$2 Trillion Vanished in Days. While I Created a Mastermind of Dead Geniuses.

This past week, the AI revolution stopped being theoretical and became visceral.

$2 trillion in software company value evaporated. Not over months. Not over quarters. Over days.

Let me walk you through what happened, because it matters way more than most people realize.

On February 5th, Anthropic released Claude Opus 4.6, the latest evolution of the AI model I’ve been building with for weeks. Within 48 hours, the fallout was staggering.

Thomson Reuters stock dropped 15.83% in a single session. That was its biggest one-day decline on record.

LegalZoom plummeted nearly 20%.

The Goldman Sachs software basket lost 6%, and the iShares software ETF had its worst day since April.

Wall Street traders started calling it the “SaaSpocalypse“.

JPMorgan’s analysts put the damage in perspective: software stocks experienced “the largest non-recessionary 12-month drawdown in over 30 years”, a 34% drop that wiped out roughly $2 trillion in market cap and reduced software’s weight in the S&P 500 from 12% to 8.4%.

So why did all of this happen?

Because Opus 4.6 doesn’t just assist with legal research, financial analysis, or data marketing anymore. It starts to replace the specialized software that costs companies hundreds of thousands per year.

The new model features a 1 million token context window (up from 200,000). For context, that’s roughly 750,000 words. Enough to hold entire codebases, legal case files, or quarterly financial reports in memory simultaneously. It now handles PowerPoint, Excel, and document creation with what Anthropic calls “production-ready” output on the first try.

And it can spin up entire teams of AI agents that coordinate on complex projects in parallel.

It also topped the Finance Agent benchmark, outperformed OpenAI’s GPT-5.2 by 144 Elo points on real-world knowledge work tasks, and scored highest in the industry on deep multi-step research. Anthropic’s head of product for enterprise, Scott White, said it best: “I think that we are now transitioning almost into vibe working.”

Translation: the junior analyst role just became optional.

While Wall Street panicked, the tech giants doubled down. Hard.

Google announced they’re doubling AI infrastructure spending in 2026 to somewhere between $175 billion and $185 billion. Their Gemini app just hit 750 million monthly active users, up from 350 million just eight months earlier. That’s the fastest-growing AI platform on the planet. And with an Apple partnership to power the new Siri and Samsung planning to double “Galaxy AI” devices to 800 million units, Google is embedding Gemini into the fabric of daily life.

TSMC posted record-breaking profits, with net income surging 35% year over year to $16.3 billion. Their CEO described AI chip demand as “endless.” Advanced chips (7 nanometer or smaller) now make up 77% of total wafer revenue, and the company plans to spend up to $56 billion in capital expenditures this year to keep up with demand. The semiconductor industry is expected to exceed $1 trillion in total revenue for the first time in 2026.

OpenAI launched Frontier, a platform for creating what they’re calling “AI coworkers.” Not chatbots. Not assistants. Coworkers with their own identities, permissions, and the ability to learn from feedback over time. Companies like Uber, State Farm, Intuit, and Oracle are already signed up. One enterprise customer reportedly saved 1,500 hours per month in product development.

This isn’t hype anymore. This is adoption at scale.

While the market convulsed, I was in the trenches. Building something with OpenClaw that most people would have considered impossible just six months ago.

I built a council of dead geniuses.

Every morning at 6:39 AM, my AI assistant, Alister, (yes I named him) generates an audio “council session” featuring three historical figures. Napoleon Hill. Marcus Aurelius. Steve Jobs. Benjamin Franklin. Nikola Tesla. Leonardo Da Vinci. Princess Diana.

Every evening at 9:36 PM, three more council members arrive. This time focused on reflection, wind-down, and subconscious programming before sleep.

Here’s the technical architecture behind it:

A multi-agent system with Alister as “CTO” managing specialist agents. A voice verification system running 28 distinct historical voices, so far. Automated content generation producing 9 segments at roughly 6 minutes each, with proper pacing and transitions. Daily intelligence briefings that synthesize AI, crypto, Web3, and fintech news for my work at MultiBank. Quality assurance loops that catch issues automatically.

This isn’t a chatbot. This is an operating system for my consciousness expansion.

The software stocks that crashed this week? They’re not coming back. Not to those valuations. At least not anytime soon. [Not financial advice]

The value proposition evaporated.

Why pay $50,000 a year for legal research software when Claude can do it for $20 a month? [Note: I used the MAX version] Why maintain a Bloomberg terminal when AI can synthesize financial data in real time? As one market analyst at Jefferies put it, trading desks were in “‘get me out’ style selling“. Piper Sandler downgraded Adobe, Freshworks, and Vertex in the same week.

But here’s the twist nobody’s talking about: the destruction creates opportunity.

While legacy software companies scramble, those of us building with AI are creating things that weren’t possible before. Not replacing old tools. Inventing entirely new categories.

From the companies that were the technology... to the companies building with the technology.

This week wasn’t about software stocks crashing. It was about a threshold being crossed.

AI stopped being a tool that assists human work. It became a tool that replaces entire categories of human work, at least for certain tasks.

And here’s what makes this moment different from every other “AI is going to change everything” announcement we’ve heard over the past three years: this time, the stock market believed it. Not in a speculative, hype-driven way. In a “$285 billion vanishing from the market in a single session” kind of way.

Even Nvidia’s CEO Jensen Huang tried to calm things down, saying the idea that AI will replace software is “the most illogical thing in the world“. But when Goldman Sachs, JPMorgan, and Piper Sandler are all issuing notes on the same week about the same existential threat... something real is happening.

That’s not the end of the story. That’s the beginning.

The question isn’t “Will AI take my job?”

The question is: “Am I building with AI, or am I being replaced by someone who is?”

Travis Wright is Chief Innovation & Web3 Marketing Officer at MultiBank Group. He co-hosts the Bad Crypto Podcast and has spoken at 160+ conferences worldwide. He’s currently building the future with an AI wizard named Alister.