Ask.com shut down today after 25 years. You probably don't care. But the search engine's closure is a signal about who survives when the cost of maintaining legacy infrastructure finally exceeds its revenue.
The company that built Ask—IAC—is "sharpening its focus." Translation: the search business stopped justifying the engineering overhead. A quarter-century of servers, indexing pipelines, licensing deals, and customer support for a product that couldn't compete with Google. At some point, you stop throwing money at a burning building and sell the land.
This matters because it's happening at exactly the moment when tech companies are facing the inverse problem: they've built AI infrastructure so expensive that they now *need* growth they don't have yet. Meta's capital allocation language ("less capital for others") signals contraction. Microsoft and Alphabet are filing insider trades while preferred stock issuances suggest balance-sheet stress. The Pentagon deals announced yesterday—integrated AI contracts with OpenAI, Google, Microsoft, Amazon—are real, but they're also the *only* large-scale non-commercial revenue stream these companies have locked in. The commercial side is still a bet.
Ask.com's shutdown is what happens when the bet stops paying. And right now, the VIX at 16.89 is elevated-but-calm, which means the market hasn't yet priced the infrastructure debt problem: companies spent 2024-2025 building AI systems expecting revenue models that either haven't materialized or are smaller than the capex required to support them. The 10Y yield at 4.4% and the inverted yield curve (10Y-2Y at 0.51%) mean the cost of borrowing to bridge that gap is no longer cheap.
The Friendster revival and the Pentagon deals both point in the same direction: the market is fragmenting. You've got legacy plays trying to resurrect community (Friendster), cutting-edge military contracts (Pentagon AI), and mid-tier infrastructure feeling the squeeze (search, legacy cloud, analytics). The companies that survive this next 12 months are the ones with either (a) a defensible moat in a high-growth segment (Nvidia, OpenAI, maybe Anthropic), or (b) revenue density high enough to justify infrastructure costs (Meta's advertising, Amazon's AWS margin structure).
The insiders know this. That's why Alphabet executives are trading now. That's why MSTR is issuing preferred stock. That's why the tech sector is in defensiveness mode while growth-at-any-price narratives—the ones that justified massive capex—are quietly being walked back.
Within the next 48 hours, look for either a correction in long-duration tech stocks (QQQ weakness) or a sharp tightening in credit spreads as the market prices the infrastructure debt question. The Fed hasn't cut rates; real yields are sticky above 1.0%. If tech can't justify capex at *these* rates, the repricing will be swift.