A Foresight article argues that 2026 is shaping up as a transition year for DePIN projects focused on AI compute. The piece says the sector is moving away from token-subsidy narratives and toward proof of protocol revenue and engineering reliability. That shift is happening as AI workloads begin to move from training-heavy demand toward inference-heavy use cases, a change the article says better fits distributed compute networks. The market picture remains mixed. As of mid-July 2026, total DePIN sector market capitalization stood at about $3.46 billion, down roughly 83% from a March 2024 peak of $20.2 billion, with a 23% decline for the year. Yet protocol revenue did not contract at the same pace. In January 2026, monthly on-chain revenue for DePIN projects reached $150 million, driven by payments from enterprises for compute, storage and connectivity. Aethir led with $55 million, followed by Render Network at $38 million and Helium at $24 million. The article says lower listed H100 rental prices on DePIN platforms do not automatically translate into enterprise adoption, pointing to service-level guarantees, engineering friction and procurement barriers. It also highlights io.net’s June 2026 launch of its Incentive Dynamic Engine and Aethir’s disclosed annualized revenue above $100 million as signs that leading projects are trying to tie token economics more directly to real network usage.

Hong Kong-based Tianfeng Securities’ overseas technology research team says the latest round of quarterly results across the AI supply chain has strengthened the case for an AI bull market comeback. In its review, the firm argued that cloud growth has reaccelerated, demand for compute continues to run ahead of supply, and customers are no longer resisting higher prices. In some cases, they are paying upfront construction costs, which Tianfeng says is improving unit economics across AI infrastructure. The report points to record cloud contract activity over the past 20 quarters and says results from Google, Amazon Web Services and Microsoft Azure support the view that long-term cloud growth remains intact. It also highlights CoreWeave and Nebius as evidence that new compute providers have gained pricing power, with strong contract growth, prepaid terms and shorter payback periods. On the supply-chain side, Tianfeng says optical communications, memory and storage are showing signs that a bottom has formed, while debate over AI infrastructure returns has been weakened by fresh earnings data.

U.S. stocks closed mixed after July consumer inflation data came in as expected, easing immediate concern that price pressures were reaccelerating. The Dow Jones Industrial Average slipped 0.04% for a third straight loss, while the S&P 500 rose 0.26% and the Nasdaq gained 0.54%. According to the report, July CPI rose 0.1% month over month and 3.4% year over year, while core CPI increased 0.2% on the month and 2.5% on the year, with the core annual reading marking its slowest pace since March 2021. Money markets subsequently cut the probability of a September rate hike to 40%, with odds of no change climbing to about 60%. The market’s leadership also shifted sharply. Capital moved out of traditional software and some mega-cap tech names and into AI infrastructure, semiconductors, storage and optical communications. NEBIUS jumped 34.14%, CoreWeave rose 19.28%, Super Micro Computer climbed 19.02% and Lumentum added 13.63%. The report also highlighted continued strength in gold and silver, a pullback in crude after a recent surge, and mounting pressure from the U.S. fiscal deficit and interest burden. Investors are now watching July PPI, upcoming 13F filings and earnings from several technology and semiconductor companies for the next directional cues.

Securitize’s first earnings report since going public has sharpened a question hanging over the real-world asset sector: can tokenization platforms turn rising on-chain asset volume into meaningful revenue? The company reported a record $4.3 billion in tokenized assets under management and $5.3 billion in on-chain transaction volume, up 16% and 147% year over year, respectively. Yet quarterly revenue fell 5% to $14.4 million, well below Wall Street’s $20.6 million expectation, while net loss widened to $21.7 million. Adjusted EBITDA also swung from a $1.8 million profit a year earlier to a $5.5 million loss. The report points to a basic tension in the RWA business model. Securitize powers issuance, transfer agency, compliance KYC, and broker-dealer functions for products such as BlackRock’s BUIDL, but it does not collect the larger pool of economics that goes to asset managers, investors, and custodians. Based on the figures cited in the article, its implied annualized fee rate works out to roughly 1.34 basis points on $4.3 billion of AUM. That has pushed investors to look past top-line tokenization growth and ask a tougher question: how much can infrastructure providers actually earn from every dollar that comes on-chain?

Hyperliquid is posting bigger trading numbers even as the revenue that supports HYPE has been shrinking. Open interest on the platform climbed above $11 billion on July 13, the highest level recorded for Hyperliquid in 2026, while 30-day perpetual futures volume approached $178 billion. Including centralized venues, the platform now accounts for about 9% of global perpetual open interest, up from less than 7% at the end of May. The revenue line has moved the other way. DeFiLlama data cited in the report shows Hyperliquid protocol revenue peaked at roughly $357 million in the third quarter of 2025, then fell quarter by quarter to $295 million, $217 million and about $202 million in the second quarter of 2026. That leaves revenue down 43% from the peak even as trading activity keeps rising. The report links the gap to HIP-3, a change that lets outside builders launch their own perpetual markets on Hyperliquid’s order book by staking 500,000 HYPE, worth about $28 million at current prices, and keep as much as half of trading fees. Those externally deployed markets have grown from 2% of perpetual volume at the start of 2026 to nearly 50% now. Much of that growth has come from RWA-linked perpetuals, including oil, gold, Nvidia, Tesla, Nasdaq 100 trackers and contracts tied to private companies such as SpaceX.

Cerebras Systems reported its first quarterly results since going public, showing a sharp shift in how the AI chip startup makes money. In the quarter ended June 30, 2026, revenue reached $180 million, up 74% from $103 million a year earlier, according to the company’s 10-Q filing with the U.S. Securities and Exchange Commission. Nearly all of that growth came from cloud and AI compute services, where revenue jumped 281% year over year to $126 million, while hardware revenue fell 23% to $54.12 million. Cloud and other services accounted for roughly 70% of quarterly revenue, reversing the company’s mix from the prior year. Profitability moved in the opposite direction on a GAAP basis. Cerebras posted a net loss of $451 million, compared with net income of $310 million a year earlier, after recognizing $377 million in stock-based compensation following its IPO. Operating expenses rose to $503 million, including $320 million in research and development. Still, on the company’s own non-GAAP measure, core operating loss narrowed to $33.61 million from $43.9 million. The filing also detailed deeper commercial ties with OpenAI and Amazon Web Services. OpenAI committed under a 2025 master agreement to buy 750MW of AI inference compute capacity and related services between 2026 and 2028, with an option for another 1.25GW by the end of 2030. Cerebras also said OpenAI exercised part of a warrant in July 2026. Separately, the company signed a global hardware leasing agreement with AWS in June and disclosed more than $2.2 billion in combined future minimum data center lease commitments from two sets of contracts.

Ethereum’s September 15, 2022 Merge cut the network’s energy use by about 99.95% and replaced proof-of-work with proof-of-stake, reshaping both ETH issuance and the chain’s security model. Nearly four years later, that decision is back in debate through a counterfactual lens: what if Ethereum had stayed on PoW through 2026 instead of moving to PoS? The discussion centers on a trade-off. On one side is the idea that Ethereum’s former GPU mining base could have evolved into one of the world’s largest distributed compute pools, potentially giving the network a stronger position in the AI era. Waterdrip Capital co-founder Jademont and crypto KOL 嗯哼 both argue that millions of GPUs once tied to ETH mining might have been reorganized into a decentralized AI cloud if the protocol and ecosystem had developed mechanisms for task scheduling, verification and revenue sharing. On the other side are the costs. Ebunker co-founder 0xTodd said Ethereum was producing roughly 13,000 ETH per day near the end of PoW, versus about 3,000 ETH after the switch, implying daily sell pressure of around $26 million versus $6 million at a $2,000 ETH price. The article also argues that staying on PoW would likely have meant more inflation, heavier miner selling, more constraints on scaling, and harder conversations around energy use, ESG and institutional adoption.

Crypto exchange Bybit has filed suit in the U.S. District Court for the District of Columbia against the North Korean government, the Reconnaissance General Bureau, and the Lazarus Group over the theft of about $1.5 billion in digital assets in February 2025. The complaint, recently unsealed, seeks roughly $1.5 billion in compensatory damages under the U.S. Racketeer Influenced and Corrupt Organizations Act, along with treble punitive damages. The case stands out because it is described as the first time a crypto exchange has used the U.S. federal court system to pursue civil recovery directly against a sovereign state and state-backed hackers. The court has already issued a preliminary injunction freezing related wallets and platform funds. Bybit said it has recovered about $48.4 million so far, while more than $30.5 million has been frozen by 28 exchanges and custodians worldwide. Even so, tracing the stolen funds remains difficult: 90.2% of the assets are now hard to track after moving through mixers, cross-chain bridges, and OTC channels, while only 9.8% can still be traced to specific wallet addresses. At the same time, Bybit is facing regulatory pressure in several jurisdictions, including Singapore, Malaysia, Canada, the United States, China, and Hong Kong.