According to SoSoValue statistics, on September 22, Eastern Time, the total net inflow of the SOL spot ETF reached $28.8692 million in a single day, with BSOL alone contributing a net inflow of $26.3814 million that day, which accounts for a very high proportion. Furthermore, its historical total net inflow has accumulated to $1.131 billion, further consolidating BSOL's leading position among similar products, with capital showing a clear characteristic of concentrated inflow in the short term. Following that, on September 23, OpenAI released two new models, GPT-6 Sol and GPT-6 Luna. According to Golden Finance reports, the API pricing for GPT-6 Luna is $0.1 per million tokens for input and $0.5 per million tokens for output, which is described as being "directly halved" compared to the previous generation. In typical calling scenarios, the cost has already fallen below that of DeepSeek V4.1 Flash, regarded as a symbolic action in the new round of price wars among large models. It is important to emphasize that the inflow of funds into the SOL spot ETF and the price reduction of the GPT-6 series are only temporally adjacent; current information only supports viewing them as two parallel developments without evidence to suggest a direct causal relationship in terms of fund flows or business cooperation.
SOL Spot ETF Net Inflow Approaches $29 Million
According to SoSoValue statistics, on September 22, Eastern Time, the total net inflow of the SOL spot ETF reached $28.8692 million in a single day, of which BSOL accounted for a net inflow of $26.3814 million, making up the vast majority of the day’s increments. Given that its historical total net inflow has reached $1.131 billion since its launch, it can be seen that this round of funding is mainly a concentrated bet on leading products, rather than being evenly distributed among multiple products. The current report does not provide the total asset scale of the SOL spot ETF or a long-term accumulated net inflow series, making it impossible to accurately rank this value in historical terms; however, for any ETF system in an expansion phase, a single-day net inflow close to $29 million is significant enough to change the short-term funding structure.
From the perspective of market sentiment and the marginal impact of institutional narratives, such a high single-day inflow can easily be interpreted as a signal that "institutional funds are entering the market," especially when funds are highly concentrated in leading products like BSOL. This will reinforce the path dependency of "leading products becoming a default allocation." However, the current data comes from a single source, and the report does not provide the performance of SOL prices or on-chain indicators for that period, making it difficult to distinguish how much of the $28.8692 million is incremental funds and how much is just the rebalancing of existing positions among products or markets. Therefore, before more samples and more complete data disclosures are available, a more prudent approach is to view this net inflow as a significant event for observing funding behavior rather than a confirmation signal that the long-term trend of SOL-related assets has undergone a structural reversal.
BSOL Attracts Over 90% of Funds in One Day
In terms of composition, this $28.8692 million net inflow was almost a "single conduit" injection. SoSoValue data shows that the day’s net inflow for BSOL was $26.3814 million, accounting for the overwhelming majority of all SOL spot ETF net inflows, translating to over 90% of the new funds concentrated in this one product, while the remaining products collectively received less than one-tenth. This extremely skewed distribution indicates that the market has almost treated BSOL as the default entry for allocating SOL in the short term, rather than diversifying allocations among multiple ETFs.
Looking at the time dimension, BSOL has achieved a historical total net inflow of $1.131 billion since its launch, which is sufficient to establish its leading position under the current sample. This also indirectly reflects the path dependency of institutions on product selection: new funds tend to follow the products with the highest historical inflows to gain relatively better secondary market liquidity and more easily evaluated operating records. The corresponding cost is that a high concentration of funds in a single ETF amplifies selection risks at the product level—whether it is rate structure adjustments, changes in redemption mechanisms, or tracking deviations, any issues arising on this “main channel” product would be magnified across the entire SOL ETF funding pool, while also creating a liquidity structure of "core products with deep liquidity, other products with shallow liquidity," making subsequent fund adjustments across products necessitate a more careful assessment of liquidity costs and the systemic exposure brought by product concentration.
GPT-6 Luna Enters at Low Price
Similar to the high concentration of funds on BSOL, the model market is also showing signs of "leading products engaging in price wars." On September 23, OpenAI launched GPT-6 Luna, and according to Golden Finance reports, its API pricing is set at $0.1 per million tokens for input and $0.5 per million tokens for output, with the same report stating that the price of the GPT-6 series has been "directly halved" compared to the previous generation. In the absence of a specific pricing list from the previous generation, "halving" is more of an official narrative about the overall price reduction, but from the input and output unit prices, it is clear that the cost composition leans significantly toward the output side: for example, 100,000 input tokens cost only about $0.01, while 200,000 output tokens cost about $0.1. The actual perceived discount in operational costs depends on the input/output token ratio configuration in application scenarios. The current report does not disclose the specific pricing and capability parameters for GPT-6 Sol, and the market’s price anchoring for the GPT-6 series remains primarily focused on the Luna tier.
From a developer's perspective, this round of "price halving" directly impacts the inference cost curve: with a nearly negligible unit price on the input side, the cost pressure for long context and complex system prompts decreases significantly, while the output side's pricing of $0.5 per million tokens determines the real spending limits for long text generation and high-frequency calling scenarios. Reports also indicate that under typical calling scenarios, the cost of using GPT-6 Luna has fallen below that of DeepSeek V4.1 Flash, meaning that among mainstream comparable products, Luna has further lowered the unit computing cost, allowing budget-sensitive teams to scale up under the same budget or test more application forms at a lower threshold. For most applications, this low price does not just mean "half the cost," but raises the overall frequency of permissible requests, user scale, and functional complexity. Once inference costs are no longer a primary constraint, the real determinants of product feasibility will depend more on the structural design of outputs and the sustainability of the business model itself.
DeepSeek's Low Price Range Under Pressure
Reports clearly state that under typical calling scenarios, the usage cost of GPT-6 Luna has fallen below that of DeepSeek V4.1 Flash, while Luna's current API pricing is set at $0.1 per million tokens for input and $0.5 per million tokens for output and has been described as "directly halved" compared to the previous generation. In the absence of detailed API pricing disclosures for DeepSeek V4.1 Flash, this comparison is still sufficient to send a signal: OpenAI is no longer satisfied with maintaining a high-end premium, but is proactively lowering the price anchor into the low-cost range originally occupied by local manufacturers and emerging players, directly engaging in "close-quarters combat" within the cost-sensitive segment occupied by DeepSeek.
From the perspective of AI manufacturers, this downward price reduction by major companies will significantly compress industry profit margins. For models that make "cheaper" their core selling point, after Luna lowers the market average price, the competitiveness of simple price comparisons is weakened, forcing manufacturers to either further dilute costs through algorithm optimizations, inference acceleration, and resource scheduling or to differentiate themselves in vertical scenarios, industry adaptations, and service integrations. In the absence of specific performance parameters of GPT-6 Sol and Luna, context windows, and without any official response or strategy adjustment information from DeepSeek, the market's judgment on technical performance will be amplified by price signals in the short term. Once the price war moves to the floor, truly sustainable participants will rely more on comprehensive efficiency rather than a single rate label.
Parallel Tracks of Funds and Computing Power
From the data, on September 22, Eastern Time, the total net inflow of the SOL spot ETF was $28.8692 million in a single day, of which BSOL alone contributed $26.3814 million. Coupled with its historical total net inflow of $1.131 billion, this indicates that funding on the crypto asset side is concentrating highly on leading products and a single track, amplifying short-term sentiment and medium-to-long-term allocation preferences on the same carrier. A day later, OpenAI released GPT-6 Sol and GPT-6 Luna, setting Luna's API pricing at $0.1 per million tokens for input and $0.5 per million tokens for output, which is described as "directly halved" compared to the previous generation. In the computing power and model service market, this reflects a passive reduction in infrastructure costs and an upgrade in the price war, affecting the cost structure and marginal expansion space of AI applications. It should be clarified that current information only indicates that these two clues are temporally adjacent, with no evidence showing a direct causal relationship between the fund flow of the SOL spot ETF and the release of the GPT-6 series in terms of funding, business, or cooperation. Therefore, investors should interpret the funding data from September 22 and the technical and pricing dynamics from September 23 as independent variables, observing the ongoing evolution of subsequent ETF subscription and redemption data, price paths, and AI service cost curves, and then assess the trend intensity and potential intersections of each track.
Join our community to discuss and grow 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support@aicoin.com，本平台相关工作人员将会进行核查。
