DeepSeek V4.1 Flash is set to launch soon, with its official claiming it outperforms the V4 Pro across the board.

According to on-chain analyst Ai Yi (handle @ai_9684xtpa), a new address chased the rally of VVV tokens during its uptrend from August 18 to September 4, purchasing 181,250 tokens at an average price of $16.69 for a total value of $3.025 million, and now has an unrealized profit of $2.002 million.

Beating AI Express (Insight) – AI coding firm Magic has released a pre-training study, claiming a new training scheme achieves pre-training performance nearly on par with DeepSeek V4 Pro Base using just ~$500,000 worth of GB200 compute, with compute requirements roughly 1/50th of the latter. "Base" refers to a foundation model that has only completed pre-training, not undergone post-training steps like reinforcement learning. Magic tested models on unseen code, math problems, and research papers to evaluate their accuracy in predicting subsequent content. The comparison here is limited to base model pre-training performance, and does not mean Magic’s models have matched the full DeepSeek V4 Pro’s chat, coding, or agent capabilities. Magic then scaled the training size 10x, requiring ~$4 million in GB200 compute. This scaled version outperformed DeepSeek, Kimi, and NVIDIA’s latest publicly available base models in the same self-administered test. Magic estimates that reaching the same performance level at DeepSeek V4 Pro’s training efficiency would cost over $100 million in compute. The firm did not disclose the full training recipe, only noting efficiency gains stem from dozens of adjustments to model architecture, optimizer, training objectives, data processing, and other areas. As model weights have not been released, the "50x efficiency" claim cannot be independently verified by external parties.

According to market data from BIT (bit.com), U.S. pre-market trading on Wednesday saw modest strength in the semiconductor, storage, and optical communications sectors. Semiconductor stocks rose broadly: Lam Research (LRCX) gained 0.91%, Applied Materials (AMAT) climbed 0.41%, and Arm (ARM) advanced 1.12%. The storage sector also trended higher, with SK Hynix (SKHY) up 1.63%, Micron Technology (MU) rising 0.84%, Western Digital (WDC) gaining 0.38%, SanDisk (SNDK) increasing 1.19%, and Seagate Technology (STX) adding 0.39%. Optical communication-related stocks led the gains: Astera Labs (ALAB) rose 1.31%, Applied Optoelectronics (AAOI) gained 0.98%, Coherent (COHR) advanced 0.79%, Credo (CRDO) up 0.95%, Lumentum (LITE) climbed 0.5%, and Nokia (NOK) jumped 4.79%.

Bernstein analysts have retained their "Outperform" rating on Robinhood Markets, along with a $160 price target, noting that the firm’s newly launched Layer 2 network has quickly become a profit driver. In a Tuesday report sent to clients, the team led by Gautam Chhugani pointed out that Robinhood Chain has grown rapidly since its July 1 launch. The network currently holds a Total Value Locked (TVL) of roughly $1.5 billion, with Decentralized Exchange (DEX) trading volume exceeding $50 billion. The analysts wrote: “This chain is now capable of generating profits.” They added that Robinhood Chain’s daily transaction fees currently range from $2 million to $4 million. Over the past 15 days, the chain ranked first among all blockchains in fee revenue, totaling roughly $33 million—surpassing Solana’s ~$11 million and BNB Chain’s ~$9 million.

Arthur Hayes announced in a post that Flop Labs will launch a KOL ranking program, assigning each participating KOL a unique referral link. KOLs can embed this link in their content, and the platform will use it to track their referred users’ contributions to FLOP and network usage. Additionally, users who create wallets via a KOL’s exclusive referral link will be eligible to participate in regular FLOP giveaways. The program will be open to all users in the future, with more event details to be announced soon.

According to market data from BIT (bit.com), Lithium Americas' US stock gained over 5% in pre-market trading, following JPMorgan Chase assigning the stock an "overweight" rating.