On November 9, 2026, Alibaba Cloud quietly rewritten a billing statement: the monthly fees for the three tiers of the Token Plan personal edition remain unchanged at 39, 79, and 139 yuan, but the corresponding quotas have been raised from 11,500, 25,500, and 45,000 Credits to 20,000, 45,000, and 80,000 Credits in one go, an increase of about 70%, even the higher-end Pro tier saw an increase from 180,000 to 200,000 Credits. This is a typical action of "increase quantity without increasing price," indicating that the domestic AI large model API and subscription services are entering a naked competition stage for cost-effectiveness—front-end prices are being pressed down, while the back-end must face the hard constraints of limited supply of high-end chips. On one side, there are market rumors that the flat-headed brother is preparing for an IPO in Hong Kong, aiming to raise 2 to 3 billion dollars to enhance self-developed computing power; on the other side, according to reports from Reuters, about 600,000 advanced Nvidia chips are stuck in Hong Kong awaiting approval for export to China. Chinese Foreign Ministry spokesperson Hua Chunying gave a restrained response to the export of chips and so-called "Chinese spy" rhetoric at a regular press conference, with the dual-track path of self-research and imports being layered on by geopolitics. At a somewhat farther technological and capital frontier, Animoca Brands announced that co-founder Yat Siu will assume the role of CEO starting December 1, and Galxe will be integrated into Gravity, shrinking its front lines, as Web3 projects shift resources back to underlying infrastructure; meanwhile, Musk admitted that orders for Cybertruck far exceed Tesla's current production capacity, reflecting the bottleneck of electric vehicle production capacity and the approval stalemate of computing power chips. It is under this multi-line tension that Alibaba Cloud chose to maintain the price unchanged by increasing the Token Plan quotas, using what appears to be a simple subscription adjustment to condense the long-term game of computing power, hardware, and infrastructure in China and even globally into a seemingly "cheap" front-end decision.
Alibaba Cloud's Quantity Increase without Price Increase Ignites Model Price War
If we pull the computing power game back to the developer console, Alibaba Cloud is genuinely putting real money into this move. In the personal Token Plan, the Lite tier's monthly fee remains at 39 yuan, but the quota has been raised from 11,500 to 20,000 Credits; the Essential tier maintains 79 yuan, while the quota has been directly pushed from 25,500 to 45,000; the Standard tier is still 139 yuan, but can increase from the original 45,000 to 80,000. Even the higher Pro tier has gone up from 180,000 to 200,000 without changing the subscription structure. In terms of units, the quota increase in the first three tiers is concentrated in the range of about 74% to 78%, yet the unit price has not been raised, meaning that the same amount of money can buy nearly double the calling space—"increase quantity without increase in price" in financial terms translates to a clear disguised price drop.
This move has struck at the sensitive points of the domestic large model API subscription market. In the past, everyone was still comparing model effectiveness and ecosystem narratives; now Alibaba Cloud has pushed the front line to the most intuitive monthly subscription quota, marking the official entry into the cost-effectiveness phase of competition: within the same budget, how many requests developers can run and how large user scale they can support has become the first filter for decision-making. For individual developers, the Lite tier at 39 yuan has transitioned from barely supporting a few small project tests to being sufficient for multiple rounds of prompt tuning and parallel scheme comparisons; for small and medium enterprises, the expansion of quotas in the Essential and Standard tiers means they can conduct stress testing and gray testing in scenarios closer to real business traffic without worrying about running out of quotas within just a few days. The space for trial and error has been practically enlarged; small teams face less resistance from finance departments when initiating projects, and the threshold for moving AI applications from PPT to online experiments has been overall lowered. When the cost curve is directly lowered, the battle of models is no longer just about who is smarter, but who can make computing power more like an affordable and repeatedly testable everyday resource.
Flat-Headed Brother IPO Rumors and Chip Approval Stalemate's Dual-Track Computing Power
When Alibaba Cloud packages computing power into a "monthly fee" Token Plan at the front end, the story of chips at the back end is being pre-priced by the capital market. Market news suggests that Alibaba's flat-headed brother plans to go public in Hong Kong next year, aiming to raise about 2 to 3 billion dollars; this unconfirmed rumor is interpreted as a signal of China's bet on self-researched paths in high-end computing power: transforming long-term, capital-intensive chip development into assets that can be traded in the secondary market, giving local computing infrastructure the chance to amplify its future with capital.
Parallel to this "self-research—IPO—reinvestment" path is an import channel that has been put on hold for approval. According to sources relaying Reuters' report, about 600,000 advanced Nvidia chips are currently stored in Hong Kong, waiting for approval for export to China; while Chinese Foreign Ministry spokesperson Hua Chunying merely emphasized that China's position on the issue of chips exported to China is consistent, suggesting that specifics should be inquired from the relevant authorities, with technological supply being categorized under policy procedures. This constitutes a typical "dual-track computing power": on one end, companies like Flat-Headed Brother seek capitalization and mass production of local chips, while on the other end, they rely on existing high-end chips from Nvidia and the progress of approval. Neither path is truly unobstructed but both are indispensable. Alibaba Cloud's large models and Token Plan services are directly tied to these two tracks, with the expansion of quotas without price increase betting on the sustainable provision of computing power from both self-research and imports while being carefully constrained by the uncertainties at both ends. When computing power is simultaneously restricted by fundraising pace and approval progress, cloud vendors' price wars can only be a long-term game of repeated trade-offs within limited resources.
Chip Game and Open Boundaries Amid Diplomatic Statements
At around the same time that Alibaba Cloud announced its increase in quantity without raising prices, the issue of computing power had already extended from data centers and warehouses to the microphones at diplomatic press conferences. Chinese Foreign Ministry spokesperson Hua Chunying reiterated that China's position on the export of chips to China is consistent, and specific circumstances should be verified with the responsible authorities; the tone seems restrained, yet it equates to acknowledging that the import of high-end computing power has already been embedded into the national-level approval chain. Meanwhile, he explicitly denied the so-called "Chinese spy" rhetoric, stating it is nothing but a wild conjecture. This strong rebuttal of security allegations, together with the Foreign Ministry's response to the meeting between the leaders of China and Russia, outlines the current contours of great power relations: technology, data, and security are no longer marginal topics, but one of the main threads appearing alongside leader meetings.
Along this main thread, approximately 600,000 advanced Nvidia chips are stuck in Hong Kong awaiting approval for export to China, serving as a concrete slice of the tight supply of computing power. When the spokesperson emphasizes a "consistent" position while shifting details to the responsible departments, the outside world sees a posture that deliberately retains space: the inflow of advanced computing power is categorized under security discourse, and data and technology are viewed as resources that need boundaries, but the exact shape of those boundaries will not be provided all at once in a diplomatic setting. Amid the intertwining of approval stalemate and security narrative, domestic cloud vendors can only bear the uncertainty upstream, leaving “certainty” as much as possible for downstream. By raising the Credit limits for all tiers of the personal version of the Token Plan and maintaining the monthly fees of 39/79/139 yuan unchanged on November 9, Alibaba Cloud exemplified this: amid the uncertainties in both self-researched and imported paths, it aims to stabilize developers' and enterprises' expectations through cost-effectiveness competition and quota expansion, trying to buy time and confidence with product-level openness before the policy boundaries are fully clarified.
Animoca and Galxe Shrink Front Lines, Resource Flow Back to the Bottom
From cloud vendors stabilizing developers through "increase quantity without increase in price" to on-chain projects directly shrinking front lines, the tugging back to the bottom layer is synchronously played out in two worlds. Animoca Brands announced in November a management adjustment, with co-founder and executive chairman Yat Siu set to take over as CEO starting December 1, 2026. This is not just a simple personnel appointment but reallocating decision-making power back into the hands of those most familiar with protocols and ecosystems, leaving space for the next round of resource redeployment. Almost simultaneously, Galxe chose to integrate its team and brand into Gravity, planning to gradually shut down a series of application-layer services like Quest, ID, Passport, only retaining the normal operation of the Gravity network, while repeatedly emphasizing in official statements that this adjustment does not involve any changes to the G token contracts or parameters, attempting to stabilize the value bearers while cutting away peripheral functionalities.
After experiencing multiple rounds of market fluctuations, with both valuation and traffic adjusting simultaneously, Web3 projects are now consciously reducing non-core businesses, redirecting capital and human resources back to the underlying networks and infrastructure. This parallels traditional tech companies increasing their investment in L1 networks, core protocols, and data centers in the context of computing power shortages: peripheral products can be cut, but the truly irreplaceable foundation carrying all transactions, identities, and assets remains. For developers, the integration of Galxe means familiar tools like Quest, ID, and Passport may disappear or weaken functionally for a period; however, the positioning of Gravity as a unified chassis is clearer in the long run and conducive to building applications on a more defined protocol stack; for brand recognition, the "Galxe" brand is weakened while "Gravity" is strengthened, which will shift users' memory from specific operational activities to the network itself; for token holders, the short-term unease mainly arises from business contraction and narrative switches, rather than contract risks. This integration clearly does not touch the G token, shifting attention from “how to change the token” back to the more critical question of “whether the underlying network can truly support the next cycle after resource flow returns.”
Cybertruck Production Capacity Crisis Reflects Hardware Bottleneck
If the shrinking front lines of Web3 projects push resources back to the "underlying network," then Musk's recent comments about Cybertruck apply that same logic to the more concrete realms of steel and batteries. Musk openly acknowledged that the order volume for Cybertruck far exceeds Tesla's current production capacity, yet did not provide any numbers, leaving a clear fact: the complexity of this high-end electric pickup regarding battery specifications, body materials, and manufacturing processes determines that it cannot "upgrade online" like a software iteration. Expanding production requires new production lines, new equipment, and a resynchronized supply chain. Capital and orders can double overnight, but the delivery pace of welding robots, battery factories, and qualified components can only be slowly extended on a quarterly or even annual basis.
This lag is not exclusive to the automotive industry. According to a report relayed by Reuters, about 600,000 advanced Nvidia chips are quietly lying in Hong Kong warehouses awaiting approval for export to China, behind which are the layered realities of manufacturing equipment, wafer plant processes, and regulatory procedures. The production pressure on Cybertruck and the “waiting state” of these chips outline the same supply-side reality from different angles: globally, whether high-end chips, batteries, or key manufacturing equipment, there are physical and institutional bottlenecks that are difficult to quickly cross. In the electric vehicle track, users who have booked their cars are queuing; in the AI track, developers wanting computing power are waiting in line for GPUs. In these two seemingly booming industries, what truly decides the pace of expansion is not the news headlines of booming orders but whether the infrastructure investment cycle can support the future capacity that has been prematurely exhausted by capital and demand.
The Endgame of the Computing Power Price War and Chinese Players' Choices
From Alibaba Cloud's increase of Token Plan personal edition quotas while maintaining the prices of 39/79/139 yuan on November 9, to rumors of Flat-Headed Brother raising 2 to 3 billion dollars on an upcoming IPO in Hong Kong, to approximately 600,000 advanced Nvidia chips being stuck in Hong Kong awaiting export approval, and finally to Animoca and Galxe's resources flowing back to infrastructures such as Gravity and Musk's comments that orders for Cybertruck far exceed Tesla's current production capacity, these seemingly scattered threads actually weave a panoramic view of the computing power game: on one side, cloud vendors launch price wars with "increase quantity without increase in price" to capture developers' minds, while on the other side, the supply constraints of computing power of self-research and imports are limited by approvals, capacity, and capital cycles. In such an environment, it is difficult for China's AI and Web3 participants to treat this as a simple multiple-choice question—small and medium teams are more likely to first embrace the cost-effective cloud services, treating Alibaba Cloud's quantity-increasing without price increase offerings as the infrastructure for survival and trial and error; larger players with capital and industrial depth will treat self-researched computing power and underlying infrastructure as long-term chips, combining and betting between cloud services, domestic chips, and overseas hardware. Looking ahead, China's computing power system is likely to proceed toward a more highly autonomous and cautiously open path in the pull of policy support, self-research investment, and limited external openness: whether approvals soften, whether self-research can genuinely be implemented at scale, and how cloud computing power and infrastructure assets can be repriced by capital will determine whether this price war evolves into a persistent battle of "low-price universal access" or whether new policies and market shocks rewrite the rules. What truly needs to be continuously monitored is how the next steps of officials and the market on the aforementioned variables will reshape the boundaries of this computing power game.
Join our community, let’s discuss and become stronger together!
AiCoin exclusive Hyperliquid benefits: https://app.hyperliquid.xyz/join/AICOIN88
AiCoin exclusive Aster benefits: https://www.asterdex.com/zh-CN/referral/9C50e2
On-chain Telegram community: https://t.me/AiCoinWhaleData
On-chain community: https://www.aicoin.com/link/chat?cid=N6OVMor5g
AiCoin on-chain Twitter: https://x.com/aicoinwhaledata
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到[email protected]，本平台相关工作人员将会进行核查。
