Lookonchain APP

App Store

The $200 billion financing behemoth behind Anthropic: Wall Street is buying chips, while Google and Broadcom provide backing.

1 days ago

Beating AI News: Behind Anthropic’s Google computing power deal, a roughly $200 billion financing and contract ecosystem has been set up. The Financial Times reports that around 80% of this structure is tied to TPU chips. Anthropic does not need to front all the capital itself; funds for chip purchases come mainly from Wall Street, while Google and Broadcom provide backing for data centers and chips respectively. The first batch of TPU chips, totaling ~$35 billion and 1GW, is first sold by Google to Broadcom, then transferred to a special purpose vehicle (SPV)—a financing entity established exclusively for this transaction. Firms including Apollo and Blackstone provide debt financing to the SPV, which buys roughly 1 million TPU chips and leases them to Anthropic. Broadcom offers residual value support for around $30 billion of this amount. If Anthropic defaults on rent payments later and proceeds from selling the TPU chips are insufficient to repay debts, Broadcom will cover the shortfall for senior creditors. Data centers are backed by Google’s guarantees: Google underwrites Anthropic’s leases, enabling developers to borrow funds for data center construction using Google’s credit. The FT notes that Google has supported 10 projects totaling 2.4GW, with a maximum potential exposure of around $440 billion. The full chain operates as: Wall Street funds chip purchases, Anthropic pays rent, Broadcom guarantees chip residual value, and Google backs data center leases. A multi-hundred-billion-dollar computing power order is thus split into layers of financing and guarantees.

Relevant content

Anthropic's over-the-counter (OTC) market capitalization is currently reported at $1.968 trillion.

HIP-3 market deployer Entropy has launched the Anthropic Pre-IPO market on Hyperliquid. ANTH is up 1.58% over the past 24 hours, currently trading at $1,968, with a 24-hour trading volume of $9.59 million and contract open interest of $9.9 million.

3 minutes ago

Analysis: Bitcoin futures demand cools, short-term trend may face further pressure

CryptoQuant analyst Darkfost wrote in a post that signs of weakening demand are emerging in Bitcoin’s futures market. Data shows the 30-day average net aggressive buy volume has started to decline after a period of relatively strong growth, a trend similar to that seen in May 2026. On August 19, when Bitcoin broke through $65,000, the Taker Buy/Sell Ratio rose to 1.21 at one point, but market optimism has faded as investors gradually increased their short positions. Currently, the 30-day average net aggressive buy volume has plummeted from $213.7 billion to $97.8 billion in just a few days, a drop of over 50%; meanwhile, the Taker Buy/Sell Ratio has remained negative since August 30. This metric, which compares the volume of aggressive buy and sell orders in the futures market, reflects investor sentiment and position changes in the futures market. The analysis notes that futures market trading volume is still significantly higher than that of spot and ETF markets, so its capital flows have a strong influence on overall market trends. In the short term, signs of deterioration are emerging in the Bitcoin market; weakening futures demand and rising short positions may further weigh on price performance.

3 minutes ago

Meta’s new ASR model tops real-time transcription rankings, costing just $0.18 per hour.

Insight Beating AI Newsflash: Meta Superintelligence Labs has launched Muse Voice Transcribe. The tool integrates real-time speech-to-text, speaker diarization, and end-of-speech detection into a single model. It can process audio up to over one hour long and distinguish between more than 20 speakers simultaneously. In English real-time transcription tests by Artificial Analysis, its Word Error Rate (WER, lower is better) stands at 3.1%, outperforming GPT Live Transcribe (3.9%) and Gemini 3.5 Transcribe Live (4.0%). Muse Voice Transcribe delivers final text roughly 0.16 seconds after a speaker finishes talking. Unlike models that apply a fixed wait time for all words, this model processes 80ms of audio at a time, deciding autonomously whether to continue listening or start outputting. Simple words are transcribed faster, while complex terms wait for additional context to confirm. Meta uses reinforcement learning to optimize both accuracy and latency. The model is trained on over 70 languages, with 25 fully validated, and can directly recognize mixed Chinese-English speech in a single sentence. It is currently integrated into Meta AI for Mac and Muse Code, and the Meta Model API is open for access. Pricing is set at $3 per 1,000 minutes, equivalent to $0.18 per hour.

3 minutes ago

Crypto-related stocks in pre-market US trading were broadly lower, with Robinhood edging down 1.28%.

According to market data from BIT (bit.com), crypto-related stocks in the US pre-market session fell across the board. Among them: CRCL dropped 2.48%, MSTR fell 2.08%, MARA declined 2.44%, COIN dropped 1.83%, BMNR fell 2.57%, and HOOD declined 1.28%.

3 minutes ago

Yilihua: Bitcoin could face a short-term correction, but remains bullish, with an expected rally to $86,000.

Liquid Capital founder Yilihua posted on social media that after Bitcoin failed to break the $81,000 resistance level a few days ago, he has been anticipating a pullback in the $75,500 to $76,000 range, after which he remains bullish toward the next resistance at $86,000. His market calls over the past few months have been largely accurate, but that does not guarantee future correctness. Every trade requires caution and a fresh approach; making mistakes is normal. Many argue that discussing market trends is too easily disproven, with a single error able to wipe out all gains, but pressing forward and continuous improvement embody the entrepreneurial spirit, and failure is an inherent part of the business process.

3 minutes ago

Binance will add 4 bStocks tokenized securities as collateral assets.

According to an official announcement, Binance will add four bStocks tokens—Seagate (STXB), ProShares UltraPro Short QQQ (SQQQB), Moderna (MRNAB), and CrowdStrike (CRWDB)—as eligible collateral assets for its Cross Margin, Unified Account, and Unified Account Pro products, effective at 20:00 (UTC+8, East Eight Zone Time) on September 2, 2026. Corresponding bStocks trading pairs will also open for margin trading. Eligible users will be able to use these bStocks tokens as margin collateral, expanding the available collateral options for margin trading.

3 minutes ago

Popular tokens

BitcoinEthereumHyperliquidSolanaTRONBNBTetherAaveXRPPepeFartcoinOndoJupiterUniswapBonkPendleEthenaArbitrumAvalancheLidoChainlinkPolygonDogecoinCardano