At 14:28 UTC on Sunday, the Liquid Federation paid out 3,996 bitcoin to a customer’s address on the main chain. Nothing about the payment looked wrong. A peg-out request had come in through SideSwap, an approved trading venue on the network, the matching L-BTC had been burned on the sidechain the way the protocol requires, and the signers signed. A small 2.5 BTC transfer had gone out just before it. The two amounts ended up consolidated in a single address holding about 3,998.5 BTC.

Fasset closed a $68 million round led by Japan’s SBI Group at a $1 billion valuation, announced Monday. The gap between this round and the last one is the detail worth pausing on: the company raised $51 million in May, which puts total funding for the year at $119 million and compresses the whole journey to unicorn status into roughly three months of calendar time.

Valuations that move that fast usually signal one of two things. Either the market is repricing a category wholesale, or the company put numbers on the table that changed the conversation. In this case it looks like both.

Museum displays of Roman building material tend to be badly labelled and quickly walked past. A broken lump of wall behind glass with a card reading “concrete, 2nd century AD” gives a visitor nothing. Which is a shame, because the cut face of that lump contains most of the answer to why Roman structures are still standing.

Here is what to look for when you find one.

The wall is a sandwich, and the pretty part is the formwork

The single most common misreading is treating the visible facing as the wall. It is not. Roman walls of this type consist of a thick core of mortar and rubble, faced on both sides with a thinner skin of stone or brick that served as permanent shuttering during construction and as a finished surface afterwards.

Cloudflare spent July building the side of the machine payments market where money gets collected. This week it shipped the side where money gets spent.

Cloudflare Wallets launched Tuesday as part of the company’s Agents Week, giving AI agents running on its network a stablecoin balance and a readable name to present when they buy something. The name lives at cloudflare.pay, where account owners can reserve a handle now. Everything else, funding the wallet, fiat onramps, letting agents actually spend, arrives “in the coming months.”

Every week I get an aggregate report across the whole network and every week the top line tells me almost nothing. This one: 15,630 visits, down 3.58%. 16,130 page views, down 4.1%. Median load time 3,252ms across 65 sites. Flat. Fine. Move on.

Except the aggregate is the least trustworthy number in the whole report, and this week made that obvious in two separate ways.

Start with the browser column, because that’s where it hurts. On analysis.org, 720 of roughly 1,300 visits came back tagged Unknown. Chrome was 390. Mobile Safari 50. On technologies.org it’s 520 Unknown against 610 Chrome. Unknown at that share isn’t a privacy-hardened reader base, it’s crawlers. Which means the headline number on my two biggest sites is inflated by something close to half, and every week-over-week percentage I’ve been reading off those two is partly a story about how often a bot felt like showing up.

ChangXin Memory Technologies begins trading on Shanghai’s STAR Market on Monday. A market for its shares has already existed for a week and a half — not in Shanghai, but on a decentralized derivatives exchange most of the company’s domestic subscribers have never heard of and cannot legally access.

Trade.xyz deployed a pre-IPO perpetual futures contract on Hyperliquid in mid-July, on the eve of subscriptions opening for what is Asia’s largest share sale of the year and the largest semiconductor offering ever run through China’s domestic A-share market. CXMT priced at 8.66 yuan, roughly $1.28, selling 6.69 billion shares for about 10 percent of the post-offering count and raising close to $8.6 billion. Within days the on-chain contract was trading near $8. Against a post-IPO share count of 66.881 billion, that implied a valuation around $535 billion — about 6.3 times what the underwriters had just settled on with institutional bookrunners in Shanghai.

Ant International has closed a Series A equity round of approximately $1.2 billion, with existing backers Ant Group and Alibaba Group joined by a group of undisclosed international institutional investors. The round follows weeks of reporting that pegged the target closer to $1 billion at a pre-money valuation near $10 billion, meaning the final raise came in above earlier expectations.

The Singapore-headquartered company was carved out of Ant Group as an independently operated business in 2024. It now connects more than 150 million merchants with over 2 billion user accounts globally, running the business through four units: Alipay+ for cross-border wallet interoperability, Antom for merchant payment acceptance, WorldFirst for cross-border trade payments, and Bettr for SME lending and treasury services.

Datavault AI (NASDAQ: DVLT) has laid out its strategic priorities for the third quarter of 2026, centered on three tracks: commercializing its Project Qestrel token program, expanding its SanQtum edge AI infrastructure, and advancing a cluster of tokenized data-exchange platforms that depend on U.S. digital-asset rules taking shape.

The token program

The centerpiece is the Qestrel token ($QEST), tied to a roughly $10 billion, ten-phase nationwide edge infrastructure buildout with a partner called Available Infrastructure. Datavault AI says the $1 billion phase-one deployment and an initial token offering are targeted for completion this quarter, with the company providing tokenization, clearing, valuation, and exchange services under contract and generating revenue from that work.

Kalshi and Polymarket built their entire identities around being each other’s opposite. Kalshi is the “regulated from the inside” exchange, holding the first federal license the CFTC ever granted an event-contract platform and pitching itself to lawmakers and institutions as the safe, onshore choice. Polymarket built the opposite brand, crypto-native and historically offshore, technically off-limits to American users even as VPN access made that limit porous, before working toward a US re-entry through its QCEX acquisition. Neither company’s survival argument works without the other looking like the risk.

Stripe and private equity firm Advent International have reportedly offered to acquire PayPal for roughly $53 billion — $60.50 per share, about a 28% premium over Tuesday’s close, backed by around $50 billion in committed bank financing. The two would hold equal stakes and, per the sources, keep the company intact rather than break it up. The approach dates to April; the formal proposal landed this month, with the parties targeting an agreement by the end of July. PayPal has not responded, and there is no guarantee it closes. This is a bid, not a done deal.