Blue Origin's New Glenn rocket exploded in a catastrophic failure at Launch Complex 36 in Cape Canaveral, Florida, during a routine static fire test ahead of the NG4 mission. The explosion was so massive that footage circulating online drew comparisons to a nuclear detonation — one observer said it looked like a scene straight out of Oppenheimer. No one was injured, which given the scale of the destruction is nothing short of remarkable. Jeff Bezos confirmed the setback on social media, writing: "Very rough day. We'll rebuild whatever needs rebuilding and get back to flying. It's worth it."
What Happened to Blue Origin's New Glenn Rocket?
The explosion occurred during a static fire test — a ground-level engine ignition meant to validate the rocket's systems before an actual launch. These tests are standard procedure, but they carry real risk, as this incident made brutally clear. The blast caused extreme structural damage to Blue Origin's only functional launchpad at LC-36, Cape Canaveral Space Force Station in Florida.
Daylight imagery of the aftermath revealed significant destruction concentrated around the pad itself, though surrounding infrastructure appeared largely intact. That's actually a somewhat encouraging sign for the team — while every bolt and beam will need thorough reinspection for heat damage and corrosion, a full rebuild may not be necessary. Still, the operational delay will be substantial.
New Glenn had completed just one successful flight out of three total attempts prior to this incident. The third test had seen the rocket reach orbit but deploy its payload into the wrong orbit — a partial success at best. Now, a fourth attempt has been set back by an unknown number of months.
Elon Musk offered his support on social media, writing: "Most unfortunate. Rockets are hard." That's an understatement, but also a genuine nod of respect. It's worth remembering that SpaceX itself had three consecutive Falcon 1 failures between 2006 and 2008 before its first successful launch. Failure is not the exception in rocket development — it's practically a prerequisite.
How Does Blue Origin Stack Up Against SpaceX?
The timing of this failure is especially painful for Blue Origin. The company actually predates SpaceX in its founding, yet it has consistently lagged behind in operational capability and launch cadence. While SpaceX is preparing for a massive IPO reportedly valued at tens of billions of dollars — buoyed by the triple engine of its launch business, Starlink, and its growing AI infrastructure play — Blue Origin is now dealing with a destroyed launchpad and a grounded rocket.
Investor commentary on CNBC framed the gap clearly: SpaceX's launch business is exceptional, but Starlink is what makes it truly formidable, and the AI layer on top of that is a compounding advantage. Blue Origin, competing primarily in the heavy-lift launch market with New Glenn, simply doesn't have that diversified business stack to absorb setbacks the same way.
That said, having a second credible American provider of heavy-lift orbital capability is genuinely good for the country and the industry. When New Glenn successfully reached orbit on its first flight, it marked only the second time that class of reusable rocket technology had been demonstrated — and both times it was an American company doing it. Even America's second-best rocket, it turns out, is better than most of the world's best.
How Much Are Companies Really Spending on AI Tokens?
While the rocket story dominated headlines, the enterprise AI spending story may have longer-lasting economic implications. Anthropic recently surpassed $4.7 billion in annualized recurring revenue and closed a massive Series H round at a reported $65 billion post-money valuation. Claude Code went viral among developers in late 2024, and by mid-2025, Fortune 500 rollouts had gone wide.
The numbers are staggering. Amazon Web Services was reportedly spending approximately half a billion dollars in a single month on AI compute, according to reporting from Axios. Uber blew through its entire annual AI token budget before the year was out. Meta has been grappling with token maxing dashboards that may have generated ROI-negative AI activity. Microsoft, Salesforce, and DoorDash have all publicly acknowledged new efforts to rationalize their AI spend.
The Wall Street Journal summarized the dynamic well: corporate AI use is exploding, but the soaring costs have executives scrambling to understand what they're actually getting for the money.
What Is Token Maxing and Why Is It a Problem?
Token maxing is exactly what it sounds like: employees or systems consuming AI tokens at the maximum possible rate, often without proportional productivity gains to show for it. It's emerged as a real phenomenon inside large organizations, sometimes encouraged — accidentally — by the very dashboards meant to track AI adoption.
Here's how it happens: a company rolls out an AI usage leaderboard to celebrate early adopters and encourage experimentation. Employees, motivated by visibility on the dashboard, start running agents overnight on low-priority tasks, churning through backlogs that were probably deprioritized for good reason. The token meter spins. The ROI does not follow.
There's also a budget utilization dynamic at play. If a manager gives a team a $100,000 token budget and that team comes back having spent only $10,000, it can look like underperformance — even if they accomplished everything they needed to. So teams spend to the budget, not to the outcome.
Two economic principles help explain what's happening at a macro level:
- Jevons Paradox: When something becomes cheaper or more efficient to use, total consumption often goes up, not down. As AI gets better and cheaper, usage explodes faster than costs fall.
- Goodhart's Law: When a measure becomes a target, it ceases to be a good measure. Token leaderboards are a textbook example — optimize for token spend and you get token spend, not necessarily business value.
Is Enterprise AI Spending Actually ROI Positive?
The honest answer right now is: it depends, and most companies don't fully know yet. Spencer Rascoff from Match Group shared that his team's AI spend is in the single-digit millions — relatively modest — but he acknowledged that understanding the ROI on that investment is a key priority for the coming year. That's a reasonable, measured take.
The bull case is straightforward: AI costs are naturally deflationary. The same task that costs X today may cost a tenth of X next year, not because of any specific optimization, but because hardware depreciates, more power comes online, and open-source models close the gap. Companies that build the muscle memory now — the workflows, the integrations, the culture of using AI tools — will be positioned to capture enormous leverage when costs normalize.
The bear case is equally real: if AI is being deployed primarily against low-priority backlog items and make-work tasks, there's no amount of cost deflation that makes that ROI-positive. The cutting room floor exists for a reason. Not every deprioritized ticket deserved to be shipped.
The smartest operators are treating AI like a scalpel, not a hammer — using $200-a-month subscriptions to build genuinely useful products by knowing exactly what they're building, rather than running unlimited agentic loops hoping something valuable falls out.
What Is Anthropic's Valuation in 2025?
Anthropic's Series H round placed the company's post-money valuation at approximately $65 billion, with the company reporting annualized recurring revenue of $4.7 billion. Claude Code has become a particularly important product, driving widespread adoption among developers and enterprise teams alike. The company is approaching a valuation that would make it one of the most valuable private technology companies in history.
Why Is a T-Rex Fossil Selling for $30 Million?
In a story the Financial Times apparently deemed cover-worthy, Sotheby's is set to auction a 67-million-year-old Tyrannosaurus Rex skeleton — nicknamed Gus, after the rancher on whose South Dakota land it was discovered — with a pre-sale estimate of $20 to $30 million. This follows hedge fund titan Ken Griffin's 2023 purchase of a Stegosaurus fossil named Apex for $44.6 million, nearly ten times its pre-auction estimate of $4 to $6 million. Apex is now on display at the American Museum of Natural History. The ultra-high-net-worth fossil market, it turns out, is quietly becoming a serious asset class.
