Trading front-ends are gradually becoming market winners, "proprietary model VS partnership model."

Goldman Sachs upgrades CoreWeave's rating, with multiple analysts defending the fundamentals, as fears of "AI capex peaking" somewhat fade. The Philadelphia Semiconductor Index ended an eight-day losing streak with a 1.6% gain. Previously, "AI stock god" Aschenbrenner fell into severe losses once again, with his seven heavily concentrated stocks averaging an 18.3% decline.

Patrick Witt departs with no successor.

The hardcore new blood network of Silicon Valley is unifying the AI and Crypto camps respectively.

The market will focus on whether AI investments can translate into actual profitability, as well as the impact of inflation, employment and other data on interest rate expectations.

High interest rates are accomplishing one thing: they are forcing the market to reprice a long-forgotten common sense—capital has a cost, and financing has boundaries. When AI construction and government deficits are simultaneously competing for money from the same capital pool, who will be the first to buckle?

Howard Marks says a real bubble requires investors to be "oblivious," and typically requires several conditions to be met simultaneously: market valuations clearly detached from investors' rational judgment, risk appetite severely disconnected from reality, and investor sentiment completely out of control. Valuation data supports his assessment: the forward P/E ratio of the top five tech stocks currently stands at about 23x, far below the 65x at the 2000 peak.

Over $1.1 trillion in "off-balance-sheet leases" are effectively disguised debt, and 2027 will be the year with the largest free cash flow shortfall; the risk of simultaneous declines in both volume and price is being significantly underestimated by the market.

The BTT from the first round of decentralized business profit buybacks has been fully burned, and BitTorrent is comprehensively transitioning to Web3 and AI infrastructure powered by its new architecture.

High-quality in-depth analysis articles and a weekly roundup of top hot topics.

Up 600% in 2 months, this year's hottest mining coin.

Inflation, war, and rate hikes have all failed to stop U.S. stocks from hitting new highs, and the upside for the S&P 500 has yet to truly unfold.

Amid panic selling that wiped out a trillion dollars from tech stocks in a single week, Jensen Huang pushed back against the pessimistic assumption that "AI compute depends entirely on two labs." He pointed out that one-third of Nvidia's compute is used for traditional computing unrelated to large models, another third flows to the open-source enterprise ecosystem, and frontier model labs account for only 30%.

On a three-year basis, net AI exposure remains in the 83rd percentile, and gross leverage is near a five-year peak — this selloff looks more like a beginning than an end.

In a week when AI stocks had just tumbled sharply and credit market alarms were ringing frequently, what this briefing delivered was not just a set of revenue figures, but a public test of the AI commercialization narrative.

High interest rates are continuing to weigh on consumption and investment, while the AI boom, compounded by high leverage and circular investment, is seen as the most vulnerable potential fracture point under a high-rate environment.