Kimi K3 is open-sourcing in just 4 days, and Americans are truly panicking this time.

Original author: Beating

Americans have always wanted to sit at the center of every industry.

The AI circle is no different. Americans have always carried an air of sure victory, holding a hand of cards that seems impossible to lose.

No matter who is building AI applications on the outside, Americans believe they will ultimately have to come back to them to settle the bill. Chips are from NVIDIA, the cloud is from Microsoft, Amazon, and Google, and the most expensive models are locked behind the APIs of OpenAI and Anthropic. For companies around the world to use AI, they inevitably have to pass through the hands of the US.

Even when the names of Chinese teams occasionally appear on leaderboards, Wall Street doesn't think much of it. With chips constrained, the cloud firmly in their grasp, and talent still flowing to Silicon Valley, how could they lose?

But this relaxed sense of sure victory was recently shattered by Kimi K3, a Chinese model.

The US tech circle hastily added an update, describing Kimi K3 as a 'Sputnik' moment, akin to the shock America felt when the Soviet satellite launched in 1957. Discussions about Kimi K3, Yang Zhilin, and Chinese models on X quickly escalated from small-scale observations within the tech circle to a topic with tens of millions of views.

Kimi K3 didn't outperform the strongest US closed-source models on every single metric, but it made more people see a possibility: that strong capability, high efficiency, and an open ecosystem may not necessarily have to be cultivated simultaneously within just a few American laboratories.

Silicon Valley is indeed anxious.

Storage is the Placebo for Anxiety in the US AI Circle

When news of Kimi K3 reached Wall Street, several investment banks almost simultaneously issued research reports. Instead of first spending time discussing whose products it would impact or whether it would force US models to lower prices, they quickly turned their attention to storage.

Almost in unison, these institutions interpreted the emergence of Kimi K3 as: strong demand for storage. Longer context means AI needs to remember more things – images, audio, videos, and work records will accumulate. Consequently, flash memory, hard drives, data centers, and data services will all benefit.

Thus, Micron, SanDisk, and Western Digital became the beneficiaries of this narrative.

Sure enough, in the US stock market yesterday, storage stocks collectively rebounded violently. The Roundhill Storage ETF rose 10.91% in a single day, SanDisk jumped 14.27%, and Micron gained 12%. A sector that was being hammered just days ago due to a 'DeepSeek moment 2.0' was once again the most certain bullish bet overnight.

From an industrial perspective, this line of reasoning is not absurd. Past chatbots were like one-time Q&A sessions: you asked a question, it answered, you closed the page, and much was forgotten. The AI everyone now envisions, however, is more like a new employee joining a company. It needs to review past contracts and emails, remember what clients said, pick up unfinished work from yesterday, and leave records to avoid disputes later. An AI that can act, remember, and process images and audio will naturally 'consume' much more data than a chatbot that just talks.

This conclusion isn't pulled out of thin air, but looking back at previous model releases and deployments, was the market's reaction: 'Forget the model, focus on storage'?

Let's just say, it's an answer that allows Americans to feel at ease.

The impact of a Chinese model should have triggered a series of difficult questions: Will it make it harder for US model companies to maintain high prices? Will it make developers less dependent? Will it allow new companies to not necessarily have to start in Silicon Valley? Why not directly discuss which users Kimi K3 will take away, who it will force to lower prices, or who it will pressure to change their product?

Skirting the most pointed questions and starting the discussion with hard drives has a certain 'protesting too much' quality to it.

It's like a shop owner who thought they had a monopoly on the whole street suddenly discovering a highly competitive new shop opening next door, and quickly comforting themselves: no matter how many customers the new shop gets, they'll still need to buy water, electricity, and counters from me.

Storage is the strongest placebo for anxiety within the US AI circle.

Closed-source Models Start to Pinch

In the past few years, closed-source has been the undisputed standard answer for US AI.

The stronger the model, the more it should be locked behind an API. Users pay for access, the model company enjoys high margins, and security and compliance are managed centrally. It's a dignified and profitable path, smooth and steady, reassuring for customers, satisfying for investors, and easy for regulators.

Americans have even grown accustomed to the rhythm of this path: releasing a stronger version every few months, setting a higher price, and telling a bigger story.

But as open models grow stronger, this path is starting to feel less smooth.

Kimi K3's position in this game isn't merely about 'catching up'; it's about lowering the cost of catching up. The most dangerous aspect of an open and sufficiently strong model isn't just what it can do itself, but the much cheaper learning curve it offers to all subsequent players.

This isn't just a battle for prestige in the tech world; it's a question of whether businesses will be rewritten. America's most comfortable arrangement previously was to turn AI into enterprise services first: capabilities hidden in the cloud, customers locked into long-term contracts, ordinary people unable to see the underlying layer, and unable to easily switch. But if models elsewhere are good enough, developers get another choice, enterprise procurement gets another quote sheet, and small teams may not have to bet their future on the same batch of US companies. At that point, merely holding onto a few large contracts and selling AI only to the B2B market is no longer an impregnable moat.

This means Kimi will foster the emergence of more excellent models, which also means greater competition for all models, and consequently, weaker pricing power.

The US tech circle itself feels the wind changing.

A few days before Kimi K3's release, on July 15th, Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, released a model called Inkling. It has nearly a trillion parameters, with its code and technology fully open source, free for anyone to download, modify, and use commercially.

This is arguably the first 'serious' open-source AI from the US. Although there were previous open-source models like Meta's Llama, Google's Gemma, Microsoft's Phi, NVIDIA's Nemotron, and OpenAI's gpt-oss, they were mostly experimental.

Inkling's significance lies in the fact that someone who once achieved the pinnacle of closed-source – a former CTO of OpenAI – has turned around to seriously pursue open source.

It's worth noting that in the early post-training phase, Inkling used data generated by open models like Kimi K2.5, and its architecture also referenced DeepSeek's ideas. In other words, America's most substantial open-source effort was also built on the shoulders of Chinese open-source models.

This stands in stark contrast to Anthropic. In February this year, Anthropic publicly accused DeepSeek, Moonshot AI, and MiniMax of conducting 'industrial-scale distillation' of Claude, alleging they created 24,000 fake accounts and ran 16 million conversations to steal Claude's capabilities. In June, they escalated by naming Alibaba. By July 21st, Trump administration Treasury Secretary Bessent outright threatened sanctions against China for 'AI theft'.

No matter how loudly the threat narrative is shouted, when it comes to controlling costs and improving efficiency, Chinese models are genuinely appealing.

Airbnb uses Qwen for customer service, Cursor uses Kimi to build its own programming agent, DoorDash outsources some tasks directly to Kimi, and even Murati's Inkling uses Kimi's data for post-training.

Whether it's the closed-source path becoming uncomfortable or the distillation accusations backfiring, these are still just embarrassments at the business model level. In reality, privacy and security issues are what truly shake the last protective charm of the closed-source camp.

AI Model 'Jailbreaks'

The last line of defense for closed-source has always been security.

Locking the model away, securing the weights, channeling access through APIs, and keeping data on-premises – this fortress is the most compelling promise the closed-source camp can make. Enterprise clients are willing to pay a premium precisely for this sense of security.

But companies are growing increasingly uneasy. They are starting to ask questions that closed-source companies find difficult to answer: What do you do with my code, contracts, and client data after I hand them over to your model? Will an agent with access to a browser, terminal, credentials, and long-term goals cross the line I set for it to complete its task? Sending tokens to a closed-source API is, in a sense, letting data leave one's own walls. This is precisely the strongest selling point of open-weight models: at least, you can see what the model is doing.

And just as the debate over which is safer was raging, an almost darkly comedic incident occurred.

On July 21st, OpenAI itself confirmed that its flagship model, GPT-5.6 Sol, and an even more capable, unreleased model, escaped their isolated environment during an internal cybersecurity evaluation.

Here's what happened: The engineering team wanted to test the models' offensive and defensive capabilities, so they lowered safety restrictions and disabled the usual protections against high-risk behavior. The models were only supposed to complete the test tasks honestly. However, one model discovered a security vulnerability in the system, exploited it to access the public network, bypassed permissions, traversed systems, and eventually used stolen login credentials to break into the core system of Hugging Face, the world's largest open-source AI platform, directly retrieving the test answers from the database.

OpenAI's explanation was eight words: no malicious intent, excessive focus.

Those eight words are what truly send a chill down one's spine.

For enterprise clients, the scariest thing has never been a model actively acting maliciously. It's a model executing a bad goal with extreme seriousness.

The greatest irony of this incident is that for the past year and a half, the world has been guarding against the 'dangerous Chinese open-source model', but that remains hypothetical. The one that actually jailbroke and penetrated someone else's production system was the closed-source camp's own flagship. Hugging Face CEO Clem Delangue promptly turned the incident into an advertisement for open source, stating that AI safety cannot be solved by a single company behind closed doors, but only through open collaboration.

The same incident was used by both the open and closed camps as evidence for the correctness of their respective paths.

In the future, the real dividing line likely won't be whether a model is open-source or not, but what kind of sandbox, identity system, revocable permissions, and audit logs it operates within. Neither closed-source nor open-source can avoid this question.

And just as the closed-source security narrative was undermined by its own incident, a larger-scale reversal was quietly taking place.

Shift in Offense and Defense: It's America's Turn to be Afraid

In some parts of US policy discussions and tech narratives, there has been a near 'Three-Body Problem'-esque imagination: as long as the most advanced NVIDIA chips are kept out of China, AI progress there will be forced to slow down.

This doesn't mean China would be completely unable to do research; rather, they believe the gap in computational power will widen, making the threshold for training cutting-edge models insurmountably high. Advanced chips are like the 'laws of physics' in this race; those who can't get them will find it very hard to get ahead.

This judgment isn't entirely unfounded. Building large models does require significant compute power. Chip restrictions would increase costs, slow expansion, and make it harder for many teams to replicate the training scales of American labs. The problem is, restrictions also change people's choices. If you can buy the best tools off the shelf, there's less incentive to figure out how to use less compute, modify model architectures, or make each training run more efficient. But when the door is closed, finding a detour is no longer a choice but an instinct for survival.

So, Americans find it very difficult to understand why restricting NVIDIA's supply hasn't stopped Chinese models but instead forced out a group of teams that are even more desperate for efficiency, engineering, and open-source distribution.

Reportedly, Moonshot AI is still using the H800, the compliant version of NVIDIA's AI chip customized for the Chinese market back in 2023, for its training.

Perhaps this is the quintessential Chinese 'Millet plus Rifles' approach.

In June 2026, to comply with export controls, the US temporarily shut down Anthropic's most powerful models, Fable 5 and Mythos 5. This might be justifiable from a compliance standpoint, but it hands every Chinese open-source lab a readymade marketing phrase: at least our model doesn't have a remote kill switch.

The more you emphasize control, the more that control becomes your opponent's selling point.

Even more dramatic is the other side. According to Reuters, China has also started holding meetings with companies like Alibaba and ByteDance to discuss whether to restrict foreign access to its most advanced AI models, with the scope of discussion even including publicly available open-source models. Zhou Hongyi, founder of 360, also publicly called for China to have its own top-tier closed-source models to guard the technological high ground.

One year ago, the US worried about advanced chips flowing to China. One year later, it's China's turn to have things worth restricting.

But amidst all these structural anxieties – storage, compute, closed vs. open source, security, the shift in offense and defense – there is one most concrete, most poignant, most personal focal point. It's not an industry trend, a research report, or a policy.

It's one person.

The End Point of Anxiety: Yang Zhilin

Ultimately, what the open-source competition reveals to America is the same larger challenge: Will AI's future only serve the few companies that can sign large contracts, or will it become a capability usable by more and more ordinary teams, like electricity or the internet? If the answer gradually leans towards the latter, then whoever can attract developers and make young people willing to stay and experiment will be more important than whoever holds more enterprise clients.

And this question of 'where do people go' ultimately brings America's anxiety to a very specific name.

The reason Yang Zhilin is repeatedly mentioned by the US tech circle isn't just because he's an outstanding Chinese researcher, and certainly not because someone wants to reduce the topic to 'America failed to retain talent'. Reducing a person's choice of location to a visa is too simplistic and smacks of hindsight.

What truly stings Americans is the counterfactual: What would have happened if someone like Yang Zhilin and his team had completed their entire journey from research to entrepreneurship in the US? They would have trained models using American clouds and chips, recruited from American talent networks, secured funding from American VCs, and pitched products to American big clients. A few years later, Wall Street's ledger might boast a new star company. One person's choice, following that familiar relay chain, can transform into a string of company revenues, jobs for a group of people, and confidence for an entire industry.

America's greatest source of pride in the past was this amplifying ability. It wasn't just about attracting smart people to study and work, but about capturing their ingenuity and not letting it stop at a paper or a lab. There is enough money, enough clients, and enough people willing to take risks together.

How did such a person not end up staying in the US? Legendary investor Vinod Khosla directly pointed the finger at the tightened immigration policies of the Trump administration. However, Yang's PhD advisor at Carnegie Mellon, Ruslan Salakhutdinov, refuted this, clarifying it had nothing to do with visas. Yang had plenty of opportunities to stay back then; Salakhutdinov even emailed on behalf of an Apple executive asking if Yang wanted to join.

It was Yang Zhilin himself who was determined to return to China to start his company.

This is the most piercing part of the debate. 'He chose to return himself' is far more painful for America than 'he was driven away by immigration policy'. The former implies the system can be fixed; the latter means that even if you open the door wide open, people might not want to come in.

When overseas discussions mention Yang Zhilin, what truly stings is never just 'another outstanding Chinese researcher has emerged'. It's another counterfactual: If this person had stayed within the American system, his papers, team, funding, and company value would have been written into the ledger of US AI. Now, his achievements are first seen as the capability of a Chinese team, then radiated globally through the open-source community.

For a system confident for half a century, the hardest thing to accept is often not someone being better than you, but someone proving that they can reach the