Something fundamental shifted in software development in late 2025 — and most organisations haven’t caught up yet. In February 2026, OpenAI published a landmark engineering blog revealing that a small team of three engineers had shipped one million lines of production code over five months without writing a single line manually. Every line was generated by an AI coding agent called Codex. The methodology behind that achievement has a name: harness engineering. For CTOs, product owners, and founders commissioning software today, understanding this shift is no longer optional — it is becoming the lens through which modern software quality and delivery speed must be evaluated.



Advertisement

From Prompt Engineering to Harness Engineering: A Brief History



The way engineers have worked with AI models has evolved rapidly through three distinct phases — and each phase reflects a deeper understanding of how AI actually produces reliable output.



Between 2022 and 2024, the dominant paradigm was prompt engineering: crafting the right instruction to get the best possible single response from a model. In 2025, context engineering emerged as the more sophisticated approach, focusing on what information surrounds a model’s context window at any given moment.

Harness engineering goes further still. Rather than optimising a single instruction or a single session, it asks: how do you design the entire environment in which an AI agent operates — across multiple sessions, multiple agents, and days or weeks of autonomous work?

As OpenAI’s Ryan Lopopolo summarised the lesson from their internal experiment: the hardest challenges in agentic software development now centre on designing environments, feedback loops, and control systems — not on writing code.

What Is a Harness, Exactly?

Advertisement



A harness is everything surrounding an AI agent except the model itself. It includes the structured documentation the agent reads, the architectural rules it must follow, the feedback loops that flag errors, and the tooling that lets it verify its own work. Without a well-designed harness, even the most capable AI model produces unreliable, inconsistent results.



The term borrows from horse tack — the equipment that both constrains and enables a horse to pull a load effectively. The metaphor is deliberate. An AI model is powerful but undirected; the harness channels that power towards a coherent, verifiable outcome.

In practice, a harness typically comprises three interconnected layers:

• Context engineering: ensuring the agent has access to the right information at the right moment — architecture documents, design decisions, product specifications, and progress logs — all versioned and stored inside the repository itself.

  Advertisement

• Architectural constraints: mechanically enforced rules that prevent the agent from drifting outside the intended code structure, regardless of how many tasks it completes autonomously.

• Entropy management: a recurring process that scans for outdated documentation, replicated anti-patterns, and accumulated technical debt, and opens corrective actions before problems compound.

What the OpenAI Experiment Actually Proved



Advertisement

The OpenAI internal experiment is the most concrete evidence to date that harness engineering works at production scale. Starting from an empty repository in August 2025, a team of just three engineers used Codex to build a fully functional internal product. By the time they published their findings, the repository contained approximately one million lines of code across application logic, CI configuration, observability tooling, tests, and documentation.



Roughly 1,500 pull requests were opened and merged. The team reported delivering at approximately ten times the speed of conventional manual development.

Crucially, early progress was slow — not because the model was incapable, but because the harness was not yet ready. Performance only accelerated as the environment was progressively improved.

Three engineering practices proved decisive:

1. A legible repository environment — rather than a single oversized instruction file, the team built a structured documentation directory with architecture maps, design documents, execution plans, and product specifications. The AGENTS.md file became an index, not an encyclopaedia.

2. Programmatic enforcement of architecture — a layered domain structure was enforced mechanically through custom linters and structural tests. If generated code violated these boundaries, the linter blocked it automatically.

   Advertisement

3. End-to-end verification tooling — the application was made bootable per git worktree, and browser developer tools were wired into the agent’s runtime, allowing Codex to reproduce bugs, validate fixes, and reason about UI behaviour without human intervention.

Why Generic Tooling Outperforms Specialised Tooling



Advertisement

One of the more counterintuitive findings from harness engineering research involves the tools given to agents. The natural instinct when building a domain-specific AI agent is to create bespoke, highly specialised tools for every task. Vercel’s engineering team discovered the opposite.



Their sophisticated internal text-to-SQL agent — built over months with complex specialised tooling — was outperformed by a dramatically simpler architecture. When they stripped it down to a single batch command tool, performance improved by 3.5 times, token usage dropped by 37 per cent, and the success rate rose from 80 per cent to 100 per cent.

The reason is straightforward: large language models have been trained on billions of tokens of standard developer tooling — bash commands, grep, npm, git. They understand these tools natively. Bespoke schemas, by contrast, introduce friction the model must work around.

This has direct implications for any organisation building or commissioning AI-powered software systems. A well-structured harness with generic, familiar tooling will consistently outperform a more complex one built around proprietary abstractions.

Advertisement

What This Means If You Are Commissioning Software in 2026



Harness engineering is not merely an internal concern for AI research teams. It is already reshaping what it means to build production software at speed and at scale — and it changes the questions that technical decision-makers should be asking of any development partner.

The role of the software engineer is evolving from writing code to designing the systems that make AI write reliable code. For CTOs and founders, the practical implications are significant:

• Delivery speed is no longer constrained by headcount in the same way — a well-harnessed agent system can scale output dramatically with a small team.

• Code quality depends increasingly on environmental design, not just on individual developer skill.

  Advertisement

• Technical debt now accumulates differently and requires active management processes, not just periodic refactors.

• Choosing a development partner who understands agentic workflows is becoming a meaningful competitive differentiator.

At Codescrum, we have been working at the intersection of software engineering and emerging AI methodologies for over 13 years. As harness engineering matures from experimental practice to industry standard, we are actively integrating these principles into how we architect and deliver software for clients across sectors — from government and finance to education and retail.



Conclusion

Advertisement



Harness engineering represents the clearest articulation yet of where software development is heading: away from code as the primary output of engineering effort, and towards environment design, feedback architecture, and structured knowledge management as the real drivers of quality and velocity.

OpenAI’s experiment proved it is possible. The broader industry — from Anthropic to Vercel — is now formalising the practices. The question for any organisation building digital products in 2026 is not whether to engage with this shift, but how quickly.



If you are evaluating how AI-assisted development methodologies could accelerate your next project — or if you want to work with a team that understands both the opportunity and the discipline required to implement it responsibly —

get a free consultation and estimate for your project

and let’s talk about what we can build together.

Gency AI, a San Francisco-based AI and blockchain infrastructure company, announced a $20 million funding round on March 17, 2026. The round includes participation from YZC Capital, MTmetaworld Holdings, Riverpark, ArkStream, MH Ventures, ViaBTC, and Basics Capital. The capital will be used to scale Gency AI’s decentralized advertising execution and settlement network, strengthen its privacy-preserving computing stack, and accelerate product deployment and ecosystem partnerships across North America, Asia, and Europe.

Key takeaways

• Gency AI raises $20 million to build and scale a decentralized advertising network anchored in verifiable on-chain data and privacy-preserving computing.
• Investors include YZC Capital, MTmetaworld Holdings, Riverpark, ArkStream, MH Ventures, ViaBTC, and Basics Capital, signaling appetite for verifiable, cross-border ad tech infrastructure.
• The project aims to shift ad tech from centralized platform trust to protocol trust through on-chain credentials and automated revenue distribution.
• The architecture centers on four core modules: policy identity, ESQ privacy computing, PSG clearing and settlement, and an AI optimization engine designed for encrypted operation.
• Industry implications point to faster settlements, clearer attribution, and increased regulatory alignment as privacy rules tighten and demand for AI-driven automated advertising grows.

Building verifiable infrastructure for the advertising economy

The digital advertising market has expanded rapidly, yet a substantial portion of its execution and settlement remains brokered on centralized platforms. Industry participants have long flagged concerns around attribution transparency, data ownership, and the friction-filled reconciliation cycles that connect advertisers, publishers, and agencies. Gency AI positions its platform as a step toward a new paradigm—a shift from “platform trust” to “protocol trust.”

By introducing on-chain verifiable credentials and automated revenue distribution mechanisms, the company envisions a workflow where ad impressions, conversion outcomes, and payments can be independently verified and settled using smart contracts. In practical terms, this could mean reduced settlement times and greater transparency for cross-border campaigns, with a built-in privacy layer that preserves user data while enabling auditable outcomes.

AI and blockchain–integrated technical architecture

Gency AI describes its network as a four-module stack designed to coordinate advertising actions in a privacy-conscious, verifiable manner:

Policy identity

On-chain policy identities create permissioned data usage boundaries, enabling transparent and traceable authorization management for data assets used in advertising campaigns.

ESQ privacy computing layer

The ESQ layer integrates technologies such as trusted execution environments (TEEs), private set intersection (PSI), and secure multi-party computation (MPC) to facilitate encrypted data processing. This enables advertisers and platforms to run analytics and optimizations without exposing raw data.

PSG clearing and settlement protocol

This component converts advertising actions and conversion outcomes into on-chain verifiable credentials and automatically triggers revenue distribution through smart contracts, delivering end-to-end verifiability of the ad economy’s financial flows.

AI optimization engine

Operating within an encrypted environment, the AI module powers campaign forecasting, audience targeting, and optimization, while supporting model training and attribution analysis without revealing underlying user data. The approach aims to balance rigorous privacy protection with practical operational efficiency.

Investor perspectives

Investors in the round argue that uniting AI automation with verifiable computing could fundamentally remodel the core infrastructure of digital advertising. By moving away from opaque, closed data platforms toward open, verifiable protocols, they see an opportunity to improve transparency, efficiency, and compliance in a space where regulatory expectations are tightening and data privacy rules are evolving globally.

As demand for AI-powered automated advertising grows, backers note that a trusted, verifiable, and autonomous ad network could become a defining trend for the industry. The funding round signals confidence that a protocol-first approach—rooted in cryptographic guarantees and on-chain governance—could become attractive to advertisers, publishers, and technology partners seeking greater interoperability and accountability.

About Gency AI

Gency AI frames itself as a sovereign advertising network designed for the agentic economy—an ecosystem where data ownership, permissions, execution, and settlement are programmable, verifiable, and user-controlled by default. Rather than relying on traditional adtech models built on opaque data aggregation and trust-based reporting, Gency AI envisions advertising as a verifiable coordination system. By combining cryptographic guarantees, on-chain policy enforcement, and measurable outcomes, the network aims to enable coordinated interactions among advertisers, publishers, AI agents, and users.

Bakkt has completed its acquisition of Distributed Technologies Research, a stablecoin payments infrastructure firm. 

The deal brings DTR’s agentic payments technology and compliance tools into Bakkt’s regulated institutional platform.

The company said the combined platform will support institutions and fintechs seeking faster digital payments. Bakkt aims to build a 24/7 digital settlement layer using stablecoin technology.

Stablecoin settlement becomes core focus

Bakkt said DTR’s AI-native engine will be added to its existing infrastructure. The company expects the integration to reduce reliance on traditional correspondent banking systems.

Bakkt CEO Akshay Naheta said, “The architecture of money movement rarely evolves at this level.” He added that the deal introduces stablecoin functionality as a bridge between legacy finance and digital assets.

At closing, Bakkt issued 11,316,775 Class A common shares to DTR’s beneficial holders. The company may issue up to 725,592 more shares tied to outstanding warrants.

The transaction was first announced in January. At that time, the deal involved 9.3 million shares. Bakkt also announced a corporate name change to Bakkt Inc. during the same period.

Bakkt stock recovers after earlier drop

Bakkt shares fell about 8% to $7.86 before the deal closed. The stock later recovered to $8.62 by Thursday’s market close.

The company has faced pressure in recent years. In 2024, the NYSE warned Bakkt over a possible delisting after its share price stayed below $1 for 30 days.

Bakkt was founded in 2018 and is majority-owned by Intercontinental Exchange. The firm has worked with major brands, including Starbucks and Mastercard. The DTR deal now places stablecoin payments at the center of its next growth plan.

SBI Holdings has started formal talks with Bitbank over a capital and business alliance. The planned transaction would make Bitbank a consolidated subsidiary of SBI.

The company said it plans to acquire shares after due diligence and internal approval. Details on timing, structure, and acquisition method will be discussed later.

The move comes soon after SBI Group absorbed Bitpoint Japan through a merger with SBI VC Trade in April 2026. The latest talks show SBI is moving quickly to expand its crypto exchange business.

Japan is also reviewing how crypto assets fit under the Financial Instruments and Exchange Act. That process may bring tighter rules for exchanges and crypto investment products.

Bitbank’s IPO path comes into focus

Bitbank had previously prepared for a possible Tokyo Stock Exchange listing by mid-2025. The company had also raised about 7 billion yen through a capital and business alliance with Mixi in 2021.

Mixi became a major shareholder with a 26.2% stake. SBI’s proposal may now draw attention to Bitbank’s listing plans and future ownership structure.

Bitbank expands crypto payment services

Bitbank has also moved further into crypto-linked payments. The exchange launched the “EPOS Crypto Card for Bitbank” with EPOS Card, the fintech arm of Marui Group.

The card allows users to settle monthly credit card bills using bitcoin held on Bitbank. It also offers 0.5% cashback in bitcoin, ether, or Aster.

Bitbank said the card is the first in Japan to allow credit card bills to be settled directly from crypto exchange balances. The service may later add more crypto payment options.

The possible SBI deal would combine Bitbank’s exchange brand and payments push with SBI’s wider financial network. Bitbank has also reported zero hacking incidents since launch, which may support its appeal to a larger financial group.

The US CLARITY Act has advanced toward enactment as the final text addressing stablecoin yields has been published, signaling a potential turning point in a long-running regulatory dispute between the banking sector and the crypto industry. The provisions are designed to deliver regulatory clarity while balancing competitive concerns with consumer incentives. According to Cointelegraph, Coinbase chief legal officer Faryar Shirzad urged lawmakers to “get CLARITY done” after Senators Thom Tillis and Angela Alsobrooks released the final language outlining how stablecoins may or may not accrue rewards.

Shirzad said the framework preserves the core ability for Americans to earn rewards that reflect real usage of crypto platforms and networks, while the negotiations secured tighter restrictions on yield-bearing features that could undermine traditional banking models.

The draft text, titled “SEC 404. Prohibiting interest and yield on payment stablecoins,” would bar crypto firms from paying any form of interest or yield solely for holding stablecoins, aligning stablecoin economics with the treatment of deposits in conventional banking. It explicitly limits such rewards to bona fide activities, aiming to prevent the sort of risk-free earnings that critics argue could distort competition with banks. Industry observers cited the provision as a central point of contention in the broader debate over whether stablecoin yields should be subject to stricter oversight or outright prohibition.

Industry participants have voiced mixed reactions. Mert Mumtaz, chief executive of Helius Labs, argued that the approach clarifies that “you don’t get risk-free yield on your dollars without using a bank,” underscoring concerns about market structure and consumer protection. The policy, while restrictive on yields, leaves room for incentive schemes tied to legitimate platform activity, a nuance that proponents say preserves user engagement without eroding banking stability.

Key takeaways

• The SEC 404 provision would bar paying interest or yield on stablecoins solely for holding them, with exceptions limited to rewards tied to bona fide activities on a platform.
• The final text represents a compromise intended to address competitiveness concerns raised by banks while preserving incentives for users who engage with crypto platforms.
• Market and industry reaction centers on whether the bill can proceed to markup and eventual enactment, with predictions about timing and likelihood continually shifting.
• Policymakers and industry insiders expect a formal markup by the Senate Banking Committee, potentially as soon as the week of May 11, signaling a path toward a full chamber vote.
• Public commentary from industry leaders emphasizes the transition from debate on yields to broader considerations of licensing, oversight, and cross-border regulatory alignment.

Context and regulatory framing: stablecoins, yields, and governance

The central issue in the CLARITY Act discussions has been the interaction between stablecoin economics and the banking system’s capital and liquidity requirements. By placing a clear line on yields tied to stablecoins, the legislation seeks to deter yield-bearing mechanisms that could emulate deposits or credit-like products outside traditional bank channels, while permitting rewards that reflect genuine platform usage. This distinction is intended to reduce regulatory ambiguity for crypto firms while preserving consumer protections and preserving fair competition with regulated financial institutions.

Legislative momentum and process implications

Industry observers view the publication of the final text as a significant milestone, potentially clearing the way for a markup and a formal vote. Cointelegraph notes that Senate Banking Committee proceedings could be scheduled imminently, with some commentators indicating the possibility of markup in the week of May 11. Among lawmakers, optimism is tempered by the likelihood of continued negotiations with banking interests, who have historically pressed for tighter restrictions on crypto yields and product features.

Traders on prediction platforms have begun pricing in the probability of legislative movement. Polymarket participants currently assign a roughly 55% chance of the CLARITY Act being signed into law in 2026, reflecting optimism about congressional action while acknowledging potential hurdles ahead. Commentators also point to ongoing advocacy from industry leaders who are urging timely markup and consideration of the broader bill’s provisions beyond the yield issue.

Regulatory and institutional implications for firms and markets

The stablecoin yield provisions are part of a broader governance framework that could influence licensing, oversight, and constitutional concerns around crypto markets. For exchanges, custodians, and banks, the draft text raises practical questions about how rewards programs will be structured, how to demonstrate bona fide activity, and how to ensure compliance with the prohibition on yield that is not tied to legitimate usage.

From a compliance perspective, the final text frames the boundary conditions for reward schemes and may necessitate changes to product design, disclosures, and customer communications. It also shapes the risk calculus for banks weighing crypto-related partnerships or onboarding stablecoin-linked products, given the perceived impact on competitive dynamics and liquidity management. The broader policy environment—encompassing licensure, anti-money laundering (AML) controls, know-your-customer (KYC) standards, and cross-border considerations—could be influenced by how the CLARITY Act ultimately is crafted and enacted.

Industry leadership has underscored the importance of clarifying the regulatory regime to support strategic planning, risk management, and compliance programs across banks, crypto platforms, and financial counterparties. As the debate shifts from whether yields should be allowed to how they should be governed, institutions are likely to adjust their internal controls, policy documentation, and external disclosures to reflect the evolving framework.

Looking ahead, the ultimate fate of the CLARITY Act will hinge on the timing and outcome of the markup process and the willingness of lawmakers to broker a final agreement that satisfies both financial stability concerns and the desire for market clarity. While the published text narrows the scope of permissible yields, it also consolidates a path toward definitive regulation, which could reshape how crypto-native finance interacts with traditional banking systems.

As discussions progress, market participants and compliance teams should monitor committee calendars, stakeholder testimonies, and potential amendments that could alter the bill’s trajectory or scope. The next steps will be critical in determining whether the promised clarity translates into a durable regulatory framework for the U.S. crypto sector.

Solana-based DeFi yield protocol Carrot has announced a permanent shutdown after losses tied to the Drift Protocol exploit left it unable to continue operations.

According to a statement posted by Carrot on X on Thursday, the April 1 attack on Drift proved “catastrophic” for the protocol, forcing the team to wind down services and set May 14 as the deadline for users to withdraw remaining funds. The team said it will continue assisting recovery efforts linked to Drift and distribute assets once they are recovered.

“We are setting May 14th as the deadline to withdraw any remaining funds from Boost, Turbo, and CRT before we will then begin to deleverage the system. Your deposited funds are still yours, but all leverage will be reduced to zero, freeing up all liquidity for CRT redemption,” the team said.

Integrated with Drift’s infrastructure, Carrot relied on its liquidity pools to generate yield, which left it exposed when the exploit drained a large portion of Drift’s total value locked. DefiLlama data shows Carrot’s TVL fell from about $28 million before the incident to $1.99 million, a drop of roughly 93%.

Drift Protocol said on April 5 that the exploit followed months of preparation, during which attackers built trust with contributors through in-person meetings and online contact before delivering malicious tools. External estimates placed losses from the attack at about $280 million, while Drift described the campaign as organized and backed by significant resources.

According to Drift’s review, contact with the attackers began around October 2025, when individuals posing as members of a quantitative trading firm approached contributors at a crypto conference and later maintained relationships across multiple industry events. The exchange said those interactions allowed the group to gain trust before compromising devices and executing the exploit.

Drift added it has “medium-high confidence” that the same actors were involved in the October 2024 Radiant Capital breach, which resulted in about $58 million in losses and involved malware distributed through Telegram.

The impact has extended beyond Carrot. Projects connected to Drift, including Gauntlet, PrimeFi, and Elemental DeFi, have also reported disruptions following the exploit.

DefiLlama data shows that April recorded nearly $630 million in crypto losses across 25 incidents, making it the largest month for exploits since February 2025, when losses reached $1.47 billion. The $293 million attack on Kelp remains the biggest exploit of 2026 so far, followed by the Drift breach at roughly $285 million, with both incidents accounting for more than 90% of April’s total losses.

Brazil’s central bank has barred virtual assets from settlement inside regulated international payment rails. 

The rule applies to eFX services, which cover certain cross-border payments and transfers. Banco Central do Brasil published Resolution BCB No. 561 on Thursday. The measure updates rules for payment providers operating under the country’s foreign exchange framework.

The new rule says payments or receipts between an eFX provider and a foreign counterparty must use foreign exchange transactions. Providers may also use movement in a non-resident Brazilian real account.

The resolution bans the use of virtual assets for those payments and receipts. This means crypto and stablecoins cannot settle transactions inside the regulated eFX channel.

Rule is not a full crypto ban

The measure does not ban crypto transfers across Brazil. It only blocks crypto settlement within the supervised eFX framework.

Transitional rules also apply to firms not yet listed as approved eFX providers. These companies may continue operating if they seek central bank approval by May 31, 2027.

However, they must follow the same settlement rule. Their payments and receipts cannot use virtual assets.

Stablecoin activity draws closer review

Brazil has increased oversight of crypto-linked payment flows as stablecoin use grows. The central bank has been adding virtual assets to its financial and foreign exchange rulebook.

In November 2025, regulators set new rules for virtual asset service providers. These included authorization requirements and rules for crypto services tied to the foreign exchange market.

BCB Governor Gabriel Galipolo previously said crypto use had risen in Brazil over recent years. He said about 90% of flows were linked to stablecoins, raising concerns over taxation, money laundering, and backing.

The central bank has also reviewed stablecoins issued outside its supervision. In a technical note to Congress, it warned that such tokenscould face bans or strict conditions in Brazil.

The note said real-denominated stablecoins issued beyond BCB oversight may affect regulatory equality and monetary sovereignty. It also said foreign-currency stablecoins may raise concerns over capital flows and payment system fragmentation.

South Korea’s Seoul Administrative Court has granted Bithumb a temporary reprieve from a six-month suspension, allowing the exchange to continue operating while the case proceeds.

According to Yonhap News Agency, the court’s 2nd Administrative Division under Judge Gong Hyeon-jin approved the stay on Thursday, pausing enforcement of a sanction imposed by the Financial Intelligence Unit, an anti-money laundering body under the Financial Services Commission.

Regulators had issued the suspension notice in March, stating that Bithumb failed to meet AML obligations that required proper identity verification of users. The penalty targeted new customers by restricting external crypto deposits and withdrawals, a move that would have limited onboarding activity if enforced.

The Financial Intelligence Unit also imposed a fine of 36.8 billion won, about $25 million, after identifying roughly 6.65 million cases where user identities were not properly verified, according to The Korea Herald. Disciplinary measures were also directed at CEO Lee Jae-won as part of the same action.

Court filings show Bithumb challenged both the suspension and its execution, submitting a lawsuit and a stay request on March 23. Enforcement had already been paused during judicial review, and the latest decision keeps restrictions on hold until a final ruling is issued, leaving day-to-day operations unaffected for now.

Company statements cited by Yonhap indicated that the suspension could slow new user growth and weigh on business activity, while the FIU maintained that any revenue impact would be limited. Payment of the fine remains pending more than four weeks after the deadline, despite a 20% early settlement discount offered by the regulator, The Korea Herald reported.

“We plan to faithfully present our position throughout the remaining legal proceedings,” Bithumb said, according to local media reports.

Regulatory pressure deepens after operational missteps

Regulatory scrutiny has intensified following a series of operational issues, including a February payout error that triggered investigations into internal controls. During a promotional campaign, the exchange mistakenly distributed a theoretical 620,000 BTC instead of 620,000 won, an error that led to unintended credits reaching external wallets.

According to local outlet Chosun Biz, Bithumb recovered about 99.7% of the assets, while the remaining portion was addressed using company reserves after some users sold the funds. Legal action has since been initiated against certain users who refused to return the assets, with provisional seizure requests filed to freeze holdings ahead of civil proceedings.

Authorities responded to the incident by tightening oversight across the sector. The Financial Services Commission directed exchanges to strengthen real-time monitoring of large transactions after an emergency inspection identified vulnerabilities in automated settlement systems.

Alongside the regulatory challenges, Bithumb has pushed its planned initial public offering timeline to 2028, citing ongoing scrutiny, while investigations by the Financial Supervisory Service continue to examine risk management practices tied to the payout error.

Bitcoin funding rates stayed negative for 46 days, the longest since 2023, forcing shorts to pay longs daily.

Bitcoin shorts didn’t just lose money when the squeeze hit; they had been losing money long before it arrived. 

For 46 consecutive days, funding rates stayed negative, forcing short traders to pay longs simply to hold their positions. 

According to CoinDesk, that stretch marked the longest negative funding period since 2023. When Strategy’s $2.54 billion purchase and Trump’s Iran ceasefire extension finally landed, the damage was already done. The catalysts were just the final blow. 

Negative funding rates made holding Bitcoin shorts expensive

Funding rates are periodic payments exchanged between long and short traders in perpetual futures markets. When rates turn negative, shorts pay longs to keep positions open. That cost runs on a clock, not a chart.

As Leverage.Trading outlines in its breakdown of funding rates, at 20x leverage, a 0.05 percent funding charge on notional exposure equals one percent of available margin per settlement. 

With three settlements occurring daily, that figure compounds fast. Leverage. Trading’s educational breakdown of funding mechanics lays out exactly how quickly those charges erode a position.

Over 46 days, that steady drain ate through an estimated 30 to 40 percent of the short margin before any major catalyst appeared. Directionally wrong traders were also paying for the privilege of staying wrong, every single day.

Shorts were already near liquidation before the news hit

By the time April’s headlines arrived, short positions were operating on a thin margin. Strategy announced the purchase of 34,164 BTC for $2.54 billion, sending Bitcoin climbing to $77,500, per CoinDesk. 

Trump’s Iran ceasefire extension added further risk appetite to the session. Both events hit close together.

The market didn’t need much upward pressure to trigger liquidations at that point. Margin had already been quietly stripped away over six weeks of negative funding. 

What looked like a sudden squeeze from the outside was actually the final stage of a much slower process.

Anton Palovaara of Leverage.Trading described it directly: 

“Forty-six days of negative funding doesn’t show up on a chart, but it shows up in your margin. By the time the ceasefire news hit, a lot of shorts were already running on fumes.” 

They added,

“The liquidations happened fast because the margin was already gone. The headline was just the match.”

Over $427 million in shorts liquidated as margin ran out

Finance Magnates reported more than $427 million in short liquidations across recent sessions. That number reflects how much trapped leverage had accumulated during the extended negative funding window. 

Shorts had been positioned for a price drop that kept not materializing.

On April 24, crypto trader CryptoBoss posted a breakdown connecting the setup to historical precedents. He noted that 50 days of deeply negative funding near the $15,500 bottom in 2022 preceded a 48 percent rally to $23,000. 

A similar pattern played out during the 2021 China mining ban, where roughly 45 days of negative funding near $29,000 preceded a rally to $48,000.

The 2025 setup matched those conditions closely, with 46 days of negative funding while price ground steadily higher, not lower.

Coinbase premium and the $80k test signal: What comes next?

Alongside the liquidations, Coinbase Premium posted its longest bullish streak since October’s $126,000 high, per CoinDesk. 

That streak pointed to consistent spot buying pressure from U.S.-based investors running parallel to the derivatives squeeze. Spot demand and a short-saturated futures market rarely stay in tension for long.

Finance Magnates noted that Bitcoin is now testing the $80,000 breakout level. 

Whether that level holds depends on continued spot support and how remaining leveraged shorts respond. Positions that survived the squeeze are still carrying funding risk if rates stay elevated.

The 46-day bleed was not visible on most price charts. However, it showed up clearly in margin balances, and ultimately in the liquidation data that followed.