News linked to both this project and an event.
Odaily News: Blockchain data analytics platform Nansen released its "Starknet H1 2026 Report," stating that in the first half of 2026, Starknet completed its strategic transformation from a high-performance Layer 2 network to a "privacy-preserving execution layer." The launch of the STRK20 privacy framework and the Bitcoin asset strkBTC became the ecosystem's most significant upgrades.The report notes that as an Ethereum-based ZK-Rollup network, Starknet generates STARK proofs off-chain and verifies them on-chain, achieving high throughput and low transaction costs. Its smart contracts use the Cairo language, specifically designed for verifiable computation, and support native account abstraction functionality.In the first half of this year, Starknet launched the v0.14.2 upgrade, introducing the SNIP-36 protocol to lay the technical foundation for private transactions. This upgrade allows the network to directly verify off-chain execution proofs, enabling confidential state transitions without exposing account balances or counterparty information. Additionally, SNIP-37 adjusted the network's economic model, increasing storage costs while lowering base gas fees to optimize incentives for long-term state growth.In terms of privacy applications, Starknet launched the STRK20 privacy framework, which allows users to convert any ERC-20 asset into encrypted balances and conduct private transfers, trades, and DeFi interactions. The system is built on zero-knowledge proof technology and implements compliant auditing through an encrypted viewing key mechanism, protecting user privacy while supporting targeted information disclosure in regulatory scenarios.On-chain data shows that in Q2 2026, Starknet recorded an average daily transaction volume of approximately 239,000 and an average of about 50,000 daily active addresses. During the period, a total of 22.5 million transactions were completed, involving approximately 71,000 users. Among these, DEX aggregator AVNU contributed roughly 14 million transactions, accounting for 62.2% of total volume; gaming infrastructure Cartridge contributed 6.57 million transactions, with the two combined accounting for approximately 91% of transaction activity.
Odaily News Alex Svanevik, CEO of on-chain analytics platform Nansen, stated that when enterprises begin effectively using Chinese large language models, the bubble in the AI industry may burst. While the U.S. regulatory environment could limit this process, the overall trend remains that Chinese models are continuously becoming more efficient, capable of running on non-cutting-edge hardware, while global GPU supply (including non-Nvidia chips) is increasing.Alex Svanevik also pointed out that the recent decline in H100 and H200 GPU rental prices reflects a shift in the supply-demand structure of computing power. He raised the question of how to interpret the market signal of declining GPU rental prices. As model efficiency improves alongside expanding computing power supply, the AI infrastructure market may be entering a phase of repricing.
Alex Svanevik, CEO of on-chain data analytics platform Nansen, posted on X stating that from a hindsight perspective, if Binance founder CZ had completed the acquisition of FTX back then, his potential asset structure would have changed significantly. He would now likely hold an indirect exposure of approximately 8% stake in Anthropic, about 5% stake in the AI coding tool Cursor, as well as some investment interests related to SpaceX.It is reported that in November 2022, CZ had disclosed his intention to acquire FTX but later abandoned the plan after due diligence uncovered issues beyond his control. Subsequently, FTX filed for bankruptcy protection.
Nansen CEO Alex Svanevik stated that open-weights models may pose greater competitive pressure on Anthropic and OpenAI in the future, because not all tasks require a "150 IQ level" frontier model. In many scenarios, a model with around 115 IQ but costing about 90% less is "fully sufficient," offering a clear cost-performance advantage. Since the AI industry has generally believed that profits would come from the most advanced frontier models, this logic may now face challenges, especially if governments impose restrictions or block access to frontier models, potentially impacting the revenue expectations of related companies.Alex Svanevik further questioned whether the business model of relying on high-end models for profitability remains viable when regulations begin to limit the capabilities or deployment of frontier models, calling it a core issue that the industry needs to reassess.