Saturday, March 28th. Bitcoin sliding toward the mid-60s as macro headwinds pile up. Traders now price in better than 50-50 odds of a Fed rate hike before year-end — not cuts, a hike. Strategy, formerly MicroStrategy, is essentially the last corporate buyer standing, responsible for 76% of all corporate Bitcoin holdings while every other treasury company has gone quiet. Meanwhile, JetBrains launches a full platform for orchestrating AI coding agents across teams, and regulators from Hong Kong to Europe are reshaping who gets to operate in crypto. Let's get into it.
Let's start with the elephant in the room. Bitcoin is trading near $66,000, down roughly 20% on the year, and about 47% off its all-time high near $127,000 set last year. Traders on options markets are pricing in a 53% chance we see sub-$66,000 Bitcoin by late April, and some technical analysts are warning of a possible weekend drop toward $61,000. The Fear and Greed Index is sitting at Extreme Fear. Not great.
But here's what's really striking. Strategy — the company formerly known as MicroStrategy — is basically the only major corporation still buying Bitcoin in size. In the past 30 days alone, they've added roughly 45,000 BTC to their stack. Their total holdings now sit at 762,099 Bitcoin, worth about $51.7 billion at current prices. That's 76% of all corporate Bitcoin reserves globally.
Meanwhile, every other company combined bought about 1,000 BTC over the same period. That's a 99% drop from corporate purchases a year ago. The concentration is extreme.
And Strategy isn't exactly thriving on paper either. The stock is down somewhere between 14% and 57% depending on which measurement window you use, and the company's market cap now trades at a discount to its net Bitcoin holdings — the mNAV has dipped below 1. That's a signal that investors are pricing in real risk around the leverage and the strategy itself.
To fund continued purchases, the company announced plans to raise $42 billion through additional stock offerings. One of Strategy's directors, Jarrod Patten, sold 700 shares worth about $96,000 this week — a small amount, but the optics aren't ideal when insider selling hits during a drawdown.
There's also a broader story here about Bitcoin miners. The average public miner spent about $80,000 to produce one Bitcoin last quarter. Bitcoin is trading at $70,000. That math simply doesn't work. So miners are pivoting hard to AI data center contracts — taking on $70 billion in new commitments — and liquidating their Bitcoin treasuries to finance the transition.
On the brighter side of institutional activity, Morgan Stanley filed for a spot Bitcoin ETF with a fee of just 14 basis points — that would make it the cheapest Bitcoin fund on the market if approved. Bloomberg's Eric Balchunas noted that Morgan Stanley's 16,000 financial advisors managing $6.2 trillion in client assets would have no problem recommending a product at that price point. But even here, timing matters. Spot Bitcoin ETFs just broke a 4-week inflow streak with $296 million in outflows last week as capital avoids directional risk.
Now let's zoom out to the macro picture, because the pressure on Bitcoin isn't coming from crypto-specific factors alone. The big shift this week is in interest rate expectations.
Traders are now pricing in over 50% odds that the Federal Reserve will raise rates at least once before the end of 2026. Let that sink in. A few months ago, the consensus was that we'd be getting cuts. Now the market is preparing for the opposite.
What's driving this? A combination of persistent inflation above the Fed's 2% target, a labor market that refuses to cool down, and rising input costs linked to new trade tariffs and spiking oil prices amid the conflict in Iran. The 10-year Treasury yield has climbed to 4.46%, its highest since last July. The Fed funds rate sits at about 3.75%, and Fed officials have been making increasingly hawkish noises.
President Trump, for his part, continues to publicly call for low interest rates and zero inflation — essentially pressuring the Fed to cut. But the gap between political rhetoric and monetary reality keeps widening. The Fed has made it clear it's waiting to see whether energy-driven price pressures are transitory or persistent before making any moves.
For Bitcoin, this is a tough setup. Higher rates mean tighter liquidity, stronger dollar, and reduced appetite for risk assets. U.S. stocks are already feeling it — the Nasdaq and Dow have entered correction territory. Economic growth forecasts are being downgraded. The risk of recession has increased, even as AI-related investment continues to prop up certain sectors.
Historically, Bitcoin has never finished a year positive after starting this badly. Whether 2026 breaks that pattern remains an open question, but the data shows that each new price low tends to add months to Bitcoin's recovery timeline. Some analysts are now suggesting $88,000 might be the ceiling for this year, while others warn that a dip below $60,000 could delay any recovery to 2027.
The one wildcard: if macro conditions deteriorate enough to force the Fed into emergency cuts, that could flip the script entirely. But right now, the base case is higher for longer.
Shifting to AI. Two big developments this week that capture the duality of where agentic AI stands right now — getting more powerful and more vulnerable at the same time.
First, the power side. JetBrains — the company behind IntelliJ and a bunch of other developer tools — launched JetBrains Central, a full orchestration platform for AI coding agents. This isn't just another copilot bolt-on. It's designed to let engineering teams distribute tasks across multiple AI agents, manage them centrally, and integrate them into existing workflows including CI/CD pipelines, version control, cloud infrastructure, and even Slack.
The platform includes a tool called Air for distributing work among agents, and Junie CLI, which supports various large language models. What's interesting is the management layer — policy enforcement, audit trails, cost tracking. JetBrains is clearly betting that the bottleneck for AI in software development isn't the models themselves anymore. It's governance, coordination, and control. They want to be the operating system for AI-powered development teams.
On a parallel track, OpenAI extended its Responses API to serve as a proper foundation for building autonomous agents. The new features include a shell tool that lets agents execute commands in Go, Java, NodeJS — not just Python. There's a containerized workspace with secure network policies, and a skill framework for reusable workflows. The core idea is an iterative execution loop: the model proposes an action, it gets executed in a sandboxed environment, results feed back in, and the loop continues until the task is done. This significantly lowers the barrier for developers who want to build agents that do real things in the real world.
But here's the vulnerability side. Security researchers disclosed significant flaws in LangChain and LangGraph — two of the most widely used frameworks for building AI agents. These vulnerabilities can expose sensitive files, secrets, and databases in applications built on these frameworks. If you're an organization running LangChain-based agents in production, this is a wake-up call. The attack surface of agentic AI is growing faster than the security tooling around it.
This tension — between rapid capability gains and lagging security — is probably the defining challenge for agentic AI in 2026. The tools are getting remarkably capable. But the more autonomy you give an agent, the more damage a vulnerability can do.
Finally, the regulatory landscape. Three regions, three different stories, all pointing in the same direction: the window for operating without proper licenses is closing fast.
Let's start with Hong Kong. The city is on the verge of issuing its very first stablecoin licenses. And the early licensees aren't scrappy crypto startups — we're talking HSBC, Standard Chartered, and a joint venture between Standard Chartered, Animoca Brands, and HKT. Hong Kong's regime isn't the first stablecoin licensing framework globally, but given the city's role in cross-border finance, it could be one of the most consequential. This is traditional finance embedding itself into stablecoin infrastructure from day one.
In Japan, the Financial Services Agency issued warnings to 4 unregistered crypto platforms including KuCoin for operating without a license and targeting Japanese residents. The FSA is going further than just warning letters — they've asked Apple and Google to remove KuCoin's app from their stores. Japan is also planning to shift its crypto regulatory framework from the Payment Services Act to the Financial Instruments and Exchange Act, which would bring significantly stronger enforcement powers and reporting requirements. With over 12 million crypto users in Japan, the stakes are high.
KuCoin in particular is having a terrible regulatory year. They've now been hit in Japan, Hong Kong, Austria, and Dubai — where VARA ordered them to cease all unlicensed activities. User funds and reserves have been declining significantly as trust erodes.
In Europe, MiCA continues to reshape the landscape. There are now 174 registered crypto asset service providers under MiCA. But here's the nuance — only 14 of those are actually licensed to operate a centralized exchange. Most MiCA licenses cover narrower services like custody or transfer. Companies like Coinbase, Kraken, and Revolut each use their licenses differently depending on the specific services they offer. It's a more fragmented picture than the headline number suggests.
And the European Central Bank is now investigating 4 DeFi protocols — Aave, MakerDAO, Ampleforth, and Uniswap — under MiCA. Their concern is centralization: the top 100 holders control over 80% of these protocols' token supply. That level of concentration could disqualify them from DeFi exemptions under MiCA, triggering stricter capital requirements and governance standards. The irony of DeFi protocols being too centralized for regulators is not lost on anyone.
Here's the thought to sit with this weekend. Strategy holding 76% of all corporate Bitcoin is not a sign of conviction spreading — it's a sign of conviction concentrating. When one entity becomes the market's backstop buyer, the resilience of that market depends entirely on that one entity's balance sheet and willingness to continue. That's not decentralization. That's a single point of failure. The question isn't whether Michael Saylor is right about Bitcoin long-term. It's what happens to the market if the one buyer who matters ever has to stop.