MegaETH trims a fifth of its workforce as strategy pivots toward AI applications
MEGA, the organisation behind the Ethereum layer-2 project MegaETH, has cut 20% of its workforce following months of internal restructuring, according to a report published on November 14, 2026. The redundancies touched teams across the project’s core functions, including staff who worked on the blockchain itself, the wallet, ecosystem support and communications.
Co-founder Shuyao Kong confirmed the departures and framed the move as part of a broader reallocation of resources. Rather than continuing to staff purely infrastructure roles, MEGA is redirecting capital and headcount toward products built on top of MegaETH, with a particular emphasis on artificial-intelligence services.
The cuts land at a moment when the competitive dynamics of the layer-2 sector have shifted noticeably. For much of the past cycle, rollup projects competed primarily on throughput, fees and technical differentiation. The emerging battleground is now applications, and AI-related services are increasingly seen as the demand driver most likely to fill blockspace and attract users. MegaETH’s restructuring suggests its leadership has reached the same conclusion and is willing to absorb the short-term optics of layoffs to fund that pivot.
There is a less flattering reading as well. A 20% reduction spanning blockchain engineering, wallet development and communications raises questions about execution capacity and staffing stability. Projects that thin out core protocol and developer-relations teams while chasing an adjacent trend can struggle to maintain the quality of the platform that the new products depend on. Kong’s public framing acknowledges the pivot, but the market will now watch whether MegaETH can ship AI-oriented products quickly enough to justify the reduced headcount on the base layer.
For the wider layer-2 market, the episode is a signal that infrastructure oversupply is beginning to bite. The number of general-purpose rollups has grown faster than organic demand, forcing teams to differentiate or consolidate. A pivot toward AI applications, where monetisation models are still being worked out, is a high-variance bet, but it is one that a growing share of the industry appears willing to take.
Readers tracking the broader trend will find more in our Ethereum coverage, where layer-2 competition and application-layer demand have been recurring themes this year.
OpenAI turns ChatGPT into a payment agent, with card networks rather than blockchains
OpenAI has introduced a ChatGPT feature that allows AI agents to make purchases on behalf of users through the American Express, Mastercard and Visa payment networks. The capability is initially available to ChatGPT Pro subscribers in the United States using eligible US-issued cards.
The mechanics matter as much as the headline. Every purchase requires biometric authorisation, such as facial recognition or a fingerprint, before it proceeds. The card networks then issue tokenised card credentials for approved transactions, which means neither the merchant nor the AI agent ever handles the user’s actual card details. OpenAI has also said it plans to support preset spending limits and other controls, opening the door to purchases that do not require approval for every individual transaction.
For the payments industry this is a significant step. For crypto, the implications are more nuanced.
On one hand, the launch accelerates a trend the crypto sector has been anticipating for years: autonomous software agents executing financial transactions without a human typing in card numbers or navigating checkout flows. Once agentic payments become normal consumer behaviour, the conceptual distance to self-custodied wallets, stablecoin settlement and on-chain rails shrinks considerably. Agent-to-agent commerce, in which software negotiates and settles with other software, is an area where blockchain-based settlement has long been argued to hold structural advantages, particularly around programmability, finality and the absence of a card-network intermediary.
On the other hand, the first iteration of ChatGPT payments runs entirely on conventional card rails. Visa, Mastercard and American Express are the settlement layer here, and tokenised credentials are the privacy mechanism, not zero-knowledge proofs or wallet abstraction. OpenAI has chosen incumbency over disruption, at least initially. That decision is a reminder that card networks still offer something crypto rails have not yet matched at scale: universal merchant acceptance,成熟的 chargeback protections and regulatory clarity in the jurisdictions that matter most to a consumer product like ChatGPT.
The competitive question now is whether crypto-native agent payment infrastructure can carve out a role before the card networks entrench themselves as the default. Stablecoin issuers and wallet providers have been building toward exactly this use case. If OpenAI’s feature demonstrates that consumers will delegate spending to AI agents, and it almost certainly will at scale, the pressure grows on merchants and agent developers to evaluate whether card fees and biometric gates are the right long-term architecture. The answer will not arrive overnight, but the starting gun has been fired.
The longer-term irony is worth noting. An AI agent that can spend money on a user’s behalf is functionally similar to a smart contract wallet with an authorised spender. Crypto has been building that pattern natively for years. What OpenAI has shown is that distribution, trust and a friendly interface can matter more than technical elegance, at least for the first hundred million users.
Jane Street nearly eightfolds Fidelity Bitcoin ETF stake while trimming BlackRock’s IBIT
Third-quarter disclosures show that trading firm Jane Street increased its holdings of Fidelity’s spot Bitcoin ETF by nearly eightfold, while simultaneously reducing its position in BlackRock’s iShares Bitcoin Trust, IBIT, by 19%. Its IBIT holdings fell from 24.88 million shares to 20.22 million.
The contrasting moves have prompted speculation about what one of the most sophisticated trading desks in global markets sees in the Bitcoin ETF complex. The honest answer is that the filings alone do not settle it.
Several plausible explanations exist, and they are not mutually exclusive. The first is straightforward portfolio rebalancing: a firm managing enormous books across asset classes routinely rotates exposure between highly correlated instruments, and the two largest spot Bitcoin ETFs are near-perfect substitutes from an exposure standpoint. The second is liquidity preference. IBIT is the deepest and most liquid spot Bitcoin ETF by most measures, which makes it the natural instrument for fast position changes, but Fidelity’s product offers its own advantages in tracking, fee structure and creation-redemption mechanics that may suit a longer-horizon holding. The third is product-level differentiation, where small differences in lending arrangements, spreads or counterparty exposure can make one wrapper more attractive than another for a particular strategy.
What the filing does not support is the conclusion that Jane Street is retreating from Bitcoin exposure. Selling 4.66 million IBIT shares is only a withdrawal if the proceeds left the asset class entirely, and the near-eightfold increase in the Fidelity position suggests the opposite. Jane Street remains one of the most significant institutional participants in spot Bitcoin ETFs, and its actions continue to shape the flow dynamics of the entire complex.
The deeper lesson is about market structure. The disclosures demonstrate just how materially a single institutional player can shift demand among spot Bitcoin ETFs. When flows concentrate in a handful of authorised participants and market makers, the differences between “the Bitcoin ETF trade” and trades in individual products become meaningful. Relative value between IBIT, the Fidelity product and the smaller funds is now a live strategy, and quarterly 13F-style disclosures have become a closely watched, if lagging, window into how the biggest desks are positioned.
For investors, the practical takeaway is caution against reading any single quarter’s repositioning as a directional signal on Bitcoin itself. For market watchers, it is a reminder that the ETF complex, covered in detail in our Bitcoin coverage, has matured into a multi-instrument institutional market with its own internal dynamics, not simply a passive wrapper around a single asset.
What the day’s three stories say about where crypto is heading
Taken together, these three developments sketch a coherent picture of the current cycle.
MegaETH’s layoffs show a layer-2 project concluding that infrastructure alone no longer guarantees relevance, and that AI applications are where the next wave of demand is expected to come from. OpenAI’s agentic payments launch shows the same logic arriving from the other direction: the largest consumer AI company in the world has now made autonomous spending a mainstream feature, using card rails for now but normalising the behaviour that crypto-native payment infrastructure was designed to serve. And Jane Street’s ETF repositioning shows institutional engagement with Bitcoin deepening into something more nuanced than simple accumulation, with active rotation between products reflecting the market’s growing sophistication.
The common thread is convergence. Crypto is no longer competing only with itself. It is competing with card networks for agent payments, with traditional finance wrappers for asset exposure, and with the AI sector for developer talent and application demand. The projects and firms that recognise this, and restructure accordingly, will define the next phase of the market.
