computer service.

Users without a cloud computer can directly purchase a bundled cloud computer plan with built-in OpenClaw, ready to use upon startup.

China Unicom's approach is also different.

Shanghai Unicom is giving them away—OPC customers can receive a 30 million Token testing quota for free.

Personal version: 15 yuan/month for 6 million Tokens. Team version: 198 yuan/month, also including a one-month free trial of AI cloud desktop.

The prices of the three operators are similar, with each having its own strategic focus.

But the core logic is exactly the same—selling Tokens like phone credit packages.

What's More Worth Noting Than the 9.9 Yuan Price

The 9.9 yuan Token plan itself is not cheaper than offerings from major tech giants.

Some market commentators note that compared to API pricing from major model providers, the operators do not have a clear price advantage.

But the key point here is not the price; it's the distribution channel.

Major tech companies sell Tokens to developers and power users—there's a ceiling there.

Telecom operators sell Tokens to everyone who has a mobile phone.

China Mobile has nearly 1 billion subscribers, China Telecom has nearly 400 million, and China Unicom also has over 300 million.

This user pool is on a completely different scale.

And most importantly, this represents the collective entry of state-owned enterprises.

The shift in the telecom industry from traffic management to computing power management is not the decision of a single company.

It is a synchronous pivot by the three major operators.

The data from the National Data Administration is also on the table—the daily Token consumption volume in China surged from 100 billion in early 2024 to over 140 trillion by March 2026.

An increase of over a thousand times in two years.

Tokens are no longer a niche unit in the AI circle; they are becoming a basic telecom commodity alongside data and call credit.

Some telecommunications industry experts judge that if all three operators follow through,

the distribution market for large model APIs might partially shift from cloud providers to the operator system.

In the future, you won't need to understand what an API or a model is to buy Tokens.

You could just go to a service center and say, "Top up 100 yuan worth of Tokens for me," and another line would appear on your phone bill.

Domande pertinenti

QWhat is the core new product that China's three major telecom operators (China Telecom, China Mobile, and China Unicom) have recently launched?

AThey have collectively launched national "Token packages" or "Token plans." These are data-like packages where users pay a monthly fee to receive a certain amount of AI model processing tokens.

QWhat is the lowest monthly price point mentioned for these Token packages, and how many tokens does it offer?

AThe lowest monthly price mentioned is 9.9 yuan per month, which offers 10 million tokens. The article cites this as China Telecom's entry-level personal plan.

QWhy is the launch of these Token packages by telecom operators considered significant beyond just pricing?

AThe significance lies in the channel and scale. Unlike AI companies selling primarily to developers, operators can sell Token packages to their entire mobile user base—totaling hundreds of millions of customers—making AI consumption as easy and common as paying a phone bill.

QHow is China Mobile specifically integrating its Token offering?

AChina Mobile is integrating Tokens into its cloud computer service. For users without a cloud computer, they can purchase a fused package that includes a cloud computer pre-installed with 'OpenClaw' AI tools, ready to use upon startup.

QWhat broader industry shift does the article suggest this move by the telecom operators represents?

AIt represents a strategic shift for the telecom industry from 'traffic operation' (focusing on data/bandwidth) to 'computing power operation' (focusing on AI processing tokens and services). This is a coordinated move by the state-backed operators, signaling Tokens are becoming a fundamental telecom commodity.

Letture associate

Jensen Huang: Vera Rubin Full Mass Production, AI Agent a Key Focus, Challenging Intel to Target the Next-Generation AI PC Gateway

NVIDIA CEO Jensen Huang delivered the keynote speech at GTC Taipei 2026, announcing several major product launches and strategic directions. The company's Vera Rubin architecture is now in full-scale production, with OpenAI, Anthropic, and SpaceX among the first customers. NVIDIA highlighted AI Agent as a key future focus, introducing the Vera CPU designed for AI agents and the Vera BlueField-4 STX for secure, chip-level AI storage processing. A significant move involves challenging Intel in the PC market. NVIDIA, in collaboration with MediaTek, is developing the RTX SPARK PC chip (manufactured by TSMC) for Windows systems, set to launch this fall for laptops and desktops. This signals NVIDIA's push into the next-generation AI PC arena, aiming to provide a vertically integrated core computing platform for the entire Windows ecosystem, similar to Apple's approach. Other announcements include the new Nemotron 3 Ultra AI model and the NVIDIA DSX platform, described as a complete "playbook" for building AI factories, allowing performance simulation and validation before physical deployment. In automotive, the DRIVE Hyperion platform was positioned as a global robotaxi platform, with major Chinese automakers like BYD, Geely, Zeekr, Xiaomi, and Pony.ai already adopting or developing autonomous driving solutions based on it. The Alpamayo 2 super open inference model for robotaxis was also introduced. For robotics, NVIDIA unveiled the Isaac GR00T humanoid robot reference platform for academic research and a large open-source agent tools and skills suite for Physical AI. The company plans to collaborate with global humanoid robot manufacturers, including China's Unitree, whose H2 Plus robot served as the reference hardware for the GR00T platform demonstration.

marsbit5 min fa

Jensen Huang: Vera Rubin Full Mass Production, AI Agent a Key Focus, Challenging Intel to Target the Next-Generation AI PC Gateway

marsbit5 min fa

Running MoE on Mobile Phones? Meta Proposes MobileMoE, Speeding Up iPhone 16 Pro by 3.8x

Meta's MobileMoE, a mobile-optimized Mixture-of-Experts (MoE) language model architecture, enables efficient on-device large language model (LLM) inference for the first time on commercial smartphones. Designed for decoder-only Transformers, it replaces dense feed-forward layers with MoE layers. Key design choices include 8 experts with granularity g=8, top-4 routing, and a shared expert. The model undergoes a four-stage training process: pre-training, intermediate training, supervised fine-tuning, and quantization-aware training. Results show MobileMoE models, with similar memory footprint, achieve equal or higher average accuracy across 14 foundational benchmarks while using only 1/2 to 1/4 of the FLOPs compared to dense baselines. After INT4 quantization, they remain competitive. Notably, on an iPhone 16 Pro, MobileMoE-S demonstrates significant speedups: up to 3.8x faster in the prompt phase and 2.2-3.4x faster in per-token generation compared to a dense counterpart, with lower peak memory usage. While MobileMoE establishes a new Pareto frontier for on-device LLMs in accuracy-compute trade-offs, particularly excelling in code and math tasks, it currently lags behind models like Qwen3.5 2B in advanced instruction following and knowledge reasoning. Future work includes improving post-training techniques, exploring NPU deployment, and managing the runtime memory sensitivity of MoE models to varying inputs.

marsbit9 min fa

Running MoE on Mobile Phones? Meta Proposes MobileMoE, Speeding Up iPhone 16 Pro by 3.8x

marsbit9 min fa

Bitcoin's Weak Rebound Fails to Mask Adjustment Trend, HYPE's Top Signal Warns of Short-Term Risks | Invited Analysis

**Title:** Bitcoin's Weak Rebound Fails to Mask Downtrend; HYPE Top Signal Alerts of Short-Term Risks | Exclusive Analysis **Abstract:** This weekly market analysis examines the current technical structures of Bitcoin and HYPE, outlining key trading strategies. Bitcoin's daily chart shows it has broken below the median line of its primary ascending channel, indicating structural weakness. It is currently experiencing a weak rebound within a short-term descending channel, targeting resistance at $75,000-$76,000. Failure to break above this zone could lead to a resumption of the downtrend, testing support at $69,500-$70,500. Trading strategies include positioning for a rebound rejection (Plan A) or a breakdown below key support (Plan B) with controlled short positions. For HYPE, the 4-hour chart reveals a potential seven-wave advance from the May 14 low, now showing signs of exhaustion. A bearish divergence (momentum weakening) has been observed, coupled with a top signal from the proprietary "Spread Trading Model" at potential endpoint 47. The key this week is to monitor if a confirmed top forms here, especially upon a breach of the $62.5-$64.57 support area. If broken, a larger corrective move towards $54-$56.30 is anticipated. The short-term strategy for HYPE focuses on cautious long entries only upon confirmed stabilization within the support zone. The report also details a successful short BTC trade from the previous week, yielding a ~5.07% profit, executed based on model signals and price action. Strict risk management rules, including dynamic stop-loss adjustments, are emphasized.

marsbit26 min fa

Bitcoin's Weak Rebound Fails to Mask Adjustment Trend, HYPE's Top Signal Warns of Short-Term Risks | Invited Analysis

marsbit26 min fa

Trading

Spot
Futures
活动图片