The AI boom is making a comeback, and South Korea's KOSPI index has entered a technical bull market.

Goldman Sachs notes two overinterpretations emerging in the market regarding the macro impact of AI capital expenditure: on one hand, investors are underestimating AI investment’s boost to the tech, energy and data center sectors; on the other, they are overstating the AI boom’s impact on the overall U.S. economy and investment in other industries. Goldman Sachs economists Jessica Rindels and David Mericle project in a new report that AI-related investment could reach roughly $600 billion this year, equivalent to about 2% of U.S. GDP, 10% of corporate fixed investment, and 15% of equipment investment. This size is enough to explain why AI infrastructure remains one of the most important trading themes in U.S. stocks, and why Nvidia, cloud providers, data centers, power equipment and the semiconductor chain continue to attract capital. However, Goldman Sachs also warns that AI investment’s direct contribution to GDP is not as large as headline figures suggest. The reasons: a large amount of AI equipment relies on imports and may not be fully counted in U.S. domestic output; at the same time, AI construction does crowd out some resources, with the crowding-out effect concentrated in three areas: cloud providers shifting internal budgets from traditional cloud services to AI, data center construction crowding out other commercial building resources, and AI-related debt financing driving up borrowing costs for other companies. In other words, AI investment is changing capital flows, but it is not strong enough to single-handedly rewrite the U.S. economic cycle. Goldman Sachs estimates that after accounting for direct and indirect impacts, AI’s net boost to U.S. GDP growth in 2026 may only be around 0.1 percentage points. This means AI remains a core driver of corporate earnings and structural stock market trends, but it should not be simply extrapolated as a reason for a broad acceleration of the U.S. economy.

Bankr developer @deployer tweeted that Sushi’s token launch platform Pools.fun will release its official protocol token. Thirty percent of all protocol fees will be allocated to buybacks and token burns; the platform also plans to conduct airdrops and launch a points system, with user trading volume and volume generated from token deployments on the platform serving as key evaluation metrics. Fees have already started accumulating, and will be used for buybacks once the token goes live. Per earlier reports, Pools.fun is a joint launch by Bankr and Sushi, designed to compete against Uniswap’s pools.trade—a Robinhood-style token launch platform.

According to data from DefiLlama, Robinhood Chain’s DEX trading volume exceeded $500 million in the past 24 hours, ranking fifth among all blockchains—only trailing Solana, BNB Chain, Ethereum Mainnet, and Base.

The Paxos-linked whale sold another 800 $BTC($50.72M) through Wintermute 8 hours ago. Over the past 2 months, the whale has sold a total of 2,500 $BTC ($154M).

Like all trading platforms regulated by the U.S. Commodity Futures Trading Commission (CFTC), prediction market firms typically aim to incentivize large traders and encourage companies to act as market makers to boost participation and trading volume. However, in guidance released Wednesday, the CFTC has raised concerns about these practices. The U.S. derivatives regulator warned event contract platforms that filings related to their incentive programs are on the rise, and these submissions are often "deficient both procedurally and substantively". This hinders the agency’s ability to assess whether platforms have "adequately notified participants of program terms and sufficiently evaluated the programs’ compliance". The CFTC noted that certain features of these reward programs "raise compliance concerns". Some rewards for high-volume participants may incentivize them "to trade solely to meet volume targets, increasing the risk of wash trades, prearranged transactions, or other fraudulent, manipulative, or disruptive trading practices". Additionally, market maker programs— which encourage firms to take both sides of a market trade— guarantee net returns or offset losses via stipends and rebates, a practice the regulator warns could also foster fraud and market manipulation.

Uniswap founder Hayden Adams said in a post that the team did not anticipate the tokens created during the pools.trade test would be discovered. Today, Uniswap has waived all creator fees from tests conducted by Uniswap employees, and allocated all such fees to automatic buy-and-burn contracts. The protocol is currently considering making this feature available to other deployers.