ChainFeeds on July 23 published a new edition of its Daily research briefing, combining headline news with a set of long-form reads on Movement Labs, public-chain business models, Japan’s crypto market structure, and the migration of crypto-era capital and infrastructure into AI.
Headlines in the daily briefing
The roundup listed several news items at the top: OpenAI launched an enterprise AI agent product called Presence; Reuters reported that Polymarket plans to challenge France’s decision to block its website through legal channels; Tesla kept its Bitcoin holdings unchanged in the second quarter and reported a $112 million impairment loss; the U.S. Senate released the latest text of the CLARITY Act, adding developer protections and digital-asset ethics provisions; and Uniswap’s DualPool Hook completed an audit and was open-sourced for team-led deployment.
Market demand shifts from narrative to product-market fit
In the first featured note, ChainFeeds cited Tiger Research on what it described as a PMF era for crypto. The argument is that market demand in 2026 is clustering around two very different poles: speculative demand for volatility and fast returns on one side, and real financial demand on the other, including secure custody, value transfer, and efficient asset management.
Tiger Research said crypto has long moved in cycles dominated by a single narrative. It mapped four major periods: DeFi in 2020, NFT / P2E / GameFi in 2021, L1 / L2 competition in 2022, and restaking in 2024. In its view, sectors that never proved real product-market fit eventually ran out of momentum once the story weakened.
GameFi was presented as the clearest example. Square Enix and Ubisoft were among the traditional gaming companies that entered the field, and the gaming segment drew $2.5 billion in capital inflows in the first quarter of 2022 alone. But Axie Infinity, used in the note as a representative project, saw its monthly active player base fall from a peak of 2.8 million in January 2022 to about 8,000 in May 2026, a 99.7% drop.
Stablecoins, by contrast, were described as a sector with measurable demand. The category’s market capitalization has reached $304.2 billion, close to the record high of $321 billion. Tether’s USDT stands at $184.08 billion in market cap, with $1.79 trillion in monthly settlement volume, up 63% from the prior month. Over the past 12 months, cumulative settlement volume reached $10.2 trillion. The report added that Tether posted more than $10 billion in net profit in 2025 and held $141 billion in U.S. Treasuries. Circle’s USDC was listed at a $73.25 billion market cap and described as the default settlement stablecoin across major exchanges including Coinbase and institutional rails.
The piece said stablecoins are no longer used only to avoid price swings during crypto trading. Their role now extends to cross-border remittances and on-chain payment infrastructure. In June 2026, more than 140 traditional companies, including Visa, Mastercard, Stripe, Coinbase, and BlackRock, jointly announced support for OUSD, or Open USD. Non-dollar stablecoins tied to the Korean won, Japanese yen, euro, and other currencies were also highlighted. Their combined market cap is still only $1.2 billion, but the number of wallets holding them rose from 40,000 in January 2023 to 1.2 million in March 2026, roughly a 30-fold increase.
Prediction markets were the other sector singled out in the report. Their total market capitalization was put at $9.58 billion, making them the youngest of the five sectors discussed. The two dominant platforms, Kalshi and Polymarket, have not issued tokens. Kalshi has raised $2 billion in total at a $22 billion valuation, and its June 2026 trading volume reached $31.5 billion, up 87.4% month over month. Polymarket has raised about $1.6 billion at a $9 billion valuation. On its main platform outside the U.S., June trading volume was $10.26 billion, up 45% from the previous month. Since obtaining a U.S. operating license, the platform’s annualized revenue has surpassed $1 billion.
The note also warned that the World Cup was both a catalyst and a test. The event drove a sharp jump in June activity, but after the July 19 final, open interest on the two platforms fell nearly 20% from an early-July peak of about $2 billion. With sports contracts making up about 80% of total volume during the tournament, the report said volumes are likely to stay subdued until the next major event, the U.S. midterm elections.
Movement Labs bankruptcy traced back to a disputed market-making contract
The second article in the briefing, from Foresight News, focused on the collapse of Movement Labs. Its framing was blunt: the company raised $38 million, but 19 months later all remaining assets were not enough to cover $1.6 million in legal fees claimed by its ousted founder.
The report said the trouble surfaced soon after MOVE began trading on Binance on Dec. 9, 2024. A CoinDesk investigation found that a deal between Movement and market maker Rentech gave Rentech control of about 66 million MOVE tokens, roughly 5% of total supply. After listing, those tokens were sold aggressively, MOVE fell sharply, and Binance and Coinbase later suspended trading in the token.
Further reporting found an undisclosed relationship between Rentech and Chinese market maker Web3Port. Movement said it may have been misled about that connection when the agreement was signed. The Movement Foundation launched a token buyback after the scandal broke, but the confidence shock did not reverse.
CoinDesk-obtained contract documents showed how the arrangement took shape. On Nov. 27, 2024, co-founder Rushi Manche sent a draft market-making agreement to the Movement Foundation over Telegram and asked for it to be signed. The contract proposed lending 66 million MOVE to an entity called Rentech, which had no public operating history at the time. Foundation legal counsel YK Pek wrote internally that it was “possibly the worst agreement I’ve ever seen,” while foundation director Marc Piano refused to sign.
One clause drew particular scrutiny. If MOVE’s fully diluted valuation exceeded $5 billion, Rentech would be allowed to liquidate the tokens and split the profits 50-50 with the foundation. Zaki Manian, cited as an industry figure who reviewed the contract, said that structure effectively rewarded a market maker for pushing the price up and then selling into retail demand. The foundation rejected the first draft, but talks continued. Rentech later told the foundation it was a subsidiary of Web3Port and promised to post $60 million of its own funds as collateral.
The report then turned to the company’s debt and control struggle. After being removed as CEO, Rushi Manche sued MVMT Labs in the Delaware Court of Chancery, asking the company to advance legal expenses under its charter for his cooperation with government investigations. He won. The U.S. Department of Justice is now conducting a grand jury investigation into the MOVE token issuance process, and Manche, as a central decision-maker during the token launch, needed legal representation. Those fees make up the bulk of his claim for more than $1.6 million.
That left the company in an unusual position. Foresight News said the business was driven into bankruptcy by the market-maker scandal, while the obligation to cover the ousted founder’s legal bills related to the same scandal became the largest debt item in the bankruptcy filing. Manche still holds 34.25% of MVMT Labs, retains the title of co-founder, and has no operating decision-making power. Under Chapter 11, the treatment of that equity stake is expected to be a central issue in restructuring talks.
Public chains are looking beyond block-space sales
The third piece, citing Castle Capital, argued that blockchains are being forced to rethink how they make money. It opened with a sharp comparison: Arbitrum generated only about $430,000 in revenue over the past 30 days, while Hyperliquid brought in about $58 million, more than 100 times as much.
Castle Capital’s point was that selling block space on its own no longer looks like a durable business model. More chains are now experimenting with roles such as product studios, application distribution platforms, payment rails, or vertical SaaS stacks. The report said the direction of travel is clear: more blockchains are likely to move away from the idea of neutral infrastructure and take direct control of specific verticals.
It used the recent Ostium exploit to show the operational risks tied to off-chain dependencies. Last week, Ostium’s LP vault was attacked after the attacker gained control of the off-chain infrastructure used to feed prices into the protocol and stole 23,752,746 USDC. The attacker submitted forged but apparently valid price reports, then opened and closed large positions against those prices to pull fake profits from the vault.
The article said this cut to the core of Ostium’s product because the protocol is built to bring off-chain markets on-chain. Stocks, commodities, and FX do not have native on-chain prices in that environment, so the protocol must import those prices and trust them. Most users, it argued, do not want to think about Greeks, expiries, strike prices, or volatility surfaces. They want yield, leverage, protection, or a simple way to express a market view. For that reason, the products with the best odds of broad user adoption may not be standardized options exchanges at all, but yield vaults, short-duration binary contracts, structured products, and prediction markets.
The article also cited Rysk’s Dan, who said options are not the product; the yield and functionality they enable are the product. Rysk said its new products accumulated more than $1 billion in open interest over the past year, with users mainly coming from deep DeFi cohorts seeking yield on assets they already hold rather than arriving as options traders.
Japan’s market structure rests on regulation, exchanges, and institutionalization
The fourth long read, from Four Pillars, examined what it called the three pillars of Japan’s crypto market: regulation, exchanges, and institutionalization. The argument was that Japan has not tried to be the fastest or most open crypto jurisdiction. It has been building, over many years, a market designed to be trusted.
According to the piece, 2026 marks an inflection point in Japan, but the shift is the product of more than a decade of regulatory iteration, often triggered by crises. Mt. Gox’s collapse in 2014 made Japan the first major economy to confront custody risk head-on. Amendments to the Payment Services Act, or PSA, in 2017 made Japan the first G20 country to impose a formal licensing regime on crypto exchanges. The 2018 Coincheck hack, which caused losses of about $530 million at the time, directly pushed the creation of the self-regulatory framework led by the Japan Virtual and Crypto Assets Exchange Association, or JVCEA. After the 2024 attack on DMM Bitcoin, Japan moved ahead with a 2025 PSA revision that added asset preservation orders and a new licensing track for crypto intermediaries.
Four Pillars described this as an “event to institution to institutionalization” pattern. Today, the Financial Services Agency, or FSA, directly registers crypto-asset exchange service providers, stablecoin issuers, and the new class of crypto intermediaries, while JVCEA and JSTOA handle parts of industry self-regulation.
The article said Japan’s system can only be understood by looking at two laws together: the Payment Services Act and the Financial Instruments and Exchange Act, or FIEA. Since 2017, the PSA has treated crypto assets as payment instruments and required trading platforms to register with the FSA. It also underpins the electronic payment instrument, or EPI, framework for stablecoins introduced in 2022. FIEA, by contrast, has long governed securities, derivatives, security tokens, and the conduct of financial institutions.
That division changed with FIEA amendments passed in April 2026. The revision did not displace the PSA system. Stablecoins remain EPIs under the PSA, and custody operations by trading platforms still require PSA registration. But the change brought 105 approved spot crypto assets under the FIEA framework and introduced securities-market tools including ongoing disclosure, insider-trading prohibitions, market-manipulation rules, and oversight powers for the Securities and Exchange Surveillance Commission, or SESC.
For market participants, that means a heavier compliance load. Starting in 2027, licensed exchanges will need to satisfy PSA obligations on custody, anti-money-laundering controls, and travel-rule compliance, while also meeting FIEA standards for disclosure, insider-trading monitoring, and market-conduct oversight. Token issuers whose assets make the 105-item list will face continuing disclosure duties similar to those imposed on listed companies, including annual reports and material event notices.
The report also pointed to a structural consolidation in the Japanese exchange market. Official FSA data showed that Japan had about 27 registered crypto-asset exchange service providers as of February 2026, more than the combined total of licensed platforms in Hong Kong, Singapore, and South Korea. But trading volume remains concentrated among a few operators, with bitFlyer, Coincheck, SBI VC Trade, and bitbank holding the largest share of domestic spot trading.
Operating costs remain a serious burden. The FSA has said about 90% of Japanese crypto platforms are unprofitable under the current compliance regime. Costs come from registration, AML systems, JVCEA membership fees, cybersecurity infrastructure, segregated custody of customer assets, and, after 2026, new requirements around disclosure and insider-trading surveillance.
In that setting, only firms with enough trading scale can sustain profitability. Four Pillars said the most likely outcome between 2026 and 2028 is structural consolidation rather than broad market failure. It outlined three possible tiers: three to four large platforms at the top, including bitFlyer, Coincheck, SBI VC Trade, and possibly a consolidated bitbank; four to six mid-tier specialists focused on areas such as institutional services, derivatives, or alternative assets; and a long tail of smaller operators that may specialize, merge, exit, or be acquired by larger institutions.
The crypto assets now feeding the AI buildout
The fifth piece, from TechFlow, tracked how crypto-native assets have been redirected into AI since 2022. It opened in Abilene, Texas, where eight H-shaped data centers are being brought online across a site of about 1,000 acres. The location is the first large campus in OpenAI’s Stargate plan, built for 1.2 gigawatts of capacity. The first two buildings are already operating and the rest remain under construction. Oracle runs the campus, while the developer behind it, Crusoe, started out in Bitcoin mining. Founder Chase Lochmiller was previously a partner at Polychain Capital.
TechFlow said Lochmiller and Cully Cavness realized in 2018 that U.S. oil fields were burning large amounts of stranded natural gas that could not be transported. They moved power equipment and mining rigs to the wellhead and used gas that would otherwise be wasted to mine Bitcoin. Seven years later, Crusoe’s customer base had shifted from the Bitcoin network to OpenAI. In 2025, Crusoe sold its Bitcoin business, which included more than 425 modular data centers, to NYDIG and focused on AI.
The article argued that the leap from crypto to AI is not as wide as it seems. Crusoe’s core capabilities remained the same: finding power, building data centers, and operating infrastructure. It said many other companies have made similar transitions. They do not belong to a single organized network, but they are linked by three types of assets left behind by the last crypto cycle: power access, land, and grid interconnection rights held by miners; engineers and founders trained inside crypto companies; and capital accumulated during the last bull market. After 2022, all three started flowing into AI.
One point the article stressed was that mining companies are not simply turning mining rigs into AI servers. Bitcoin miners largely use ASIC chips built for specific hash functions and cannot train large AI models. Even GPUs left behind by Ethereum mining farms often fall short of current AI cluster requirements around networking, memory, liquid cooling, and reliability. The real asset miners hold is grid-connected data-center infrastructure.
Building an AI data center, the article said, is often constrained less by buying GPUs than by securing hundreds of megawatts of stable power, land, substations, transmission lines, and permits. Those processes can take years. Mining firms had already done much of that work in their attempt to lower mining costs and improve compliance. Once Bitcoin mining margins weakened and AI companies were willing to sign long-term, high-value data-center contracts, the shift in customer base followed.
CoreWeave was presented as one of the earliest examples. Three commodities traders started mining Ethereum with GPUs in 2016, then bought large volumes of used graphics cards during the crypto winter and expanded into film rendering and machine learning. The company was first called Atlantic Crypto and later renamed CoreWeave. Public filing materials showed that before 2022, most of its revenue still came from crypto mining. After that, the crypto business was shut down. CoreWeave is now one of the leading AI cloud firms backed by Nvidia, and TechFlow said more miners are following a similar path.
The report then shifted to capital. Jed McCaleb was one of the clearest examples cited. The early crypto billionaire, known for creating Mt. Gox and helping found Ripple and Stellar, put about $500 million through the Navigation Fund in 2023 to buy 24,000 Nvidia H100 GPUs in one move and set up Voltage Park, which rents GPU capacity to AI companies and research organizations.
It also revisited an investment tied to FTX founder Sam Bankman-Fried. In 2022, SBF invested $500 million in Anthropic and received about 13.5% equity when the company was still far less visible. After FTX entered bankruptcy, the estate sold those shares in 2024 and recovered about $1.3 billion. Anthropic is now valued in the tens of billions of dollars, and the article said the upside would have been far larger had the stake been retained.
Cursor was used as another example. In 2022, Alameda participated in Anysphere’s early financing with a $200,000 investment. Anysphere later launched the AI coding tool Cursor. After FTX entered bankruptcy, the estate sold that position at close to cost. As Cursor grew rapidly, the value of that stake may later have reached into the billions of dollars.
TechFlow’s conclusion was not that crypto investors had some special ability to predict AI. Its narrower claim was that before ChatGPT’s release, some of the highest-risk-tolerance capital produced in the crypto bull market had already started looking for AI projects. Capital accumulated in crypto is now becoming a funding source for AI infrastructure, model development, and startup experimentation.
