Tuesday, March 31st, 2026. Here's what matters right now. AI infrastructure spending is going completely vertical, with billions flowing into new data center projects on 3 continents. Bitcoin is grinding near $67,000 as it faces a potential 6th straight monthly loss, geopolitical tension from the Iran conflict, and a fresh $2.2 billion FTX distribution hitting the market today. Google's quantum research team just dropped a paper arguing that breaking Bitcoin's cryptography may be far easier than anyone assumed. And in the startup world, a former a16z investor just raised $10 million to build a stablecoin clearinghouse. Let's get into it.
The scale of AI infrastructure buildout right now is honestly hard to wrap your head around. Let's start in Texas. Google is backing a $5 billion data center campus developed by Nexus Data Centers that will be leased to Anthropic. The site covers 2,800 acres, aims for 500 megawatts by late this year, and has a theoretical expansion path to 7.7 gigawatts. To put that in perspective, 7.7 gigawatts is roughly the output of 7 nuclear power plants. And because Texas grid capacity is already strained, they're going with on-site natural gas generation to bypass the grid entirely. That tells you something about where we are. The demand for AI compute is so intense that companies are literally building their own power plants.
Meanwhile in Arizona, Nvidia's fingerprints are all over a $3.8 billion junk bond deal funding a 30,000-acre data center project backed by a relatively unknown firm called Tract Capital. Despite Tract's limited track record, investor demand was overwhelming, with $14 billion in orders for a $3.8 billion offering. That's nearly 4 times oversubscribed. The Nvidia brand is essentially functioning as a credit guarantee at this point.
Across the Atlantic, France's Mistral AI secured $830 million in debt financing from a consortium of 7 banks to build a data center outside Paris housing 13,800 Nvidia GPUs with 44 megawatts of capacity. Mistral is positioning itself as Europe's sovereign AI alternative, with plans for a second facility in Sweden and a goal of 200 megawatts across Europe by 2027. This is significant because it marks a shift from equity to debt financing in AI infrastructure. When companies start issuing debt for data centers, it means the revenue models are maturing enough that lenders see predictable cash flows.
And here's a Bitcoin angle to all this. Bitfarms, once a dedicated Bitcoin miner, announced it's targeting zero Bitcoin on its balance sheet as it pivots entirely to AI data centers. They're actively selling their Bitcoin and redeploying capital into AI infrastructure. It's a telling sign of where the economics point right now. Data center operators see more reliable margins in AI compute than in Bitcoin mining, at least at current prices and difficulty levels.
Bitcoin is having a rough stretch, and the pressures are stacking up from multiple directions at once. The price is sitting around $67,000 to $67,500, and if it closes March below $67,300, that would confirm 6 consecutive monthly losses. That would tie the longest losing streak in Bitcoin's history.
Let's unpack what's weighing on it. First, the macro picture. Oil prices hit a 3-year high above $105 a barrel, with WTI crude closing above $100 for the first time since 2022. The Strait of Hormuz remains shut amid the Iran conflict, and while Trump signaled he may seek to end hostilities, the damage to energy markets is already done. High oil means higher inflation expectations, which means the Fed stays restrictive longer. Real yields on 10-year TIPS are surging, and that's a direct headwind for zero-yielding assets like Bitcoin.
Second, today specifically, FTX begins its 4th creditor distribution, releasing approximately $2.2 billion through BitGo, Kraken, and Payoneer. That money hits eligible customers within 1 to 3 business days. Now, not all of it will be sold into the market, but it's a meaningful liquidity event arriving at a fragile moment.
Third, David Bailey's Nakamoto, one of the newer Bitcoin treasury companies, sold roughly 284 BTC, about 5% of its holdings, at a loss. They'd bought Bitcoin at around $118,000, and they're selling near $67,000. The stock has collapsed. This is exactly the risk that skeptics of the treasury company model have been warning about. When you lever up on Bitcoin at cycle highs and the price drops 45% or more, the whole structure comes under pressure.
On the bullish side, there are some interesting signals. Whales holding between 10 and 10,000 BTC have quietly accumulated around 62,000 coins during this drawdown. MicroStrategy bought roughly 40,000 BTC during March alone. And spot ETF inflows have remained positive on a monthly basis. So institutional conviction hasn't broken. But on-chain data tells a more cautious story. Long-term holder profitability has dropped from 58% to just 3%, and short-term holder profits have cratered. The $59,000 level, defined by Bitcoin's 200-week moving average, is the line that separates a deep correction from a full bear market. That's the number everyone is watching.
Let's shift to what's being built. A few notable funding rounds dropped this week that tell you where smart money is flowing in the Bitcoin and stablecoin ecosystem.
First up, Sam Broner, a former a16z crypto investor, raised $10 million in seed funding to launch The Better Money Company. It's a stablecoin clearinghouse designed to facilitate low-cost exchanges between different dollar-backed tokens. Think of it as plumbing for the stablecoin economy. Right now, moving between USDC, USDT, and the growing number of newer stablecoins involves friction, slippage, and unnecessary cost. A dedicated clearinghouse that can handle those conversions efficiently is exactly the kind of infrastructure the market needs as stablecoin supply approaches $200 billion.
Then there's Valinor Digital, which raised $25 million in seed funding led by Castle Island Ventures. Valinor is building what they call an Open Credit platform, essentially bringing private credit markets on-chain. The founders come from Blackstone, and they're using smart contracts to automate things like revolving loans that currently run on spreadsheets and email chains. Private credit is a $2 trillion market heading toward $3.5 trillion, and bringing even a fraction of that on-chain could be transformative. The real-world asset tokenization market has already hit $24 billion, and private credit is emerging as one of its biggest categories.
On the cross-border payments front, OpenFX raised $94 million in a Series A to expand its stablecoin-powered foreign exchange network. They act as a bridge between traditional banking and digital assets, enabling faster and cheaper FX conversions for businesses moving large sums internationally. This is one of the clearest product-market fit stories in crypto right now. Stablecoins are genuinely better than the existing correspondent banking system for cross-border payments, and the companies building that bridge are attracting serious capital.
And in a regulatory tailwind that could matter enormously, the U.S. Labor Department proposed a rule change that would allow crypto investments in 401(k) retirement accounts. This follows an executive order from President Trump directing regulators to expand digital asset access in retirement portfolios. If this goes through, it could open trillions of dollars in retirement savings to Bitcoin and crypto exposure. That's a structural demand shift that would dwarf anything we've seen from spot ETFs.
Let's close with the AI geopolitics angle, because it's evolving fast. China's AI strategy is entering a new phase. The initial wave was about building frontier models to match GPT-4 and Claude. Now the focus is shifting to industry-specific AI applications and monetization.
Alibaba is integrating AI deeply into its sourcing platform with a product called Accio Work that automates customs calculations, profit projections, and supplier matching. Chinese companies are increasingly seeking partnerships with U.S. firms to build specialized AI for law, finance, and HR. It's a pragmatic pivot. Rather than trying to out-compete OpenAI on general reasoning, they're going after vertical use cases where AI can drive immediate revenue.
But the infrastructure challenges are real. DeepSeek, China's leading AI platform, suffered a major 7-hour outage over the weekend that disrupted developer workflows globally, including for Chinese government agencies. The cause hasn't been confirmed, but speculation points to backend migration work ahead of their V4 model launch. During the blackout, competitors like Alibaba's Qwen saw a spike in traffic. It's a reminder that even the most efficient AI systems are only as good as their infrastructure.
On the geopolitical front, China proposed establishing a global AI regulatory body at the upcoming APEC summit. This is a direct challenge to U.S. influence over AI governance. The proposal emphasizes international cooperation and responsible development, but the subtext is clear. Beijing wants a seat at the table, or better yet, to set the table, when it comes to how AI is governed globally.
Meanwhile, the data backs up China's growing influence. According to one analysis, nearly 80% of U.S. AI startups are utilizing Chinese open-source technology, primarily models from Alibaba's Qwen family. That dependency cuts both ways. It gives Chinese AI companies enormous distribution, but it also gives U.S. policymakers a reason to worry about supply chain vulnerabilities in AI, much like the concerns around Huawei in telecom years ago.
And speaking of supply chains, U.S. senators introduced the Mined in America Act, aimed at boosting domestic Bitcoin mining hardware manufacturing. Right now, the U.S. hosts 38% of Bitcoin's hashrate, but 97% of mining machines are made by 2 Chinese companies. It's the same pattern playing out across tech. American innovation running on Chinese hardware.
Here's the thought to sit with. The AI infrastructure buildout is now measured in gigawatts and tens of billions of dollars. Bitcoin is measured by whether it can hold $59,000. These two worlds are deeply connected. The same energy constraints, the same capital markets, the same geopolitical tensions shape both. The question isn't whether AI or Bitcoin wins. It's whether the physical infrastructure of the world can keep up with what we're trying to build on top of it. That's the bottleneck that matters. I'll talk to you tomorrow.