Mistral AI, Digital Realty Partner to Scale European AI Infrastructure
The French startup has secured 10 megawatts of compute at Digital Realty’s Paris South campus.
Fundraising
Hut 8 AI landlord data center strategy turns Bitcoin collateral into bridge capital
Hut 8 is pushing even further into AI infrastructure than most other Bitcoin miners are. Its latest disclosures show a company using power access, data center leases, project debt, and BTC-backed liquidity to build the financing stack for that move.
The company’s latest disclosures put numbers around that transition. Hut 8 reported $16.8 billion in triple-net, take-or-pay contracted lease revenue across two hyperscale AI campuses, then separately refinanced a $200 million Bitcoin-backed credit facility with FalconX.
The new facility cut the fixed rate to 7.0% from 9.0% and unencumbered roughly 3,300 BTC from the prior collateral package.
Taken together, the disclosures show a miner identity changing into something closer to an infrastructure landlord. Hut 8 is turning megawatts, lease commitments, project debt, and Bitcoin holdings into the machinery for a business that depends less on mining alone.
The result is a case study with more substance than a generic AI pivot. Hut 8 is showing a funded path into data center infrastructure, though the model still needs operating proof. The test is whether contracted AI cash flows arrive on schedule and become durable enough that Bitcoin collateral becomes a bridge instead of a recurring source of balance-sheet dependence.
The lease base turns power into finance
The strongest number in Hut 8’s first-quarter disclosure sits outside the Q1 income statement: $16.8 billion of contracted lease revenue across River Bend and Beacon Point, covering 597 MW of AI data center capacity.
Hut 8 generated $71 million of revenue in the first quarter, including $66 million from Compute, and posted a $253 million net loss that included $295 million of primarily unrealized digital-asset losses.
The $16.8 billion figure represents long-term contracted lease value that Hut 8 is presenting as the foundation for a different kind of business.
The pieces are specific. Hut 8’s Beacon Point lease added 352 MW of IT capacity and $9.8 billion of base-term value. Its earlier River Bend lease added 245 MW and $7 billion of base-term value, with Google providing a financial backstop for the base lease term.
Hut 8 is commercializing scarce power and data center capacity under long-term lease structures. The appeal comes from contracts and power access rather than a token, a cloud slogan, or a vague compute promise.
Triple-net and take-or-pay terms are designed to make those cash flows more financeable because the tenant obligation is less tied to day-to-day mining economics.
Hut 8’s disclosures line up across four moving parts:
| Model component | Hut 8 evidence | Reader impact | Risk still live |
|---|---|---|---|
| Power and sites | 597 MW of contracted AI data center capacity across two campuses | Turns miner infrastructure into leaseable digital infrastructure | Delivery, interconnection, construction, and tenant concentration |
| Contracted demand | $16.8 billion in base-term contracted lease revenue | Creates a financing story beyond hashprice exposure | Lease value depends on execution over long timelines |
| Project finance | $3.25 billion River Bend notes, non-recourse to Hut 8 | Reduces the need to fund all growth from equity or BTC sales | Large projects still carry cost, schedule, and market risks |
| Bitcoin balance sheet | $200 million FalconX BTC-backed facility and 3,300 BTC unencumbered | Gives liquidity without immediately selling coins | Collateral value still moves with BTC |
Hut 8’s AI transition has more to it than most, but each component still carries a different kind of risk.
The leases reduce some revenue uncertainty. The bond financing reduces some parent-level funding pressure. The Bitcoin facility improves liquidity. Still, all three leave Hut 8 with the task of building, delivering, and operating infrastructure for customers whose requirements differ from Bitcoin mining.
Bitcoin becomes bridge capital
The FalconX refinancing is the clearest sign that Bitcoin is becoming part of the financing machinery rather than only the asset being mined.
The full Hut 8 release distributed through Nasdaq described the facility as a 364-day Bitcoin-backed loan with limited recourse to pledged BTC, a no-rehypothecation covenant, fixed loan-to-value thresholds, and no loan-to-value ratchet triggered by declines in Bitcoin’s price.
Those terms blunt part of the obvious criticism. The deal improves the terms of a miner’s coin-backed borrowing instead of worsening them to chase a new market.
Hut 8 lowered its fixed cost of debt by 200 basis points and increased Bitcoin held outside collateral covenants. The release valued the newly unencumbered coins at roughly $260 million as of May 1, 2026, giving Hut 8 more balance-sheet room without selling the asset.
That makes the facility a better tool, but not a risk-free one.
Hut 8’s own balance sheet shows why the distinction is important. Its 10-Q said the company held about 16,332 BTC as of March 31, 2026, including about 9,311 BTC held by Hut 8 and about 7,021 BTC held by American Bitcoin.
The aggregate fair value was about $1.11 billion, based on approximately $68,222 per BTC. The same filing tied the first-quarter digital-asset loss to Bitcoin’s decline during the period.
Today, Bitcoin trades near $75,782 on CryptoSlate’s price page, down 2.1% over 24 hours and roughly 40% below its October 2025 all-time high. The market-price channel is the relevant risk.
Bitcoin can provide liquidity without a sale, but the borrowing value, covenant comfort, and refinancing backdrop still depend on the asset’s market behavior.
That is why the AI landlord strategy cannot be separated from the Bitcoin treasury strategy. If AI leases produce reliable cash flows, BTC collateral can be transitional capital. If delivery slips, financing markets tighten, or Bitcoin weakens at the wrong time, the same collateral can keep the pivot tied to the volatility it was meant to escape.
The miner label is becoming less useful
Earlier coverage of miners’ AI pivot showed the broader identity split facing the sector. Miners are moving toward AI and high-performance computing because power access, cooling infrastructure, land, interconnection work, and industrial operations can be worth more under contracted dollar revenue than under compressed mining margins.
Hut 8 fits that broader sector shift. Public miners built businesses around converting power into BTC, and AI data center demand is now giving some of them a second possible use for the same physical footprint.
The difference is that AI customers do not buy the same thing the Bitcoin network buys. Mining can tolerate interruption when economics or grid conditions change. AI tenants want uptime, delivery certainty, dense power, cooling, network architecture, and creditworthy execution.
A miner with megawatts still has to become a hyperscale landlord. It has to turn a power position into infrastructure that lenders and tenants will treat as dependable.
Hut 8’s disclosures show both sides of that transition. The company describes itself as an energy infrastructure platform integrating power, digital infrastructure, and compute. It also still reports digital-asset losses, BTC holdings, and exposure to mining economics.
Some Compute revenue and BTC holdings are held by American Bitcoin, a consolidated subsidiary, making Hut 8’s strategy less straightforward than a clean exit from mining.
That complexity is part of the shift. The market is watching whether miners can stop being pure BTC proxies without losing the balance-sheet optionality that made their treasuries valuable in the first place.
The strongest argument in Hut 8’s favor is that the AI pivot uses more than Bitcoin-backed debt. The company said it closed $3.25 billion of fully amortizing 16.5-year investment-grade senior secured notes to finance River Bend.
Hut 8 described the financing as non-dilutive and non-recourse to Hut 8, with loan-to-cost increasing to about 95%.
That weakens the crutch argument. If project-level debt funds the campus and long-term leases support the debt, then Bitcoin collateral is one part of the structure rather than the whole. It is a liquidity tool alongside project finance and contracted revenue.
The caution is that the financial structure still has to become operationally sound. River Bend is still advancing toward delivery, Beacon Point still has to be built out, and the company still has to convert an 8,375 MW development pipeline into real contracted capacity.
Hut 8 also warned investors about risks tied to data center construction, financing, power expansion, permitting, supply chains, technical challenges, and market conditions.
Hut 8 is showing that miners can finance a route into AI infrastructure when they have scarce power, credible tenants, project-finance access, and a Bitcoin balance sheet lenders will underwrite. It has yet to show that the route is self-sustaining.
The next test is whether AI infrastructure cash flows become strong enough to push Bitcoin collateral into the background. If they do, Hut 8’s BTC-backed financing will look like bridge capital for a miner that successfully monetized its power footprint.
If they fail to do so, the pivot will remain tethered to the same balance-sheet asset that made the strategy possible in the first place.
The post Hut 8 AI landlord data center strategy turns Bitcoin collateral into bridge capital appeared first on CryptoSlate.
US and China pour record capital into AI as funding race intensifies on both sides
The artificial intelligence sectors in both the United States and China are being flooded with unprecedented capital. However, the AI funding is now becoming a competition between the two largest economies.Â
Global AI startups raised $255.5 billion in Q1 2026 alone. Meanwhile, Chinese AI ventures separately pulled in over 110 billion yuan ($16.2 billion).
How is the Chinese government fueling the AI boom?Â
Pan Xiaodong, the secretary general of the Ministry of Science and Technology, made some huge announcements at a Beijing press conference back in February. He mentioned that the Chinese government had launched a national venture capital guidance fund. It focuses on early-stage, small, long-term, and hard-tech enterprises. This includes AI, semiconductors, and advanced manufacturing. The estimated total scale of the fund is around 1 trillion yuan ($144.45 billion).
Investors linked to the Chinese government reportedly participated in more than 140 AI deals in 2025. It is a huge jump compared to the 10 deals per year seen before 2018. The authority has also joined hands with financial institutions and local governments. This is done to establish various funds totaling over 350 billion yuan. This includes tech-industry integration funds and secondary market funds.
DeepSeek will reportedly have its first outside investment round led by China’s Integrated Circuit Industry Investment Fund (the “Big Fund”). The AI company came into the light for its cost-efficient models. The startup’s valuation surged from $10 billion to $20 billion in April. It later reached an estimated $45-$50 billion by early May.Â
Other investment rounds backed by the government include Moore Threads, a Beijing-based GPU designer, which raised $720 million at a $4.1 billion valuation in February 2025. Moonshot AI secured $700 million at a $10 billion valuation in January 2026, while StepFun reportedly raised $717 million.
Linkerbot, a robotic-hand startup, is targeting a $6 billion valuation backed by Ant Group and Bank of China Asset Management, while Unitree Robotics has filed for a Shanghai listing seeking up to $7 billion.
The US and China are protecting their industries from outside influence
Washington banned American investors from backing Chinese AI and chip companies back in January 2025. In late April, China applied its own version of the same restriction. The National Development and Reform Commission instructed Moonshot AI, StepFun, and ByteDance not to accept US capital without explicit government clearance after Meta acquired Manus for $2 billion.Â
Cryptopolitan previously reported that the Trump administration accused Chinese labs of “industrial-scale” distillation of American AI models. Distillation is a method where a developer uses data from a larger AI model to train a smaller one.Â
The White House memo outlined four measures to stop this, including sharing intelligence on distillation tactics and coordinating defenses with US AI companies.
Anthropic also previously accused DeepSeek, Moonshot AI, and MiniMax of exploiting its models.Â
Despite the tensions, the total deal activity for the Chinese private equity market reached 2,568 transactions worth 234.4 billion yuan in Q1 2026, and foreign-currency deals in China more than doubled year-on-year to 210 in the same period. The disclosed investment value jumped by 495% to 67.3 billion yuan ($9.9 billion).Â
In the US, investments from the first quarter of 2026 surpassed the $254.4 billion deployed across all of 2025. OpenAI, Anthropic, and xAI accounted for more than two-thirds of the total.
Despite the differences in the US and Chinese governments’ approaches, they are both producing results. Chinese large-model companies have shortened iteration cycles to under three months by 2026, and a Stanford University report suggested that the performance gap between top US and Chinese AI models has “effectively closed.”
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It’s free.
Cardano founder warns network could lose its scientists in Input Output’s $46.8M funding vote fails
Cardano could lose a core group of scientists if Input Output fails to secure treasury funding for a slate of research and infrastructure proposals that are still awaiting approval.
Last month, Input Output, the development firm behind the Cardano network, revealed that it was seeking $46.8 million to finance its operations for the 2026 development cycle.
However, the funding request has encountered significant resistance as the May 24 voting deadline approaches. A look at the major proposals shows that they face weak support, heavy abstentions, and large blocs of votes left uncast, leaving the network’s technical future hanging in the balance.
The escalating tension prompted a stark warning from Cardano founder Charles Hoskinson, who cautioned that failing to approve the treasury withdrawal could trigger an exodus of top talent and potentially shutter the network’s flagship research laboratory.
Trailing the threshold
The $46.8 million request is fractured across several specialized workstreams, each requiring a 67% ratification threshold from the network’s DReps. As the voting window narrows, almost none are on track for approval.
The largest line item is the Cardano Maintenance Initiative, an ask of more than 62.1 million ADA designed to cover continuous core maintenance from the third quarter of 2026 through the first quarter of 2027.
The proposal covers nine functional areas, including bug fixing, disaster recovery, mainnet monitoring, and incident response.
Despite its critical nature, described by developers as the protective foundation that ensures network uptime and security, the proposal currently holds just 46.58% affirmative votes. A massive 9.25 billion ADA is logged as abstaining, while 45.61% of voting power has yet to weigh in.
Other critical infrastructure proposals are faring even worse. A 10.4 million ADA request to fund Layer 2 scalability solutions, including a data availability solution and the launch of Midgard, the network’s first permissionless optimistic rollup, sits at just 16.08% approval.
Layer 2 architecture is widely considered the only viable path to achieving the 10,000-plus transactions per second and sub-cent fees necessary to attract high-frequency decentralized finance and artificial intelligence micropayments in the current cycle.
A $2.95 million pitch to build “Pogun,” an end-to-end Bitcoin liquidity and credit engine meant to capture a share of the $1.5 trillion Bitcoin asset class, is polling at 19.04% in favor, weighed down by 24.15% active rejections and overwhelming abstentions.
Cardano’s development and tooling on the brink
The hesitation among DReps extends to proposals that target developer experience and smart contract capabilities, areas where Cardano has historically struggled to gain ground against rivals like Ethereum and Solana.
A 13 million ADA proposal to bring automated formal verification to decentralized applications has performed the best thus far, but still trails the needed supermajority at 57.79%.
The initiative aims to extend the Blaster verification tool across multiple smart contract languages, lowering the barrier for developers to mathematically prove their code’s correctness.
Similarly, an 11.8 million ADA request to expand the capabilities of Cardano’s native Plutus smart contract language, aimed at reducing script costs and improving expressiveness, is hovering around 32% approval.
A separate 3.6 million ADA pitch explicitly designed to boost developer growth by 30% over the next year, by streamlining onboarding and documentation, sits below 30%.
Also struggling is Project Cayley, a 7.92 million ADA initiative aimed at decentralizing data indexing. Currently, indexing the entire Cardano blockchain dataset requires massive computational resources, a burden that will only grow as the network scales.
Project Cayley introduces decentralized slice indexing, allowing node operators to index only specific portions of the chain. This lowers the barrier to entry and prevents data-serving infrastructure from centralizing around a few well-funded providers.
Yet, the proposal is languishing at 13.83% approval, with nearly 30% of active voters rejecting it outright.
Finally, a 13.1 million ADA proposal that would introduce Babel Fees, allowing users to pay transaction costs in any native asset, such as stablecoins, rather than holding ADA, has garnered nearly 60% support but remains shy of the 67% hurdle.
The upgrade is widely viewed as essential for removing onboarding friction for new users.
A clash over the “science coin”
For years, Cardano has staked its reputation on rigorous, peer-reviewed academic research and formal methods. This methodical approach has occasionally drawn criticism for moving too slowly, but it has cultivated a fiercely loyal community.
However, the stakes of the treasury vote also appear to be impacting a research-focused proposal called “Cardano Vision 2026: Human Centered, Scalable, Post Quantum Secure – IO Research.”
This proposal seeks nearly 33 million ADA tokens, approximately $8 million, to “preserve Cardano’s evidence-based approach” and “ensure that research outputs translate more reliably into measurable ecosystem growth.”
However, YUTA, a Cardano Drep, stated that the “proposal is a mix of a waste of funds and a potentially excellent proposal for Leios and quantum resistance research.”
In response, Hoskinson stated:
“We are deeply saddened that some Japanese dReps voted against our research proposal…If this proposal does not pass, we want the entire Japanese community to fully recognize that Cardano will lose its scientists, and our lab will be forced to close.”
Hoskinson emphasized that building the organization’s research apparatus took more than a decade and hundreds of millions of dollars.
He warned against dismantling the world’s strongest cryptocurrency research group over “piecemeal funding support,” asserting that the organization’s scientists would simply leave for ecosystems offering greater certainty and professional respect.
He added:
“This doesn’t have anything to do with me. This has to do with destroying the entire core of our ecosystem. Cardano is the science coin. That’s our brand. We spent hundreds of millions of dollars and a decade to earn the right to say that. You don’t throw it away.”
As of press time, the proposal has secured only a 13% support, and its voting is expected to close on June 8.
Cardano’s decentralized governance experiment tested
The funding friction is particularly notable given that Input Output explicitly scaled back its financial demands for this cycle.
The 2026 treasury request represents a nearly 50% reduction from the previous year’s budget, signaling an intent to transition the ecosystem toward long-term self-sufficiency.
Yet, even this tightened fiscal belt has failed to win immediate favor from the newly empowered governance body.
The gridlock illustrates the double-edged sword of Cardano’s decentralized governance era. By placing the keys to the treasury directly in the hands of token holders and elected DReps, the ecosystem has achieved a level of financial decentralization rarely seen in major blockchain networks.
However, the current voting impasse highlights the vulnerability of that model. With large swaths of voting power either abstaining or remaining dormant, funding for vital infrastructure is essentially frozen.
For Input Output, the proposals represent a bare-minimum operational baseline to keep the network secure and competitive. For the DReps, the vote is an exercise in budget discipline and accountability, requiring the software laboratory to justify every dollar.
If the proposals fail to cross the 67% threshold by the May 24 deadline, Cardano faces an unprecedented scenario. Without the requested treasury disbursements, key upgrades could be delayed, and essential maintenance operations may be forced to scale back.
More critically, as Hoskinson warned, the talent pipeline that built Cardano’s sophisticated, academically rigorous architecture could begin to fracture, fundamentally altering the ecosystem’s trajectory.
The post Cardano founder warns network could lose its scientists in Input Output’s $46.8M funding vote fails appeared first on CryptoSlate.
China rejects claims it blocks startup funds but foreign capital continues to exit Chinese data centers
China has denied claims that it is discouraging local tech companies from taking foreign investment, despite continued withdrawals by international investors from major sectors.
Li Chao, a National Development and Reform Commission official, stated on May 22 that the government has never instructed Chinese IT companies to steer clear of foreign funding.
He continued by saying that China is in favor of international collaboration and intends to continue opening its economy to foreign companies and investment.
Meta deal blocked amid security concerns
According to reports, Chinese regulators had discreetly instructed local tech companies to refuse US money unless they first obtained government approval.
ByteDance and AI startups Moonshot AI and StepFun were among the businesses listed.
Concerns grew after the commission said in late April that it had blocked Meta Platforms from acquiring the $2 billion AI startup Manus.
Although Manus is registered in Singapore, its products are made in mainland China.
Citing national security risks, the regulator ordered the deal to be canceled.
Following this, Manus is now reportedly trying to raise nearly $1 billion from external investors to comply with Beijing’s requirement to reverse the takeover.
This unofficial guidance, which falls between official policy and administrative counsel, is frequently referred to as “window guidance” in Chinese regulatory practice.
In addition to monitoring cross-border transactions for threats to national security, the commission is in charge of the Negative List for Market Access, which places restrictions on foreign investment in specific industries.
According to Li, foreign investment must abide by Chinese law and not jeopardize national security or other interests.
Beijing maintains that it is not shutting down the market, but the national security approval procedure is still erratic, making it challenging for foreign investors to determine what degree of participation is appropriate.
International investors are wary of regulatory risks because the commission’s actions have sent mixed signals, despite its claims to support international investment.
Foreign investors exit China’s data center sector
After years of investing heavily in China’s cloud computing sector, foreign private equity firms are now divesting from the data center industry.
Growing political and regulatory pressures are making it increasingly hard for overseas investors to maintain control over digital infrastructure.
Princeton Digital Group, which has backing from Warburg Pincus, is putting its China assets up for sale in a deal that could bring in as much as $1 billion, according to three sources.
A sale of the group, which owns data centers in six Chinese cities, would basically end a ten-year effort by global buyout firms to invest directly in China’s digital infrastructure.
Major private equity firms like Bain Capital, Warburg Pincus, and The Carlyle Group began making significant investments in China’s data center sector in 2017.
With the expectation of steady, infrastructure-like long-term returns, they were drawn to the growing demand from cloud providers associated with Alibaba, Tencent, and ByteDance.
However, Beijing’s tougher cybersecurity and data management regulations have made foreign ownership of crucial digital infrastructure more delicate and challenging, even though China’s cloud industry is still expanding.
Several international investment funds have already exited due to this shift, selling their stakes to domestic investors.
Last year, Bain sold its Chinese data centre assets for $4 billion to a consortium led by Shenzhen Dongyangguang Industry, while keeping Bridge Data Centers outside China.
Similarly, Carlyle has gradually reduced its exposure over the past two years after investing in VNET Group in 2020.
It did so through refinancing supported by state-backed funds, and fully exited when CATL acquired the company.
Global private equity firms are shifting billions of dollars into other Asian economies, including Malaysia, Japan, and India, as they withdraw from China’s data center industry.
These nations are becoming more appealing for long-term investment due to strong AI-driven demand and more stable legislation.
Despite China’s assertions that it welcomes international investment, stricter cybersecurity laws and restrictions on IT transactions have alarmed foreign businesses.
Many firms now see owning sensitive infrastructure in China as too risky and are moving their investments elsewhere.
Don’t just read crypto news. Understand it. Subscribe to our newsletter. It’s free.
a16z emerges as big spender, directing cash into Anduril, US mid-terms
Rather than wait for the regulatory environment to shape itself, Andreessen Horowitz (a16z) has gotten directly involved in funding AI technology, lobbying for the rules, and helping choose the people who write them.Â
On the same day that reports surfaced that a16z had become the largest donor in the 2026 U.S. midterm election cycle, AI defense technology Anduril also revealed that it received billions in investments from the firm.Â
$115 million in political donations from a16z
Andreessen Horowitz’s checkbook (a16z) has been busy as one of the largest political donors of the 2026 midterms, also confirmed a massive investment into defense tech company Anduril. The company’s autonomous weapons factory is now operational in Ohio.
Cryptopolitan reported that the firm’s crypto division announced its fifth fund earlier in May, raising $2.2 billion for Web3 startups. Through these five funds, a16z’s total dedicated crypto capital is about $10 billion.Â
According to a New York Times analysis, a16z and its co-founders Marc Andreessen and Ben Horowitz have donated more than $115 million in disclosed federal contributions to midterm election efforts so far.Â
This puts them ahead of Democratic megadonor George Soros, who has donated $102.9 million, and Tesla (NASDAQ: TSLA) CEO Elon Musk, who has put in about $85 million. The firm gave more than $23 million to the political action committee Fairshake and its affiliates to support crypto-friendly candidates during the 2026 midterm elections.Â
On the investment side, Anduril announced a $5 billion Series H funding round led by Thrive Capital and a16z, bringing its valuation to $61 billion.Â
Cryptopolitan previously reported that Andreessen Horowitz raised over $15 billion across multiple funds in its 2025 fundraising rounds and currently has over $90 billion in assets under management. The company raised $6.75 billion for growth, $1.7 billion each for apps and infrastructure, and $1.176 billion for American Dynamism, its national security and government tech vertical.Â
Why is a16z spending big on elections?
Marc Andreessen has stated he is a “single-issue voter,” meaning he casts his ballot purely on what is good for tech startups. After feeling hostility from the Biden administration on crypto and AI policy, a16z backed the Trump campaign and deepened its presence in Washington.Â
The firm donated $47.5 million to crypto super PAC Fairshake and $50 million to Leading the Future, a pro-AI super PAC. Andreessen personally spends significant time advising the Trump administration on tech policy.
On defense, Anduril’s investor letter argues that the U.S. defense base is built around expensive, long-cycle platforms like $3 billion warships rather than cheap, mass-produced autonomous systems.Â
The $5 billion raise is aimed at scaling Arsenal-1 to fix that. By the end of 2026, the Ohio factory will be producing Fury combat drones, Roadrunner interceptors, Barracuda cruise missiles, and a classified platform. At full capacity, the Fury production line alone can produce 150 aircraft per year, running three shifts daily.Â
The Trump admin has been good for U.S. tech firms. Anthropic is in early talks with investors to raise at least $30 billion in fresh investments at a valuation of more than $900 billion. The round is expected to close by the end of this month, despite the fact that the deal is not finalized and no term sheet has been signed.Â
Anthropic previously closed a $30 billion Series G in February 2026, led by GIC and Coatue, valuing the company at $380 billion. At the time of Series G, Anthropic’s run-rate revenue had reached $14 billion due to demand from enterprises and developers.Â
Despite April 2026 being the lowest monthly crypto raise in a year at $662.4 million, a16z has pressed ahead with its $2.2 billion Crypto Fund 5. The figure is notably smaller than the $4.5 billion raised in 2022. Chris Dixon, a16z crypto founder, has said the focus is on “practical applications” like stablecoins, which continue to grow even in downturns.
If you’re reading this, you’re already ahead. Stay there with our newsletter.
Forgd launches liquidity transparency — and it could change how every token launches
Market infrastructure · Token launches · Web3 capital markets
Forgd launches liquidity transparency — and it could change how every token raises capital
For years, crypto founders have been choosing market makers based on reputation and promises. Forgd is ending that with the first dataset that lets projects compare firms head-to-head using real historical performance.
1,000+
Projects on platform
30+
Market makers integrated
500+
TGEs powering benchmark data
Token launches in 2025 were brutal. Altcoins launched, spiked, and collapsed — often within hours. Most founders blamed macro. Most missed the real problem: they were flying blind on market making, picking firms based on anecdotal case studies and aggressive pitch decks, with no way to verify whether any of it was true.
Forgd, a Web3 investment bank and capital markets advisor, is changing that. The firm has spent years quietly building what is now a first-of-its-kind dataset: standardized, historical market maker performance data drawn from live engagements across hundreds of token projects. That data is now powering a new layer of the Forgd platform — letting any project compare market makers head-to-head on the metrics that actually matter.
The problem that’s been hiding in plain sight
Market makers are integral to every token launch. They’re responsible for providing liquidity — the resting orders on both sides of a trade that let buyers and sellers transact without massive price swings. Choose the wrong one, or structure the engagement badly, and a launch can crater in its first hours and never recover.
But for most of crypto’s history, founders have had almost no information to make that choice. Market makers submit proposals with tight spreads and big depth commitments. They share case studies. They reference successful projects. And because there’s no standardized record of what actually happened post-launch, founders can’t verify any of it.
“For years, market-maker selection in crypto has been driven more by perceived reputation than by verifiable performance. Forgd forces market makers to earn — and continually re-earn — their reputation through actual results.”
Shane Molidor, Founder & CEO, ForgdForgd’s platform monitors live market maker engagements across its user base, collecting three core metrics for every engagement: bid–ask spreads, order-book depth at various thresholds from mid-price, and fill order volume broken down by maker and taker activity. Over hundreds of engagements, these data points accumulate into a longitudinal record of how each firm actually performs — not just on launch day, but across weeks and months of live trading.
Two models, two very different risk profiles
Before a project can even think about which market maker to pick, it needs to understand the two engagement structures that define the industry.
Structure A
Loan + call option
The market maker receives a loan of tokens and holds a call option to purchase more at a fixed price. Requires no stablecoins — only tokens. Currently dominates the market, especially for cash-constrained teams.
Structure B
Retainer + working capital
The project provides both tokens and stablecoins as working capital. More flexible for the market maker, but significantly more expensive. Requires the project to be stablecoin-rich — increasingly rare today.
Forgd strongly advises against mixing the two structures within a single project’s market-making setup. The incentive misalignments between the two models can compound in ways that make performance evaluation nearly impossible. Most projects today work with two to three market makers, all under the same structure, across three to four centralized exchanges.
Why launches fail — and why they rarely come back
The structural shift that made 2025 so brutal for altcoins wasn’t just macro. It was a fundamental change in what retail and speculative investors expect from a token at the moment of launch. In 2021, it was acceptable to launch pre-revenue, pre-product, even pre-business model. That window has closed.
Today’s market demands proof. Users want to see a product that’s live, tested, and attracting organic activity before the token enters secondary trading. If demand isn’t there to absorb the natural sell pressure that comes from airdrop recipients, exchange counterparties, and early investors, price decays — and it rarely recovers.
Why recovery is so difficult
- Negative price momentum drives away retail buyers, the only natural source of organic demand for most early-stage tokens
- Switching costs are low in crypto — users move to the next launch immediately
- Market makers that profited at launch have no incentive to rebuild depth at lower prices
- Airdrop recipients who didn’t sell at launch often become permanent overhang as conditions deteriorate
- Exchanges begin deprioritizing low-volume tokens for prominent placement
What Forgd’s data makes visible
The new analytics capabilities center on four standardized indicators that Forgd tracks across every market maker engagement on the platform.
Bid–ask spread measures the gap between the best available buy and sell prices in real time. Tight spreads signal a healthy, active market maker; wide spreads indicate withdrawal or failure to perform. Order-book depth captures how many resting orders exist at various price thresholds — the difference between a market where liquidity evaporates on moderate volume and one that absorbs selling pressure without significant price impact.
Uptime tracks whether the market maker is actually active and quoting during live trading hours, including through volatile periods. And fill order volume breaks down trading activity into maker and taker categories — revealing whether a market maker is building depth or extracting value from it.
“Transparent performance data raises the standard for everyone. It creates a market where firms are evaluated on outcomes, not marketing.”
Source quoted in Forgd announcement, GSRWhat makes this dataset genuinely new is its longitudinal dimension. Forgd isn’t showing a snapshot of how a firm performed for one project on one launch day. It’s showing patterns across hundreds of engagements, across different tokens, market environments, and venue types. That allows founders to ask not just “did this firm perform well?” but “does this firm consistently perform well when conditions get difficult?”
The RFQ process — and where founders get trapped
The standard way a project hires a market maker is through an RFQ process: the project circulates a brief to multiple firms, receives proposals with spread targets, depth commitments, and fee structures, and picks a winner. The problem is structural. RFQ dynamics create strong incentives for firms to submit aggressive bids — tight spreads, large depth commitments — that they may not be able to sustain once trading goes live.
When founders compare proposals through Forgd, they can now see each firm’s historical adherence to the kinds of KPIs they’re proposing. A firm bidding 0.15% spreads that has historically delivered 0.4% spreads on comparable tokens is a very different proposition from one with a consistent track record of hitting or beating its targets. That context changes the negotiation entirely.
Signals every founder should track before signing
Pre-launch market maker evaluation checklist
- Has the firm delivered comparable spreads and depth on tokens of similar size and liquidity profile?
- How did the firm’s depth and spread behavior hold up during volatile periods in prior engagements?
- Did performance hold beyond the first week post-launch, or did it decay as initial attention faded?
- Is the firm proposing the engagement structure that genuinely fits the project’s treasury position?
- What is the firm’s venue coverage — does it include both centralized and decentralized exchanges relevant to the project’s target user base?
East and west, CEX and DEX
One structural reality Forgd emphasizes with every founder: geography matters, and so does venue type. The speculative trading ecosystem in Asia — particularly Southeast Asia — generates significantly more volume and activity for newly listed altcoins than Western markets. A project with a North American founder base that skips Asian exchange listings is leaving a major portion of its potential buyer base untouched.
Similarly, Forgd advises against launching exclusively on centralized or decentralized venues. CEXs bring scale and distribution — large quantities of users who provide volume. DEXs attract a smaller but more sophisticated, protocol-native audience — the early adopters most likely to become power users of the underlying product. Market makers must cover both, and they approach the two venue types differently: centralized exchanges use continuous limit order books, while most DEX liquidity operates through concentrated liquidity AMM pools that require active management as price moves.
Interview bonus
Google invests up to $40B in Anthropic
According to April 24, 2026 reports, Alphabet, which is the parent company of Google, plans to invest up to $40 billion in Anthropic. This investment will consist of $10 billion upfront and $30 billion in additional funding if certain performance goals are achieved.
Anthropic is part of the multi-cloud ecosystem, including its deal with CoreWeave that spans several years. This is as well as its plan to harness close to 1 gigawatt of accelerators powered by Amazon’s GPUs.
Google invests up to $40B in Anthropic
Alphabet had previously made an investment of around 14% before this new funding round. Google will also be backing Anthropic in scaling up its computing capacity using TPU chips. The revenue run rate of Anthropic has now exceeded $30 billion, as compared to $9 billion at the end of 2025.
In response to the announcement, Alphabet stock was slightly buoyant. At midday, GOOGL stock rose about 1.2%, while the market was up about 3.77% overall, on an otherwise positive tech stock market.
Anthropic is not a publicly listed firm; hence, it lacks tradable shares. The stock price before the Forge Global announcement was strong, with the Forge Price as of April 23, 2026, at $259.14 per share.
However, recent transactions on the secondary market have raised the implied value substantially to levels between $800 billion and $1 trillion, or more, prior to the acquisition (such as a listing by one shareholder valued at $1.15 trillion). The commitment made by Google at around $350–$380 billion had no impact on the secondary pricing.
Anthropic continues to roll out major Claude updates
On April 16, 2026, Anthropic released Claude Opus 4.7. The model enhances software engineering, handles long-running code-generation projects effectively, and performs high-definition vision tasks.Â
There is no change in pricing, as Opus 4.7 costs $5 per million input tokens and $25 per million output tokens.
One day later, Anthropic introduced Claude Design. This experimental product helps create prototypes, slide presentations, one-page designs, and more with Claude. This is specifically designed for founders and product managers who don’t have a design background. The product works on Opus 4.7 for Claude Pro, Max, Team, and Enterprise.
In the early part of 2026, Anthropic released multiple updates. They include Claude Opus 4.6 and Sonnet 4.6 in February, with context window sizes of 1 million tokens; Claude Cowork for working collaboratively; and Claude Code for agentic coding.Â
The company has also released the Claude Mythos Preview at the beginning of April for complex cybersecurity activities; however, it is only available to select partners via Project Glasswing.
Global AI investment spike in April 2026
2026 is one of the most successful years for artificial intelligence funding. The first quarter alone saw $297 billion to $314 billion invested globally in venture capital, of which AI accounted for 80% to 81%. This includes the four largest venture deals of all time: $122 billion from OpenAI, $30 billion from Anthropic, $20 billion from xAI, and $16 billion from Waymo.
Capital inflow didn’t stop in April either. On April 3, Microsoft pledged $10 billion in investments in Japanese artificial intelligence and cybersecurity. Another commitment from Microsoft was to pledge $5.5 billion to the development of artificial intelligence in Singapore.Â
Then came Amazon’s announcement that it would invest $5 billion in Anthropic, with the potential for more in the near future. Finally, Cerebras Systems filed to go public at a $35 billion valuation on April 17. Overall capital expenditure intentions of Big Techs remain huge, at $635- $665 billion by year-end.
Still letting the bank keep the best part? Watch our free video on being your own bank.
DeepSeek is seeking funding at a valuation above $20 billion
DeepSeek is now chasing a valuation above $20 billion as Tencent Holdings and Alibaba Group discuss possible investments in the Chinese AI startup.
The Information reported that on Wednesday, citing four people who knew about the talks. DeepSeek, which is owned by hedge fund High-Flyer Capital Management, had only just started talking to outside investors for the first time.
By Friday, the reported target was at least $300 million at a valuation of at least $10 billion. Now the asking price has climbed fast as interest builds around DeepSeek.
The talks are still going on, so the final number could still change. The amount DeepSeek wants to raise could also change. Some U.S. venture capital companies may be cautious because DeepSeek is a Chinese startup.
Earlier this year, Cryptopolitan reported that DeepSeek did not show U.S. chipmakers its flagship model for performance tuning. We also reported that one of DeepSeek’s newer models was trained on Nvidia’s most advanced banned chip.
Back in January 2025, the first big DeepSeek release helped trigger a global tech selloff and pushed Chinese rivals to upgrade their own models.
Tencent and Alibaba push DeepSeek into a bigger money race
Meanwhile, on the Dwarkesh Podcast on Wednesday, Nvidia chief executive Jensen Huang said it would be “a horrible outcome” for the United States if DeepSeek optimized its new AI models to run on Huawei chips instead of American hardware.
Jensen said, “If future AI models are optimised in a very different way than the American tech stack,” and as “AI diffuses out into the rest of the world” with Chinese standards and technology, China “will become superior to” the United States.
On chip performance alone, Huawei still trails Nvidia. The Ascend 910C, which came before the 950PR, delivers about 60% of the inference performance of Nvidia’s H100. That H100 is already two generations behind Nvidia’s current best chip.
American chips are about five times more powerful than Chinese rivals today, and that gap is expected to widen to 17 times by 2027. Huawei is targeting 750,000 AI chip shipments in 2026, but its total production amounts to only about 3% to 5% of Nvidia’s combined computing power.
“A lot of work has to go into it to change. But go to the global south, go to the Middle East. Coming out of the box, if all of the AI models run best on somebody else’s tech stack, you’ve got to be arguing some ridiculous claim right now that that’s a good thing for the United States,” said Jensen.
AI funding surges as DeepSeek and Vast Data chase bigger checks
Jensen said his real concern is not just the gap in chip strength. He said China could still catch up in AI because it has “abundant energy” and a “large pool of AI researchers.”
If DeepSeek V4 runs well on Ascend chips, that would give China another route in AI development that does not depend on Nvidia across the supply chain.
That same funding rush showed up elsewhere on Wednesday. Vast Data announced a $1 billion funding round at a $30 billion valuation, and Nvidia was one of the backers.
The company says it supports projects that power millions of GPUs. Its customers include CoreWeave, Mistral, the U.S. Air Force, and Cursor. The new round more than tripled Vast’s $9.1 billion valuation from 2023. Drive Capital and Access Industries led the Series F. Fidelity Management and Research Co., NEA, and Nvidia also joined.
The financing included both primary and secondary capital. Dealroom said AI companies globally have already raised $280.5 billion this year, with more than $170 billion going to OpenAI, Anthropic, and xAI.
Chris Olsen of Drive Capital said, “The scale and speed of AI adoption are creating a new class of infrastructure company.” Chris added, “VAST is emerging as the clear leader in this category, with the architecture and momentum to support the world’s most demanding AI environments.”
The smartest crypto minds already read our newsletter. Want in? Join them.
Strategy Raises $1.76B War Chest As Saylor Signals Bigger Bitcoin Buy
Michael Saylor’s company has already lined up the money. Now the question is how much Bitcoin it plans to buy with it.
Saylor’s Signal Fires Up The Market
Strategy’s executive chairman posted his well-known “Orange Dots” chart on X over the weekend, adding just three words: “Think even Bigger.”
The chart maps every Bitcoin purchase the company has ever made. In crypto circles, its appearance has become a reliable preview of an imminent acquisition announcement — and Monday is the day Strategy most commonly makes those announcements public.
The post landed after a string of major purchases. On April 13, Strategy spent $1 billion on Bitcoin. The week before that, it dropped $330 million.
Both buying rounds were preceded by the same chart. This time, Saylor’s caption suggests the next move could top them both.
A War Chest Already Sitting Ready
The fuel for that purchase appears to already be in place. Strategy’s STRC instrument has raised enough capital to fund up to $1.76 billion in Bitcoin acquisitions, based on reports tracking the company’s fundraising activity.
The company routinely uses proceeds from STRC to bankroll its Bitcoin buying program, so the timing of that capital raise lines up with the weekend post.
At the time of writing, Strategy holds 780,897 Bitcoin across its corporate treasury. The company’s average purchase price sits at $75,577 per coin.
At current market prices, the entire stash is valued at roughly $58 billion — a figure that would shift significantly with any large new purchase.
Bitcoin Price Holds Flat Despite The News
The market has not moved much on Saylor’s hint. Bitcoin was trading around $75,500, down less than 1% in the 24 hours following the post.
Geopolitical pressure has been a drag on price action, with US President Donald Trump accusing Iran of violating ceasefire terms — a development that has kept risk appetite subdued across financial markets.
One signal watched closely by analysts did break out over the weekend, though. Bitcoin Dominance — the share of total crypto market value held by Bitcoin — pushed above a key resistance level on the three-day chart, clearing a descending trendline it had been stuck under for some time.
Reports from crypto analysts indicate that if the breakout holds, more capital could rotate into Bitcoin at the expense of smaller coins.
For Strategy’s playbook, that kind of market shift would not be unwelcome.
Featured image from MetaAI, chart from TradingView