Killer Bees — Market Brief February 12, 2026

The Sting

The auto industry just admitted the EV dream cost $53 billion more than it was worth. Meanwhile, software stocks are getting destroyed by the very AI that was supposed to save them. These stories look unrelated. They're not.

Deep Dive #1: The $26 Billion Confession

Stellantis dropped a bomb on Friday: a €22.2 billion ($26.5 billion) write-down, the largest one-time charge in automotive history. The stock cratered 24% in a single session, closing at $7.28 wiping out roughly $6 billion in market cap.

Here's what the headlines said: "Stellantis takes charge, retreats from EVs." Here's what actually happened: new CEO Antonio Filosa stood in front of investors and called his predecessor's entire strategy a capital allocation catastrophe. The charges cover canceled EV programs (including the once-vaunted Ram 1500 REV), broken supplier contracts, and the cost of bringing back V8 engines that were never supposed to return.

What most coverage missed: this isn't about one company. Stellantis joins Ford ($19.5 billion in EV write-downs) and GM ($7.6 billion) in what analysts are now calling "the single biggest capital allocation mistake in the history of the automotive industry." Combined, that's over $53 billion in shareholder value vaporized because executives collectively misread how fast consumers would abandon internal combustion. The irony: EVs still represent just 7.7% of U.S. car sales.

The market reaction tells you something. STLA is now trading at a P/E of 4.3x. Analysts still rate it a "Buy" with a $12 median price target 65% upside. Either the Street thinks this kitchen-sinking marks the bottom, or they haven't updated their models yet. The smart money question isn't "is EV dead?" (it's not). It's this: what happens when the pendulum swings too far back? Stellantis is betting on range-extended hybrids and bringing HEMIs back to Ram trucks. If EV adoption accelerates in 2027, they'll be writing another check.

Deep Dive #2: The Claude Crash

Something broke in software stocks this month, and it has a name: Claude.

The iShares Expanded Tech-Software ETF (IGV) has fallen roughly 30% from its October high, with a 14.5% January drawdown its worst month since October 2008. The 14-day RSI hit 19, the lowest since the U.S. credit downgrade crisis in August 2011. Tuesday's volume was the highest in the ETF's 25-year history.

The catalyst? Anthropic's Claude Cowork announcement. The AI tool automates complex professional workflows in legal and financial sectors exactly the tasks that justify SaaS subscription pricing. Suddenly, investors asked a question they'd been avoiding: if AI can do the work, why pay per-seat licensing?

The damage is widespread. Microsoft is down 26% from its peak. Oracle has been "taken to the woodshed" down 56% from September. Palantir, Salesforce, Adobe all in deep drawdowns. Goldman Sachs called the selloff "too broad" and JPMorgan sees "overly bearish outlook on AI disruption." But here's the thing: even the bulls aren't arguing the thesis is wrong. They're arguing the timing is wrong.

The weirdest part: the S&P 500 is within 2% of all-time highs. This isn't a systemic crash. It's a sector rotation and IGV is trading at 40x trailing earnings still not cheap by historical standards while regional banks, homebuilders, and energy stocks quietly rally. The market isn't panicking. It's reallocating. The question nobody's asking: if AI really does obsolete SaaS business models, where does the productivity gain accrue? Someone captures that value. So far, the answer seems to be "not the companies selling AI" (even Nvidia is treading water). That's either a buying opportunity or a warning that AI's economic story is more complicated than anyone modeled.

Chart of the Week: Software vs. Everything Else

IGV's divergence from the S&P 500

Software stocks have completely decoupled from the broad market. The S&P 500 has hit 58 all-time highs since the start of 2024. IGV is 30% below its peak and just recorded its highest weekly volume ever. Either this is the buying opportunity of a generation, or the market is telling us something about AI's winners and losers that the index composition hasn't caught up to yet. For context: IGV returned 41% in 2023, 52% in 2024, and -8% in 2025. The mean reversion was always coming the question was whether it would be gradual or violent. Now we know.

The Hive Mind

FinTwit is split between "software is uninvestable" and "this is 2022 energy all over again time to buy." Reddit's WallStreetBets is mostly distracted by bitcoin's collapse to $67,000, but the few software threads are surprisingly bearish. The interesting signal: options flow on IGV is actually balanced despite record volume. Retail is loud. Institutional positioning is quieter. That usually means something.

The Thread

Two stories this week: a $26 billion admission that the EV transition was oversold, and a software selloff driven by fears that AI was undersold. In both cases, the market is saying the same thing we got the timing wrong. The auto industry moved too fast toward a future consumers weren't ready for. Software investors moved too slow to price in a future that's arriving faster than expected. What both really reveal: markets are terrible at predicting adoption curves, and the cost of being early is indistinguishable from being wrong. The question for the rest of 2026 isn't "EVs or software" it's whether any consensus forecast deserves our confidence at all.