Wednesday, July 29, 2026 Trend Press · Cloudflare Pages

The Trend Tribune

"All the trends that are fit to read" Morning Edition Free of Charge
TODAY'S LEAD STORY

The day it was called "AI Communism"—the full significance of Kimi K3 was revealed, and the industry atmosphere changed.

On July 27th, Moonshot AI released all the weights for its massive 2.8 trillion-parameter model, "Kimi K3," for free. This article explains the sharp drop in Chinese AI company stocks immediately after the announcement, the surge in demand that led to a suspension of new registrations in just three days, the rapid growth of six times the daily revenue, and why releasing weights can be a strategy that transforms other companies' computing resources into a competitive advantage. It also touches upon the controversy that an OpenAI executive described as "complete AI communism."

The day it was called "AI Communism"—the full significance of Kimi K3 was revealed, and the industry atmosphere changed.
(Photo: illustrative)

The Day It Was Called "AI Communism"—The Full Release of Kimi K3's Weights Changed the Industry Atmosphere

On July 27th, Moonshot AI released all the weights (the model parameters themselves) of its massive model, "Kimi K3," for free and unlimited download. Anyone can download, run, and modify it on their own server. This incident has had far-reaching consequences, going beyond simply being another "China has released a new model." This article will examine how the industry is being shaken by this weight release.

An "Extraordinary" Scale of 2.8 Trillion Parameters

First, let's grasp the numbers. Kimi K3 is a Mixture-of-Experts (configuration that switches between multiple expert networks) model with 2.8 trillion parameters. Of its 896 expert networks, only 16 per token are actually active. This "sparse activation" design keeps inference costs down relative to the overall size of the model. It also boasts a massive context window of 1 million tokens, allowing it to load entire codebases and lengthy documents at once.

It is the largest open-weight model to date. According to charts presented by Moonshot itself, K3 significantly outperforms competitors such as DeepSeek (1.6 trillion), Xiaomi (1.02 trillion), and Alibaba (397 billion). While it lags slightly behind Anthropic's Fable 5 and OpenAI's GPT-5.6 Sol in benchmarks, it reportedly outperforms their older models in coding and agent-related evaluations.

Stock Market Moves Immediately After the Announcement

Before discussing the technical details, let's look at the market reaction this announcement triggered. On July 17th, the day K3 was announced, the stock prices of Chinese AI companies plummeted. Z.ai fell by as much as 30% in the Hong Kong market, MiniMax by as much as 16%, and Alibaba by 4%. In the US market, Nasdaq fell 1.4% and NVIDIA by 2.2%. What's interesting is the divided opinion within the industry regarding this reaction. Patrick Moorhead, CEO of Moor Insights and Strategy, commented to CNBC that the market reaction was "a chilling overreaction, similar to the DeepSeek shock." The volatility in stock prices itself was attributed to other factors such as corporate earnings, oil prices, and geopolitical risks, suggesting that K3 is more accurately described as "one of the triggers" rather than the "sole cause." In fact, some technology stocks recovered their value the following week.

Demand So High It Had to Suspend New Registrations in Just 3 Days

Even more impressive than the technical performance is the sheer scale of the demand. Just three days after K3's launch, Moonshot stopped accepting new subscriptions. This was because demand completely exceeded the computing resources the company had secured.

Business figures have also shown rapid growth. It has been reported that Moonshot's daily revenue grew at least sixfold after the K3 launch. Monthly annualized revenue (ARR) rose from $200 million in April to $300 million in June, and the company is reportedly considering new fundraising at a valuation of $50 billion, and an initial public offering (IPO) in Hong Kong as early as this year.

Why "Weight Disclosure" Has Such Strategic Significance

This is the technical point I want to emphasize most. Disclosing weights is not simply a "generous free provision." It is also a shrewd strategy of transforming the computing resources held by other companies into a competitive advantage for the company.

Hosting massive models on their own servers requires enormous computing resources and operational costs. However, by disclosing the weights, the costs of procuring and operating the GPUs needed to actually run them become the responsibility of the companies and individuals using them. The ability for anyone to self-host K3 on their own GPU cluster creates a situation where "anyone with an NVIDIA GPU can become a direct competitor to OpenAI and Anthropic." This scenario poses a significant threat to closed-model companies like OpenAI and Anthropic.

The powerful expression "AI communism"

Interesting in this context is a comment posted on X by Dean W. Ball, OpenAI's head of strategy. He describes a world dominated by open weight models as a "dystopian hellscape" that could lead to "full AI communism."

While this is a powerful expression, it reveals how seriously closed-model companies perceive this move towards weight disclosure as a threat. The situation where free, high-performance software can rival paid, proprietary products could shake the very foundations of their business models.

What Engineers Should Consider

From an engineer's perspective, the combination of a 1 million token context window and cost-effectiveness through sparse activation is a design that deserves genuine praise. However, as AI researcher Nathan Lambert points out, the true significance of this incident lies in the possibility that the performance gap between closed and open models, and between US-made and Chinese-made models, has narrowed from the previously discussed 6-9 month gap to around 3-5 months.

Furthermore, behind the scenes of this weight disclosure, another case involving Moonshot (illegal distillation of Anthropic models and unauthorized access to regulated chips), which I previously discussed in this column, is also unfolding. With both technological capabilities and the surrounding geopolitical tensions intertwined, the competition surrounding open weight models is likely to intensify even further.

Kimi K3Moonshot AIオープンソースAI/ML中国MoE

The highest rating in the industry was "C+"—the AI ​​safety report card reveals truly important numbers.

This article explains the "Summer 2026 AI Safety Index" released by the Future of Life Institute on July 7th. Nine companies were scored across 37 items and 6 categories, and even the industry's top performer, Anthropic, only received a C+. However, what's truly important is the reality of the "regression"—the weakening of the four major companies' previous commitments to temporarily suspending development. This article will examine the restructuring of OpenAI's safety team after the evidence collection period, the industry's shift towards military AI use, and the limitations of the index itself.

The Industry's Highest Rating Was "C+"—The Truly Important Numbers Revealed by the AI ​​Safety Report Card

One of the industry's most authoritative independent scoring systems, covering nine AI development companies, was released on July 7th. The result was a "C+" even for the highest industry rating. However, this article will focus not on the ranking itself, but on another, more significant fact pointed out in the report. We will examine the "Summer 2026 AI Safety Index" by the Future of Life Institute (FLI).

Scoring Mechanism: 37 Items, 6 Categories, 7 Independent Reviewers

First, let's introduce the framework of the survey. FLI is a non-profit organization that has been tracking AI risks for over 10 years. This year's index covers nine companies—Anthropic, OpenAI, Google DeepMind, Meta, Z.ai, Alibaba Cloud, xAI, DeepSeek, and Mistral—and scores them across 37 items in six areas: risk assessment, current harm, safety framework, existential safety, governance and accountability, and transparency.

The evaluation was conducted by a panel of seven independent experts, including David Krueger of the University of Montreal and Stuart Russell of UC Berkeley. Grades are assigned using the US GPA (Grade Point Average) system, from A to F, and the evidence collection period is until June 3, 2026. This is the fourth edition of the index, which began in 2024 with six companies, and this year marks the largest number of companies included to date.

The Numbers Themselves: Even the Highest Rating is "C+"

Let's look at the crucial results. Anthropic received the industry's highest rating of C+ (2.66 out of 4.0), followed by OpenAI and Google DeepMind, both with C ratings (2.28 and 2.01 respectively). Meta rose from 6th to 4th place but received a D+ rating. xAI, DeepSeek, and Mistral—one each from the US, China, and Europe—all received a failing grade (F). Z.ai and Alibaba Cloud received a D- rating in the middle.

The lowest-rated area was "Existential Safety," where no company could achieve a C-, with many receiving a D or lower. Even Anthropic, considered an industry leader, received a harsh evaluation in this area.

The Real Headline is "Backsliding"

This is the most important point I want to convey. Multiple reports unanimously point out that the core of this index is not the rankings themselves, but the fact that top companies are quietly retracting their past promises.

Specifically, Anthropic, OpenAI, Google DeepMind, and Meta all previously committed to unilaterally halting development if their systems approached a certain risk threshold. This review points out that these pledges have been weakened or even withdrawn. One report describes this as "moving the goalposts"—that is, the promise to stop when approaching a safety risk zone has retreated into more lenient language, conditional on the actions of competitors.

Furthermore, it's concerning that events that occurred after the evidence collection period (June 3rd deadline) are not reflected in this scoring. According to TechTimes, after this period, OpenAI integrated its safety team under its research division and eliminated its independent reporting line. This is reportedly the sixth departure of a senior safety officer from the company in two years. Therefore, it should be kept in mind that the current scoring results may already be a snapshot of a somewhat "lenient" point in time.

A New Issue: The Shift Towards Military Applications

Another issue highlighted in this review is the industry-wide shift towards military AI use. Between 2024 and 2026, Anthropic, OpenAI, Google DeepMind, and Meta are reportedly gradually shifting away from their previous policies prohibiting military applications and, alongside xAI and Mistral, are actively seeking defense-related partnerships.

While Anthropic is commended for maintaining restrictions on domestic surveillance and autonomous weapons, it has been criticized by the review panel for "questionable military involvement." Specifically, the report points to involvement in the "Minab School attack," which resulted in numerous civilian deaths. However, it's crucial to understand that this is merely the review panel's opinion on "reported relevance," and not a determination of legal responsibility.

How should we interpret this indicator?

As a researcher, several reservations are necessary when introducing this type of third-party evaluation. As the AI ​​Weekly analysis accurately points out, this indicator does not include detailed item-by-item rebuttals from each company or comparisons with evaluations conducted by AI Safety Institutes in the UK and the US on the same model. It is purely a panel judgment based on publicly available information (model cards, research papers, benchmark results) combined with responses to company questionnaires designed to bridge transparency gaps.

Nevertheless, for policymakers and companies selecting frontier AI vendors for highly confidential projects, the fact that "the best in the industry is C+" and "existential safety is at a D level across the industry" implies that published safety frameworks should be treated as mere "marketing material" until independent audits and quantitative thresholds catch up.**

Thoughts from a Researcher

Compared to the flashy announcements of new products, these low-key third-party evaluations tend to go unnoticed as news. However, I believe that this continuous accumulation of scoring, which visualizes the gap between "self-reported safety" and "independently verified safety" across the entire AI industry, is an indispensable source of information for accurately understanding the industry's reality. I will continue to closely monitor how events like the "loss of independence of the safety team," which were missed during this evidence collection period, will be reflected in the next index announcement.

AI安全性Future of Life InstituteAnthropicAI政策企業公式発表ガバナンス

NVIDIA offers its creditworthiness as "collateral"—the precarious accounting loop behind the $250 billion guarantee.

This article explains the WSJ report from July 26th regarding NVIDIA's negotiations to provide a $250 billion funding guarantee for OpenAI's 10-gigawatt data center project in Ohio. It analyzes, from an accountant's perspective, the mechanism by which NVIDIA's creditworthiness will underpin OpenAI's fundraising, given that OpenAI does not have an investment-grade credit rating; the new financial product structure in which a chip vendor guarantees a customer's real estate debt; the "cyclical" risk pointed out by Michael Burry; its relationship to the existing $30 billion investment; and its connection to US-Japan policy.

NVIDIA Offers Its Creditworthiness as Collateral—The Accounting Dangers Behind the $250 Billion Guarantee

A proposed deal reported by the Wall Street Journal on July 26th has sent ripples through the financial world. NVIDIA is reportedly in talks to provide a approximately $250 billion financial guarantee (backstop) for OpenAI's Ohio data center project. From an accountant's perspective, what's noteworthy is not just the sheer magnitude of the amount, but the structural "circularity" of this transaction. This article will carefully unravel this point.

First, let's clarify the details of the deal.

According to the report, the target is a 10-gigawatt data center campus being constructed by SoftBank's energy subsidiary, SB Energy, on land formerly used for uranium enrichment in Pike County, Ohio. One gigawatt is roughly equivalent to the annual electricity consumption of 8 million average US households, illustrating just how enormous the 10-gigawatt scale is. The total project cost, including chip costs, is said to exceed $500 billion, making it the largest single data center project announced to date.

The $250 billion guarantee currently under discussion covers the financing of the campus's lease and construction debt, but does not include the cost of the NVIDIA chips themselves that will be installed in the data center. Separately, discussions are reportedly underway for a loan of up to $350 billion for chip purchase costs. In total, this means NVIDIA could potentially be involved in a project worth $600 billion.

Why is a "guarantee" necessary? – The barrier of credit ratings

Returning to the basics of accounting and finance, it's crucial to understand why such a massive guarantee is necessary. The answer is simple: OpenAI does not yet have an investment-grade credit rating. As a company that is not yet profitable, attempting to secure long-term leases worth hundreds of billions of dollars on its own would result in extremely stringent interest rates and loan terms from lenders.

By having NVIDIA act as a guarantor, lenders can effectively base their lending decisions on "NVIDIA's creditworthiness" rather than "an individual AI campus or individual customer." This allows for more favorable interest rates and repayment periods. As AI Weekly's analysis accurately points out, this is a new type of financial product on an industry scale where "a chip vendor guarantees a customer's real estate debt," and it is being evaluated as potentially becoming "a template" for modeling similar deals in the future.

The "Circularity" Issue that Catches Accountants

This is the point I want to emphasize most. Diagramming the structure of this transaction looks like this: NVIDIA guarantees the debt of its customer (OpenAI) who purchases NVIDIA chips, using NVIDIA's own creditworthiness.

This, to borrow AI Weekly's expression, is a "troublesome circularity." It is because of NVIDIA's guarantee that OpenAI can raise funds on favorable terms, and use those funds to purchase a large quantity of NVIDIA chips. NVIDIA's sales are growing, and its creditworthiness is further strengthened—creating a self-reinforcing loop. The sarcastic comment by prominent investor Michael Burry upon hearing this news, "Around and Around We Go," seems to accurately capture the precariousness of this cyclical structure.

Such guarantees are generally treated as "contingent liabilities" in accounting. While it's uncertain whether the payment obligation will actually arise, if it does, it could represent a massive burden. How this will be disclosed in the notes to the financial statements and how it will be reflected in the risk assessment on NVIDIA's balance sheet remains to be seen, and we must await the finalization of the formal contract details.

Already a 3 Trillion Yen Investment—The Depth of the Relationship

Another point to note is that, separate from this guarantee, NVIDIA has already invested $30 billion in OpenAI. In other words, the relationship between the two companies is evolving beyond a simple "buyer and seller" dynamic, becoming an extremely close capital and business relationship involving multiple roles such as shareholder, guarantor, and largest component supplier.

OpenAI's objectives also become clear. Until now, the company has leased infrastructure from cloud providers such as Microsoft, Amazon, and Oracle, but if this deal is realized, it is reported to be the first step towards directly controlling its own infrastructure. For NVIDIA, on the other hand, it effectively secures demand for its chips for many years to come. While this appears to be a rational transaction for both parties, it should be noted that the very closeness of this relationship could raise transparency issues for external investors and regulators.

Another Layer: Government Involvement

There is another element in this deal that cannot be overlooked. The power supply for this data center is controlled by the US government, and Japan has reportedly contributed separate funds as part of a recent trade agreement. It has also been reported that Commerce Secretary Howard Lutnick is involved in deciding which companies have access to this power. In other words, this massive infrastructure project is not merely a transaction between private companies, but has a multi-layered structure that is also connected to policy agreements between Japan and the United States.

Points to Watch Practically

At the time of this report, it is still in the "discussion stage," and a formal contract has not yet been finalized. From an accountant's perspective, here are some points to keep in mind as we follow the development of this case:

• How will this guarantee be accounted for and disclosed in the financial statements of both NVIDIA and OpenAI when the formal contract is signed?

• How will the scope of the guarantee (lease and construction obligations) and the chip financing (up to $350 billion) currently under negotiation be ultimately resolved?

• To what extent does this "cyclical funding" accurately reflect the actual demand for AI infrastructure investment as a whole, or is it leading to an overestimation?

• To what extent will the relationship with policy funding contributions between Japan and the United States be disclosed in detail in the future?

In the world of AI infrastructure investment, where figures in the hundreds of billions of dollars are exchanged, the role that financial technology in the form of "guarantees" plays is likely to become even more important in the future. At the same time, we must continue to pay close attention to the risks inherent in this cyclical structure.

NVIDIAOpenAIデータセンターファイナンス循環取引投資
Advertisement300 × 250