Category: English

Let that number sink in for a moment. Nine hundred billion dollars.

That is the staggering figure that Anthropic, the startup behind the Claude chatbot, has reportedly agreed to in its latest funding round. According to the Financial Times, the company has finalized terms for a massive $30 billion fundraising effort that would catapult its valuation to approximately $900 billion.

And here is the kicker. That valuation does not just make Anthropic the most valuable AI startup on the planet. It officially puts the company ahead of its arch-rival, OpenAI, whose valuation stands at roughly $852 billion after its record-breaking round earlier this year.

The AI king has a new face. And its name is Anthropic.

How Did This Happen So Fast?

Just three months ago, Anthropic completed a Series G funding round at a valuation of $380 billion. Three months. That is not growth. That is a rocketship strapped to a supernova. In a single quarter, the company has nearly tripled its valuation.

What makes this even more mind-boggling is how quickly the deal came together. Investors first approached the company last month. Anthropic’s Chief Financial Officer, Krishna Rao, began testing market demand. And within a few weeks, the terms were locked in. For those keeping score at home, that is about the time it takes most startups to schedule a board meeting.

The lead investors in this round read like a who-is-who of Silicon Valley royalty. Dragoneer Investment Group, Greenoaks Capital, Sequoia Capital, and Altimeter Capital have all agreed to co-lead the funding. Each of these heavy hitters is reportedly committing at least $20 billion to the deal. The final round is expected to close as early as this month.

There is something poetic about that lineup. Three of those investors are also major backers of OpenAI. Dragoneer, Sequoia, and Altimeter have all poured billions into Sam Altman’s company over the years. Now, they are writing even bigger checks to its biggest competitor. If that does not tell you everything you need to know about where the smart money is moving, nothing will.

The Numbers That Broke Silicon Valley

Valuations are one thing. But in the world of venture capital, nothing speaks louder than revenue. And on that front, Anthropic is delivering numbers that are almost impossible to comprehend.

The company’s annualized revenue is expected to cross $45 billion soon. To put that in perspective, at the end of last year, that figure was just $9 billion. In less than six months, Anthropic has more than quadrupled its annualized revenue.

Let me run those numbers again. Nine billion to forty-five billion. In half a year.

Even more stunning is the comparison with OpenAI. According to reports, OpenAI’s annualized revenue sits at about $24 billion. That means Anthropic is now generating nearly twice the revenue of the company that started the generative AI revolution.

If you zoom out even further, the growth trajectory becomes almost mythical. Just a few years ago, in 2023, Anthropic’s annualized revenue was around $100 million. From a hundred million to forty-five billion in three years. That is a four-hundred-and-fifty-fold increase.

Wall Street is not just noticing. Wall Street is panicking to get in.

The Secret Sauce Called Claude

So what is driving this insane growth? The answer is a product called Claude Code.

At a recent developer conference in San Francisco, Anthropic CEO Dario Amodei dropped a bombshell. In the first quarter of this year, the company’s annualized revenue and usage grew by eighty times compared to the same period last year. Eighty times. And here is the crazy part. Anthropic had only planned for ten times growth.

Amodei even joked on stage that the hyper-growth was too difficult to manage. He said he wished the growth would slow down to just ten times so he could catch his breath. That is the kind of humblebrag that only a CEO whose company is on fire gets to make.

The engine behind this eighty-fold explosion is Claude Code, an AI-powered programming assistant that launched in late 2025. The product spread through the developer community faster than anyone anticipated. Software engineers, known for being the fastest adopters of new tools, flocked to Claude Code in droves.

What they found was an AI that could handle coding, debugging, and long-context reasoning tasks with a level of sophistication that made other tools look like toys. And once developers started using it, they never looked back.

But Anthropic is not just winning in the coding space. The company’s Claude models are being deployed across enterprise software, research, legal analysis, and countless other industries. More than five hundred customers are now spending over a million dollars annually on Anthropic’s services.

The Infrastructure Arms Race

Of course, all this growth comes with a massive catch. Running cutting-edge AI models requires an almost obscene amount of computing power. And when your usage grows eighty times faster than planned, your infrastructure tends to explode.

Over the past month, Anthropic has been open about the pressure on its infrastructure. Users complained about rate limits and outages during peak hours. The company was scrambling to keep up with demand.

Then came the plot twist that nobody saw coming.

Elon Musk stepped in. Anthropic announced an exclusive partnership with SpaceX to tap into the full computing power of the Colossus 1 data center. That facility boasts over 300 megawatts of computing capacity and a staggering 220,000 of Nvidia’s top-tier GPUs.

Those chips were originally built for Musk’s own AI venture, xAI. Now, all 220,000 of them are being redirected to Anthropic. And they will be fully deployed within a month.

This is not just a partnership. This is a declaration of war. Musk, who was one of OpenAI’s co-founders before leaving the board, is now throwing his full weight behind its biggest rival.

Anthropic has also locked down major infrastructure deals with Amazon Web Services, Google Cloud, and Akamai. These partnerships are designed to secure the vast computing power needed to train and run the next generation of AI models. In the AI arms race, compute is ammunition. And Anthropic is loading up.

The Leadership Shuffle

While the numbers are impressive, the story of how we got here is just as compelling. Anthropic was founded in 2021 by a group of former OpenAI researchers led by Dario Amodei and his sister Daniela Amodei.

They left OpenAI because they believed the company was moving too fast and paying too little attention to safety. They wanted to build AI that was powerful but also aligned with human values. At the time, most of the industry dismissed them as idealistic defectors.

Fast forward five years, and that bet has paid off in ways nobody could have imagined. While OpenAI has been forced to redraw its product roadmap twice in the past six months in response to competitive threats first from Google and then from Anthropic, the so-called safety-first upstart has been quietly eating their lunch.

Some of OpenAI’s own investors are now questioning the company’s strategy. According to the Financial Times, OpenAI’s shift toward the enterprise market has left some backers wondering if the company is losing its focus. One early investor asked publicly why a company with a billion-user consumer business is pivoting to talk about enterprise and code.

That question is particularly pointed because enterprise and code are exactly the markets where Anthropic is crushing it.

The Secondary Market Frenzy

The excitement around Anthropic is not just coming from institutional investors. On secondary markets like Forge Global, where private company shares trade before an IPO, Anthropic’s valuation has already surpassed $1 trillion. That is nearly twenty percent higher than OpenAI’s secondary market valuation of about $880 billion.

Business Insider reports that buyers are scrambling to get their hands on the dwindling supply of Anthropic shares. Demand has become so intense that investors are willing to buy at almost any price. The CEO of Rainmaker Securities, an investment bank that specializes in private securities transactions, described it as an epic period for Anthropic.

But he also admitted that much of the demand is driven by FOMO rather than market fundamentals. Venture capital firms and family offices feel like they need to own Anthropic stock regardless of price. They are terrified of missing out on the next trillion-dollar company.

On the flip side, traders are noticing a drop in demand for OpenAI stock. It is now trading at a discount compared to Anthropic on secondary markets. Just a few months ago, that would have been unthinkable.

What This Means for the AI Industry

This is the part where I try to pull back and look at the bigger picture.

Anthropic’s rise signals something profound about where the AI industry is headed. For the past few years, the narrative has been that OpenAI is the undisputed leader. ChatGPT captured the world’s imagination. Sam Altman became the face of the AI revolution. Everyone else was just playing catch-up.

That narrative is dead.

What we are seeing now is the emergence of a genuine duopoly. Instead of winner-takes-all, the AI market is shaping up to be a two-horse race between OpenAI and Anthropic. And right now, the second horse is pulling ahead.

Investors are increasingly treating AI startups not as experimental bets but as foundational infrastructure companies. They are comparing them to the cloud computing giants of the previous decade. And they are willing to pay prices that would have been laughed out of the room just two years ago.

But there is also a cautionary note buried in all this euphoria. These valuations raise serious questions about sustainability. Training advanced AI models requires enormous computing power, expensive GPUs, and long-term infrastructure investments. Companies like Anthropic are growing quickly, but they are also spending billions just to stay in the race.

Can Anthropic maintain its momentum? Can it keep its safety-focused culture intact while scaling at warp speed? Can it fend off not just OpenAI but also a resurgent Google and a deeply determined Meta?

Those are the questions that will define the next chapter of this story.

The Road Ahead

If the current funding round closes at the $900 billion valuation, Anthropic will become one of the most valuable private companies in the world. And with an IPO reportedly on the horizon, perhaps as early as this year, the company could soon make its public market debut at a valuation that rivals the biggest names in tech.

To give you some context, a $900 billion valuation would make Anthropic larger than Meta, larger than Tesla, larger than every company in the world except a handful of giants like Apple, Microsoft, and Nvidia.

That is the scale we are talking about.

For Dario Amodei and his team, the journey from exiled OpenAI researchers to leaders of the AI industry has been nothing short of extraordinary. They bet that safety and capability could go hand in hand. They bet that enterprise customers would pay a premium for AI they could trust. They bet that the developer community would embrace a tool that made them more productive.

All of those bets have paid off in spectacular fashion.

As for OpenAI, the company is not going anywhere. It still has ChatGPT, which commands nine hundred million weekly active users and over fifty million paying subscribers. It still has the deepest bench of AI talent in the industry. And it still has a massive war chest after raising $122 billion at an $852 billion valuation just a few months ago.

But for the first time since ChatGPT launched, OpenAI is looking up at someone else.

The AI king has a new face. And the race for the trillion-dollar crown is just getting started.

When it comes to cancer treatment, few breakthroughs have generated as much excitement over the past decade as CAR-T therapy. It’s been a genuine game-changer for people with blood cancers like leukemia and lymphoma—patients who had run out of options suddenly found themselves in remission.

But there’s a catch.

The current model of CAR-T therapy is incredibly complicated, painfully slow, and eye-wateringly expensive. We’re talking about a process that takes two to three weeks, involves collecting your own immune cells, shipping them to a centralized lab, genetically engineering them, and then infusing them back into your body. And the price tag? Hundreds of thousands of dollars for a single treatment.

That’s not a typo.

Now, a radically different approach is emerging. It’s called in vivo CAR-T, and it promises to turn your own body into the factory. Instead of removing your T cells and modifying them outside your body, in vivo therapy delivers the genetic instructions directly into your bloodstream, where your T cells get reprogrammed right then and there.

Sounds like science fiction, right? Well, it’s already happening in clinical trials across the United States, Australia, and China. And the early results are turning heads.

Let me walk you through what’s happening, because this could fundamentally change how we think about cancer immunotherapy.

The Traditional CAR-T Problem No One Likes to Talk About

First, let’s give credit where it’s due. Traditional ex vivo CAR-T therapy has saved thousands of lives. Since the FDA approved the first CAR-T product back in 2017 for acute lymphoblastic leukemia, these therapies have become a standard option for certain blood cancers.

But the manufacturing process is a nightmare.

Here’s what a patient currently goes through. First, you sit for several hours while a machine pulls your blood and separates out your T cells. That’s called leukapheresis. Those cells get frozen and shipped to a manufacturing facility somewhere across the country. At the facility, technicians use a viral vector to insert the CAR gene into your T cells. Then those engineered cells get expanded into the hundreds of millions. Finally, they’re shipped back to your hospital and infused into your veins.

Oh, and before the infusion, you also need a round of lymphodepleting chemotherapy to wipe out your existing immune cells so the new CAR-T cells have room to expand. That chemotherapy leaves you vulnerable to infections.

The entire process takes two to three weeks. For patients with rapidly progressing cancer, that waiting period can be fatal.

As one oncologist put it during a recent medical conference, the current ex vivo model is effective but limited by complexity, cost, and logistics. The in vivo approach, by contrast, is designed to function more like gene therapy. You basically write a prescription, the patient gets an infusion, and their own lymphoid organs become the bioreactor that manufactures the CAR-T cells.

That’s the dream, anyway.

What Exactly Is In Vivo CAR-T?

The concept is deceptively simple. Instead of doing all the genetic engineering in a clean room facility, you deliver the CAR gene directly into a patient’s bloodstream using some kind of delivery vehicle. That vehicle could be a lentiviral vector, an adeno-associated virus, or a lipid nanoparticle similar to what was used in the mRNA COVID vaccines.

Once inside the body, these delivery vehicles find their way to T cells and insert the CAR gene. Those T cells then start producing the chimeric antigen receptor on their surface. And just like that, you have functional CAR-T cells roaming around, hunting down cancer.

No cell collection. No weeks-long manufacturing. No lymphodepleting chemotherapy. Just a single intravenous infusion, and the patient’s body does the rest.

Researchers are exploring multiple delivery platforms for this approach. Lentiviral vectors are the most mature option, given their established safety profile and ability to achieve stable gene integration. Non-viral systems like lipid nanoparticles are also gaining traction because they offer transient CAR expression, which could provide better safety control.

Stanford researchers recently showed that they could generate CAR-T cells inside mice using the same mRNA-based lipid nanoparticle technology used in COVID vaccines. The results were impressive. About seventy-five percent of mice with B cell lymphoma saw their tumors eradicated after several doses. And crucially, the approach didn’t require any pre-treatment to deplete existing immune cells. The researchers reported seeing no toxicity even with a fairly large number of injections.

The Clinical Data That’s Got Everyone Excited

The real proof, of course, comes from human trials. And in the past year, we’ve seen some genuinely encouraging data.

Take the ESO-T01 trial conducted in China. This Phase 1 study enrolled five heavily pre-treated patients with relapsed multiple myeloma. These were patients who had already failed a median of three prior lines of therapy. They received a single infusion of an immune-shielded lentiviral vector designed to generate anti-BCMA CAR-T cells directly inside their bodies.

No leukapheresis. No ex vivo manufacturing. No lymphodepleting chemotherapy.

The results were striking. Four out of five patients achieved objective responses. Three of those four reached stringent complete remission, meaning no detectable cancer. And all four evaluable responders achieved minimal residual disease negativity by day sixty.

CAR-positive T cells were detectable in the blood as early as day seven after infusion, with peak expansion around days ten to fourteen. The responses deepened over time, with patients moving from partial response to deeper remission categories. At a median follow-up of six months, all responses were still ongoing.

One patient died before the first efficacy assessment due to spinal cord compression from progressive disease, but investigators attributed that to the underlying myeloma rather than the treatment.

Now, I should mention that this trial was stopped early in 2025, and no further enrollment was performed. Still, the data demonstrated that in vivo CAR-T generation is feasible in humans.

Then there’s KLN-1010 from Kelonia Therapeutics, which made headlines when the first US dose was administered at Winship Cancer Institute of Emory University in May 2026. This Phase 1 trial for relapsed multiple myeloma has shown promising early results, with four patients achieving MRD-negative status. The longest response lasted five months at the time of reporting.

The patients didn’t need lymphodepletion. They didn’t need to wait weeks for manufacturing. And all four saw CAR-T cell expansion around day fifteen after infusion.

The momentum has been so strong that Eli Lilly agreed to acquire Kelonia Therapeutics for up to seven billion dollars. That’s billion with a B. The deal includes a three-point-two-five-billion-dollar upfront payment, with additional milestone payouts tied to clinical and regulatory success.

If that doesn’t signal industry confidence, I don’t know what does.

But Let’s Talk About the Elephant in the Room

For all the excitement, in vivo CAR-T therapy is not without risks. And the safety data so far has been mixed.

In the ESO-T01 trial, all five patients experienced grade three or higher adverse events. Cytokine release syndrome occurred in four patients. Three patients had grade two infections. One patient developed grade one neurotoxicity. Transient cytopenias and reversible liver enzyme elevations were common.

A separate presentation at a major cancer conference highlighted these safety concerns. One patient in an early trial died, though the cause wasn’t entirely clear. The attending physician suggested the death may have been inflammatory or cardiac-related.

The early toxicity is thought to be related to an innate immune response to the viral particles themselves. But alternative mechanisms, such as T cell activation driven by the delivery vector, may also play a role.

Another open question is how well these vectors can access T cells beyond the bloodstream. Most of the T cells in your body aren’t actually circulating in your blood. They’re sitting in your lymphoid organs like the spleen and lymph nodes. In vivo delivery systems may or may not reach those cells effectively. On the flip side, being able to access that larger pool of cells could actually be an advantage if the technology works properly.

So the jury is still out.

The Industrial Scale of This Shift

One thing is clear: Big Pharma is betting heavily on in vivo CAR-T.

Beyond the Lilly-Kelonia deal, we’re seeing major investments across the board. Umoja Biopharma’s UB-VV111 became the first in vivo CAR-T therapy to receive FDA clearance for study back in 2024. The FDA has since granted fast track designation to this therapy for relapsed large B-cell lymphoma and chronic lymphocytic leukemia. AbbVie holds an exclusive option to license Umoja’s CD19-targeted candidates.

AstraZeneca acquired EsoBiotec, the Belgian company behind the ESO-T01 vector, in a deal worth up to one billion dollars.

Sana Biotechnology continues to advance its in vivo CAR-T candidate for blood cancers and B cell-mediated autoimmune disorders.

Oricell Therapeutics recently raised over one hundred ten million dollars in a pre-IPO round to advance its CAR-T platform, including in vivo technologies aimed at solid tumors.

Even smaller players like ParcelBio are entering the space with programmable mRNA platforms designed for in vivo CAR-T applications in autoimmune diseases.

The common thread? Everyone is trying to turn CAR-T from a bespoke, patient-specific manufacturing ordeal into an off-the-shelf injectable drug.

So Who Wins?

If you’re asking me whether in vivo CAR-T will completely replace traditional ex vivo CAR-T, I’d say that’s unlikely in the near term. The two approaches will probably coexist for years.

Traditional CAR-T has a proven track record, especially for hematologic malignancies. The safety profile is well understood. The long-term outcomes are documented. For patients who can tolerate the waiting period and the lymphodepleting chemotherapy, it’s a highly effective option.

But for patients with rapidly progressing disease who don’t have weeks to wait? For patients who live in regions without access to advanced cell therapy manufacturing facilities? For patients who can’t afford the astronomical costs of personalized cell therapy? In vivo CAR-T could be a genuine breakthrough.

The field is moving fast. Clinical trials are expanding. The FDA has already granted fast track status to multiple candidates. Major acquisitions are happening at valuations that would have seemed absurd five years ago.

Will there be setbacks? Almost certainly. Safety remains a significant concern. The optimal delivery vector hasn’t been settled. Dosing control is an open question. And we still don’t know how durable the responses will be compared to traditional CAR-T.

But the direction is clear. The patient’s body as the bioreactor isn’t just a catchy phrase anymore. It’s becoming a clinical reality.

And for cancer patients who are running out of options, that reality can’t come soon enough.