OpenAI just closed the largest private funding round in history at one hundred and ten billion dollars, backed by Amazon, Nvidia, and SoftBank. The OCC dropped its proposed rulebook for the GENIUS Act, meaning federally regulated stablecoins are now a concrete policy discussion, not a hypothetical. Bitcoin spot ETFs snapped a five-week outflow streak with over a billion dollars flowing in across three days. Metaplanet reported a staggering 95 billion yen loss on its Bitcoin treasury position, but somehow managed to post a seventeen-fold increase in operating profit. And Jack Dorsey announced Block is cutting half its workforce to reorganize around what he calls an AI-era operating model. Plenty to dig into. Let's go.
So one hundred and ten billion dollars. Let that number sit for a second. OpenAI has now raised more in a single round than most countries spend on defense in a year. Amazon put in fifty billion, Nvidia and SoftBank each contributed thirty billion. The pre-money valuation sits at seven hundred and thirty billion dollars. This isn't a funding round in any traditional sense. This is industrial policy conducted through private capital.
What's fascinating is the structure behind the money. OpenAI is getting two gigawatts of Amazon's Trainium compute capacity and three gigawatts of Nvidia's dedicated inference infrastructure. ChatGPT now has nine hundred million weekly active users and over fifty million paying consumer subscribers. The projected operational costs through 2030 are six hundred and sixty-five billion dollars. That's not a typo. Running frontier AI models at this scale is absurdly expensive, and OpenAI is essentially pre-funding half a decade of operations.
Meanwhile, Microsoft rushed to put out a joint statement making clear that nothing has changed in their partnership. Azure remains the exclusive cloud provider for OpenAI's stateless APIs. The IP access stays exclusive. The revenue share is untouched. You can read that statement two ways: either as reassurance or as anxiety. When you feel the need to publicly declare that everything is fine, it usually means someone was worried it wasn't.
On the other side of the field, Anthropic is carving out a genuinely distinct position. Valued at three hundred and eighty billion, which is remarkable for a company that barely existed five years ago, Anthropic just launched customizable plugins for Claude that work directly inside enterprise tools like Excel and Gmail. But the more interesting story is Anthropic's standoff with the Pentagon. The Defense Department wanted unrestricted access to Claude for quote all lawful use unquote, and Anthropic said no, citing concerns about autonomous weapons and domestic surveillance. That takes real conviction when you're competing against companies that are cozying up to every government contract they can find.
Meta is pushing its own AI agents into Ads Manager, automating campaign research and analysis. OpenAI is forming alliances with consulting firms to accelerate enterprise adoption. The pattern is clear: every major AI player is racing not just to build better models, but to embed those models so deeply into enterprise workflows that switching costs become prohibitive. The moats aren't being built in the model weights. They're being built in the integration layer.
Bitcoin is stuck in an uncomfortable range, trading around sixty-seven thousand after briefly touching the low sixties earlier this month. The spot ETFs saw over a billion dollars in inflows across three days, snapping that brutal five-week outflow streak. BlackRock's IBIT led the charge. The Coinbase Premium index flipped positive, which signals renewed U.S. demand. So there are real signs of dip-buying.
But the macro headwinds are real. January PPI came in hot, which sent gold to a one-month high while punishing risk assets. Credit stress is mounting. U.S.-Iran tensions aren't helping. And there's a deeper structural concern: Bitcoin's rally has been riding record margin debt of one point two seven nine trillion dollars in U.S. markets. If that leverage starts unwinding, and there are signs it might be, the drawdown could accelerate.
Glassnode identified a key level that will determine whether Bitcoin's bounce holds or fades. Analysts are split. Some see the selling pressure as nearly exhausted, pointing to on-chain data showing that the bearish selloff is running out of steam. Willy Woo thinks Bitcoin will trade sideways for weeks but won't fully recover until the fourth quarter. Others are more bearish, with forecasts of a cycle low between thirty thousand and forty-five thousand backed by rising exchange reserves.
One genuinely bullish signal: the number of wallets holding at least one hundred Bitcoin is approaching twenty thousand, which some analysts view as a sign of improving distribution. Whales are accumulating, even as the price action looks ugly. ETF holders and treasury companies are simultaneously buying downside protection through options on Deribit against a crash below sixty thousand. That's the posture of long-term conviction hedged against short-term pain.
Mining profitability has cratered. CNBC reported that Bitcoin mining is no longer profitable at current prices for many operators. MARA posted a one point seven billion dollar quarterly loss on Bitcoin markdowns, though their pivot to AI data centers, including a new deal with Starwood, sent the stock up seventeen percent. TeraWulf missed Q4 estimates as mining revenue fell, but has twelve point eight billion dollars in AI and high-performance compute contracts lined up for 2026. The mining industry is transforming in real time. If you can't mine Bitcoin profitably, you mine AI compute instead. On the sustainability front, Cambridge University reported that over fifty-two percent of Bitcoin mining energy in 2024 came from sustainable sources, primarily hydro, wind, and nuclear. That narrative continues to strengthen.
Metaplanet is the most interesting and most divisive Bitcoin treasury story right now. The Japanese company holds thirty-five thousand one hundred and two Bitcoin, making it the fourth-largest corporate holder globally. Their average acquisition cost is around one hundred and seven thousand dollars per coin. With Bitcoin trading near sixty-eight thousand, that's an unrealized loss of roughly one point two billion dollars, which translated into a net loss of ninety-five billion yen for fiscal year 2025.
But here's where it gets interesting. Operationally, Metaplanet is thriving. Operating profit surged seventeen-fold to six point two nine billion yen, driven almost entirely by Bitcoin option premiums. Ninety-five percent of their revenue comes from selling options and derivatives against their Bitcoin holdings. Total revenue from these operations jumped seven hundred and thirty-eight percent. The company is essentially self-financing its Bitcoin accumulation through options income, which means they're not constantly diluting shareholders to buy more coins.
Management's target is breathtaking: they want to grow from thirty-five thousand to two hundred and ten thousand Bitcoin by 2027. That requires acquiring roughly one hundred and seventy-five thousand more coins. At current prices, that's roughly twelve billion dollars in additional purchases. The company has a ninety point seven percent equity ratio, which provides some buffer, but governance questions are mounting. Critics have pointed to undisclosed Bitcoin borrowing against existing holdings, and the CEO has been defending the company against accusations of dishonest disclosure.
The broader digital asset treasury space is showing the same tension. Strategy, formerly MicroStrategy, essentially created this playbook, and now a growing cohort of companies is running the same trade. When Bitcoin is going up, these companies look like geniuses. When Bitcoin drops thirty-seven percent from your average cost basis, the unrealized losses generate alarming headlines. But Metaplanet's operational model, generating real cash flow from derivatives while accumulating the underlying asset, is actually more sophisticated than most Bitcoin treasury strategies. The question is whether that sophistication scales and whether governance can keep up with ambition.
Circle's stock surged nearly fifty percent this week, but analysts noted it was driven by a short squeeze rather than strong fundamentals. Overcrowded bearish positioning got wiped out. Meanwhile, Citi and Morgan Stanley are expanding their Bitcoin and crypto custody, trading, and tokenization efforts, bringing institutional infrastructure further into the mainstream.
The OCC just dropped its proposed rules for implementing the GENIUS Act, and this is a genuinely significant moment for stablecoins. The GENIUS Act, signed into law in July 2025, established the first federal regulatory framework for payment stablecoins. Now we're seeing what the actual implementation looks like.
The key provisions: all payment stablecoins must be backed at least one-to-one with highly liquid reserve assets. Issuers must redeem stablecoins at par value within two business days. The OCC will have supervisory authority over subsidiaries of national banks, federal savings associations, and foreign stablecoin issuers seeking U.S. market access. There's a capital and operational backstop requirement, plus amended capital adequacy rules.
This matters for a few reasons. First, it creates regulatory clarity that major players have been waiting for. Barclays is actively looking for a technology partner to support blockchain-based payments and deposits, directly citing accelerating stablecoin adoption. PayPal, MoonPay, and M0 just launched PYUSDx, which lets developers issue app-specific stablecoins backed by PayPal USD. Japan is pushing yen-denominated stablecoins for DeFi. Germany's AllUnity launched a regulated stablecoin tied to the Swiss franc. The global stablecoin infrastructure is being built right now, in parallel across multiple jurisdictions.
But there's a potential dark cloud. The OCC's proposed rules include apparent limits on rewards, which could directly impact companies like Coinbase that currently share stablecoin yield with their users. If you can't offer competitive returns on stablecoin holdings, the value proposition changes significantly for retail users.
The deeper shift is what CoinDesk flagged in their analysis of Block's layoffs. While Dorsey cited AI productivity gains, the real pressure on fintech payment companies is stablecoin settlement threatening to compress the fee stack that acquirers have relied on for years. When you can move a dollar on a blockchain for fractions of a cent and settle in seconds, the entire payments margin structure comes under pressure.
World Liberty Financial's USD1 stablecoin supply has topped four point seven billion dollars, and they're now tying voting power to staking, which redirects arbitrage profits from institutional market makers to large token holders. The stablecoin landscape is evolving rapidly, and the GENIUS Act rules will shape which models survive and which get regulated out of existence.
Here's the thing to sit with today. OpenAI just raised more money than most of us can conceptualize. The OCC is building real regulatory infrastructure for stablecoins. Bitcoin miners are pivoting to AI compute to survive. Metaplanet is generating real operating profit while sitting on a billion-dollar paper loss. All of these stories share a common thread: the institutions and companies that will matter five years from now are the ones making irreversible bets right now, not waiting for clarity but building through the uncertainty. The question for everyone watching is simple. Are you positioned for where things are going, or where they've been?