In a report aimed at retail investors, Deutsche Bank pointed out that after the Moonshot AI Kimi K3 release, AI stocks experienced another "DeepSeek moment" last Friday: the market began to question anew whether the substantial AI infrastructure expenditure by U.S. tech giants, amounting to billions, is still justified, if China's open-weight models can approach cutting-edge closed-source models at a lower cost.

This is not merely a technical debate. The report mentioned that last Friday, the "Big Seven Tech" saw a decline of around 1.8%, the Philadelphia Semiconductor Index fell by about 1.6%, leading to a weekly cumulative drop of over 10%. Alphabet's Google, Microsoft, Amazon AWS, Meta, and Oracle are expected to invest around $700 billion this year in building AI capabilities, representing a growth of around 70% from last year.

The reason for investor nerves is straightforward: the AI bull market of the past few years has been built on an assumption—that stronger models require more chips, more data centers, more power, and higher capital expenditure. However, post-DeepSeek R1, Kimi K3 once again reminded the market that cheap, lightweight, downloadable, and modifiable models are catching up with closed-source behemoths.

Why did a model release hit at NVIDIA and cloud providers?

The impact triggered by DeepSeek R1 in January 2025 remains part of market memory. The report recalls that at that time, the market was concerned that a Chinese team had created a "good enough" model with older chips and lower training costs, leading to NVIDIA's market cap evaporating by around $600 billion in a single day.

The issue brought by Kimi K3 is similar: if open-weight models can reach a usable level in many applications, do enterprises still need to keep paying for the most expensive closed-source models? Do cloud providers still need to build data centers at the current pace? Will the chip demand shift from "more is better" to "cheaper, more efficient, more distributed"?

The crux here is not whether Kimi K3 has completely surpassed OpenAI or Anthropic, but rather whether it is impactful enough to shake one of the default assumptions of the AI business model in the market. As long as "good enough" models become increasingly affordable, pricing power of closed-source model manufacturers will be compressed, and the growth narratives of cloud providers and chip companies will face more scrutiny.

This is also why a model launch event ripples into the stock market. Among the Big Tech Seven, several companies serve as both providers of AI models and applications and the world's largest data center buyers; the semiconductor index directly bears the market's expectations for changes in demand for GPUs, network equipment, and AI servers.

What truly impacts valuation is not the word "open source"

Many AI models referred to by the market as "open source" are more accurately described as having "open weights."

Closed-source models are most similar to "plug-and-play" services. Flagship models such as OpenAI's GPT, Anthropic's Claude, etc., are usually controlled by developers in terms of model weights, training methods, pricing, updates, and security restrictions, with users accessing them through applications, APIs, or cloud services. Their advantages are convenience, stability, and high integration; however, the downside is that users find it challenging to grasp underlying control and must also accept the provider's pricing and rules.

Models with open weights release the trained parameters, allowing users to download, deploy, fine-tune, and even technically modify some security settings. DeepSeek R1, Meta's Llama, and several models from Mistral are closer to this category. They do not necessarily disclose complete training data, training code, and replication details, so they are not strictly equivalent to open-source systems.

This distinction is crucial for enterprise customers. Models with open weights imply greater control: companies can run models on their own servers, reducing reliance on external APIs; they can also customize models for specific industries, languages, or tasks. For institutions such as those in finance, healthcare, or government, where data cannot easily leave the local environment, on-premises deployment is more attractive.

However, "downloadable" does not mean "free." Companies still have to pay for computing power, electricity, engineering, monitoring, maintenance, and security costs. Open models may not necessarily possess the same level of general capability, product experience, and service commitment as closed-source flagship models. In other words, their impact lies not in overnight replacement of closed-source models but in providing customers with another choice.

Four Areas of Inquiry into AI Valuation

Models like Kimi K3 first impact "scarcity."

If powerful models can be replicated, modified, and hosted by more developers, AI models themselves may increasingly resemble basic software infrastructure rather than high-margin products enjoyed by a few companies. Closed-source model providers can still make money through products, data, security, and ecosystems, but relying solely on "model capability leadership" to charge a premium will become more challenging.

The second area impacted is pricing power. As more AI agents and enterprise employees use tokens, model invocation costs are shifting from small bills in pilot stages to significant items in business budgets. When AI subsidies decrease and leading model providers more explicitly charge based on tokens, cost-sensitive customers will be more willing to try open models.

The third is capital expenditure. The current investment scale in AI by U.S. tech giants is already very large. Google, Microsoft, AWS, Meta, and Oracle have announced a combined AI capability investment plan of about $700 billion this year, which is one of the most closely watched numbers in the market. If companies can use smaller, cheaper, and more specialized models to solve most tasks, investors will naturally ask: Is the spending growth on chips, networks, power, and data centers accelerating too quickly?

The fourth is the U.S. technology moat. China's continuous release of open-weight models has weakened the market's imagination of the U.S.'s AI leadership position as "unshakable." The report mentions that this year on the OpenRouter developer platform, the number of tokens processed by Chinese models has surpassed that of U.S. models. This does not mean that the U.S.'s AI advantage has disappeared, but it does mean that developer usage and ecosystem diffusion are becoming more diversified.

All these impacts point to one result: Closed-source models are still the benchmark for cutting-edge capabilities, but the narrative of "closed-source models monopolize everything" has weakened.

Open Models Are Not Only Bearish for AI; They May Also Amplify Demand

The market is most likely to interpret such events as "cheap models impacting chip demand." However, the report provides a more balanced assessment: Powerful open models will indeed challenge the business model of closed-source models and semiconductor demand expectations, but they may also benefit AI adoption, application development, and enterprise customization.

The reason is that price reductions usually lead to increased usage. Cheaper and more efficient models will allow more companies to integrate AI into customer service, offices, R&D, advertising, coding, data analysis, and internal processes. Companies that previously thought API costs were too high, data could not leave the premises, and models were uncontrollable may now start deploying AI due to open-weight models.

This is close to the logic of Jevons Paradox: after technological efficiency improves, unit costs decrease, and ultimately consumption increases. The improvement in steam engine efficiency did not reduce the demand for coal; instead, it expanded the use of steam power. AI may also experience a similar situation—individual inferences become cheaper, but the number of calls, application scenarios, and end users increases significantly.

Therefore, computational power demand may not have peaked, but the demand structure may change. A small number of the most cutting-edge closed-source models will still require the most powerful training and inference clusters; a large number of open models may be deployed on the cloud, on-premises servers, smartphones, cars, and other edge devices. Chip demand may also shift from a single pursuit of the most powerful GPU to a greater emphasis on inference efficiency, networking, power, storage, and edge computing.

For application companies, open models may actually reduce input costs. Companies can use cheaper models for vertical scenarios, software companies can develop more specific products on top of models, and cloud providers can continue to make money by hosting open models, providing inference services, and enterprise support.

Coexistence of Closed Source and Open Source, but Regulatory Responsibility Remains Unclear

The more likely scenario is not open source defeating closed source or closed source re-monopolizing, but rather both coexisting in the long term.

The closed-source model will continue to take on the roles of cutting-edge capabilities, industrial-grade quality, security assessment, and deep product integration. The open-weight model will compel closed-source vendors to control prices, improve flexibility, and shift competition from "whose model weight is stronger" to data, product design, memory capabilities, security, efficiency, and ecosystem services.

Model companies offering open weight are not necessarily acting out of charity. Developers can make money through hosted APIs, premium subscriptions, faster inference, enterprise support, fine-tuning, security services, and service level agreements; cloud service providers can sell the compute power needed to run open models; consumer platforms can embed model capabilities in recommendations, ads, and user experiences; and application companies can build features at a lower cost.

What remains truly unresolved is the boundary of responsibility. The open-weight model gives users more control but also requires users to take on more testing, security protection, cybersecurity, updates, and reliability responsibilities. Once a model is downloaded, modified, and deployed locally, determining accountability for misuse, malfunctions, or security incidents becomes more complex than in the closed-source API model.

Regulation could also alter the competitive landscape. Closed-source model vendors like Anthropic advocate for catastrophic risk testing, external evaluation, and ongoing disclosure requirements for cutting-edge AI developers; critics, however, are concerned that heavy regulation will give a competitive edge to well-funded large companies, stifling open ecosystems and small team innovation.

Therefore, the "Second DeepSeek Moment" brought by Kimi K3 is not merely a simple AI bearish proposition. It is more like a stress test: a $700 billion AI spend, closed-source model fees, U.S. tech dominance, and chip demand all need to readdress the same question—where will the money come from in the AI industry when "good enough" intelligence becomes cheaper.

Welcome to join the official BlockBeats community:

Telegram Subscription Group: https://t.me/theblockbeats

Telegram Discussion Group: https://t.me/BlockBeats_App

Official Twitter Account: https://twitter.com/BlockBeatsAsia