On May 13, 2026, several seemingly unrelated news stories aligned on the same timeline: In Shenzhen, at Tencent's shareholder meeting, Ma Huateng used the metaphor “a year ago thought we had already boarded the AI ship, later found that the ship was leaking; now feel like just stepping on the ship, can't sit down yet, hope the ship can go a bit faster” to describe the company's AI progress, admitting that early foundational capabilities were not outstanding, and that it could only hurriedly repair the ship through filling gaps in talent, management, and training; in Hangzhou, Alibaba continued the momentum after launching Qwen3.6-Plus in March, publicly disclosing that the large model HappyOyster and video model HappyHorse were still in the grayscale testing phase, presenting a posture more like that of a follower rather than a decision-maker. Across the Pacific, the White House press officer confirmed that NVIDIA CEO Jensen Huang was “last minute” added to Trump's visit to China list, and would travel to Beijing on Air Force One alongside Elon Musk, with a delegation of more than a dozen CEOs being repainted with the colors of computing power and chips; meanwhile, in the on-chain Hyperliquid HIP-3 market, data from a single source shows that Micron Technology (MU) contracts had an open interest of approximately $145 million, surpassing gold's approximately $130 million, directly squeezing into the top five OIs, ranking only behind SP500, XYZ100, BRENT OIL, and CL, and currently, there is no public information explaining the specific drivers behind this surge in positions. On that day, all these scenes overlapped like a shutter pressed simultaneously, framing the narrative of AI hardware amidst China's tech self-critique, the showcase of American diplomacy, and speculative activities in cryptocurrency derivatives.

Ma Huateng said the ship has just set sail: Tencent's AI makeup lessons

On the same day in Shenzhen, Ma Huateng used a series of “boarding the ship” metaphors to put Tencent's position on AI in the spotlight. He stated that a year ago he thought they had already boarded the AI ship, later finding that the ship was leaking, and that now it felt like just standing on the ship, unable to sit down, only hoping the ship can go a bit faster. Facing a room full of shareholders, he directly admitted that Tencent's early foundational capabilities in AI were not prominent, and that this ship did not belong to Tencent by nature, but rather they had hurriedly boarded it due to external waves pushing them along. This expression serves not only as a response to investor inquiries but also as an indirect reflection on the internal path dependency over many years—past reliance on social, content, and gaming to build their moat does not automatically translate into leading advantages in a new round of AI infrastructure competition.

He immediately provided the path for “makeup lessons”: over the past two years, the company relied on bringing in talent, adjusting management structures, and systematically training internal teams to gradually strengthen the originally weak AI foundation. Ma Huateng summarized Tencent's current state as “gradually entering the AI development stage,” deliberately avoiding any terms like “leading” or “completed transformation,” instead emphasizing that they are still in the process of catching up. This public acknowledgment of “the ship used to leak, and only now just standing firm” is itself a response to long-standing external doubts about Tencent's slow AI progress: it is not about lack of desire but rather starting late with a poor foundation, and having to earnestly work from the ground up. On the same May 13, another giant frequently appeared with large models, but Ma Huateng chose to clarify the story of “makeup lessons” at the shareholder meeting instead of “overtaking.” This difference in mentality reflects the slow turn of Chinese large platforms in the global AI competition from flashy narratives to engineering realities, from claiming to lead to acknowledging shortcomings and then seeking to catch up.

Alibaba releases large models: from Qwen3.6 to grayscale video

If Tencent was still repeatedly emphasizing “makeup lessons” at its shareholder meeting, Alibaba was providing a series of product rhythm announcements. In March 2026, Alibaba released the Qwen3.6-Plus large language model, described externally as a new version with significantly improved performance, meaning it chose to respond to external doubts through high-frequency iterations on the backbone of large models: not telling a story of an “ultimate large model,” but rather, through the continuous advancement of version numbers, proving that it has at least entered a normal engineering phase. By May 13, Alibaba further disclosed that the advanced text model HappyOyster and video model HappyHorse had entered grayscale testing, with text and video advancing in parallel, pushing these two mainstream AI narrative entry points together into the experimental field.

The definition of grayscale testing itself carries a cautious yet aggressive flavor: limited range, limited users, limited scenarios, indicating that Alibaba does not dare to easily announce “full launch,” yet unwilling to keep the models locked in the laboratory until perfect before unveiling. In this sense, it stands in stark contrast to what Ma Huateng said on the same day about “just standing on the ship, can't sit down yet”—Tencent chose to acknowledge its weak foundation and slowly make up for it, while Alibaba tends to push half-finished products into controlled environments to observe reactions. Together, these two paths form the same picture of the era: Chinese platform companies in the AI race are simultaneously chasing global leaders with iterative versions while experimenting with product forms in grayscale, exploring how to progress in the intersection of trial and following.

Jensen Huang boards for a diplomatic flight: AI chips take center stage in diplomacy

On the same day, on another flight route outside Beijing, the story was elevated to a higher level. A White House press officer confirmed to the media that NVIDIA CEO Jensen Huang would visit China with U.S. President Trump and would board Air Force One to Beijing alongside Tesla CEO Elon Musk—on this plane, about “more than a dozen” CEO-level executives were named. Bloomberg's report particularly emphasized that Jensen Huang was added to the delegation at the “last minute,” and this “last-minute boarding” arrangement itself sent signals to both sides: the topics of AI computing power and chips are no longer just technical chapters in industrial policy, but elevated to a strategic priority that can influence seat arrangements.

From the composition of the lineup, this is no longer a traditional economic and trade visit, but rather a direct transfer of the competitive relationship between the global AI hardware and electric vehicle industries to the highest level of diplomatic stage: on one side is Huang, who holds the discourse power of GPU computing, and on the other is Musk, who bets on autonomous driving and electrification. Their co-presence with heads of state means that the “chip—computing power—whole vehicle” industrial chain is bundled into the symbolic narrative of this visit. The specific meeting agenda and potential technology agreements have yet to be made public, but just from this last-minute candidate adjustment, it is clear that in the new round of Sino-U.S. competition and cooperation, AI chips are no longer a backdrop but are pushed to the forefront as protagonists.

MU contracts surpass gold: crypto traders bet on chips

On the same day, AI chips were pushed to the center of the Sino-U.S. negotiation table, and in the on-chain derivatives market, they quietly took on a new form on the stage. According to data from a single source, on the Hyperliquid platform, Micron Technology (MU) contracts had an open interest of approximately $145 million, while gold contracts had around $130 million. For the first time, MU surpassed gold in this primarily crypto-native venue, squeezing into the top five OIs in the HIP-3 market, ranking only behind SP500, XYZ100, BRENT OIL, and CL. For a traditional tech stock, this is not just a simple ranking change, but rather reclassified by the entire market as a new type of chip that can be traded around the clock like Bitcoin: the narrative of AI storage chips is compressed into a string of contract codes, becoming a risk exposure that crypto traders are willing to hold for long periods.

Unlike offline capital markets, it is hard to reverse engineer the emotional context here through roadshows, announcements, or policy trends. Research briefs clearly state that there is currently no public information to explain why MU contract positions surged in a short time, nor whether it was due to a concentrated build-up from a single large account or dispersed follow-ups from a batch of small and medium accounts; these details are buried in anonymous addresses and matching engines. The only relatively certain thing is that crypto-native traders have begun treating tech stocks related to the AI industrial chain as a kind of “digitally tradable asset,” but in the absence of more specific driving information, hastily interpreting the surge in MU OI as a certain industrial signal is more a projection of emotion rather than a verified rule.

From the shareholder meeting to Air Force One: Next chapter in the AI race

Rewind the timeline back to May 13, 2026: on one side is Tencent's shareholder meeting, where Ma Huateng used the metaphors “the ship is leaking” and “just stepped on the ship” to acknowledge that early foundational capabilities were not outstanding, and could only catch up by making up for lessons; on the other side, Alibaba continued to disclose that HappyOyster and HappyHorse were still in grayscale testing following Qwen3.6-Plus, attempting to tie itself to the latest developments with faster iterations of large models and video models. Meanwhile, the White House confirmed that Huang had joined the visiting delegation at "the last minute," board Air Force One alongside Musk, placing AI chips, computing power, and electric vehicle industry chains at the forefront of Sino-U.S. diplomacy; within the same time frame, Hyperliquid reported that the open interest for MU contracts surpassed that of gold, squeezing into the top five of the HIP-3 market, with crypto derivative funds structurally shifting between traditional safe haven assets and tech stocks. In this montage of scenes, one can see that while Chinese platform giants are still trying to “catch up” with large and video models, it is still American companies, represented by Jensen Huang, that hold the steering wheel of AI computing power and chips, with capital markets betting in advance on each link through derivatives like MU. What remains to be seen is whether China's AI products can transition from grayscale testing to large-scale implementation, whether Sino-U.S. competition and cooperation in AI chips and computing power could find temporary relief through similar visit scenarios, and whether the rotation of crypto derivative funds between gold and tech stocks will continue to strengthen the long-term narrative around AI hardware.

Join our community, let's discuss and grow stronger together!
Official Telegram community: https://t.me/aicoincn
AiCoin Chinese Twitter: https://x.com/AiCoinzh
OKX benefits group: https://aicoin.com/link/chat?cid=l61eM4owQ
Binance benefits group: https://aicoin.com/link/chat?cid=ynr7d1P6Z

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support@aicoin.com，本平台相关工作人员将会进行核查。