• Tech shares push higher even as 10-year Treasury yields hit 5.306%, a 24-year high.
  • Micron's blowout earnings and $32 billion in supply commitments fuel confidence in AI infrastructure demand.
  • Investors rotate within tech, with sector funds seeing $3.79 billion in outflows while broad large-cap funds attract $19.33 billion.

AI Enthusiasm Trumps Valuation Pressure

Investors are shrugging off the highest long-term borrowing costs in nearly two decades, betting that the artificial-intelligence boom will deliver earnings power strong enough to justify stretched valuations. The Nasdaq Composite reached record highs last week even as the 10-year Treasury yield climbed to 5.306%, its highest since mid-2007, according to market data.

The immediate catalyst was Micron Technology (MU)'s fiscal fourth-quarter results. The memory-chip maker reported $54.23 billion in revenue and guided to approximately $61.5 billion for the current quarter. More importantly, Micron disclosed $32 billion in financial commitments under long-term supply agreements, up from $22 billion in June. That figure is being read as hard evidence that AI data-center demand remains insatiable.

"If you think the benefits of AI are essentially infinite, so what if you have to pay more to borrow money?" said Steve Sosnick, a strategist at Interactive Brokers (IBKR), capturing the prevailing mood.

Bond Market Tightening Intensifies

The yield surge is not trivial. The 10-year Treasury yield climbed 87 basis points during the July–September quarter, the largest quarterly increase since 1994. Several forces are at work: Brent crude rose 42% amid stalled Middle East peace talks, the Federal Reserve raised rates by a quarter-point in September to a 3.75%–4.00% range, and the U.S. budget deficit is expected to hit roughly 6% of GDP, or $1.9 trillion, this year.

Adding to the supply pressure, AI hyperscalers have issued nearly $220 billion in bonds by mid-August, more than double the $108 billion issued across all of 2025, according to LSEG data. AI infrastructure is not just driving equity demand—it is also competing for capital in debt markets.

Selective Buying Emerges

Despite the bullish tone, fund flows reveal a more nuanced picture. U.S. equity funds took in $20.6 billion in the week through September 30, their second consecutive weekly inflow. But technology-sector funds saw $3.79 billion in net outflows, while large-cap funds attracted $19.33 billion. Investors appear to be growing more selective about valuation, profitability, and balance-sheet exposure within the tech space.

The market's tolerance for high yields rests on the belief that AI will lift productivity and corporate earnings enough to offset higher discount rates. But higher real yields—with 30-year real yields near 3%, an 18-year high—make future cash flows less valuable today and offer more attractive inflation-adjusted returns from bonds.

Regulatory Clouds Gather

A potential counterweight emerged this week as the Federal Trade Commission opened an industry-wide inquiry into Anthropic, OpenAI, and other AI laboratories over consumer dangers from agentic AI. The agency plans information demands and executive testimony, according to people familiar with the matter. New compliance or safety requirements could affect deployment speed and costs, though it is too early to quantify the impact.

Internationally, the divergence is playing out across markets. Japan's Nikkei rose more than 3%, South Korea's KOSPI gained 1.7%, and Taiwan's benchmark increased 0.8% following Micron's results, while markets with less AI exposure lagged.

The Delicate Balance

The central question is no longer whether AI investment is large—it clearly is—but whether eventual earnings and cash flows can justify valuations if real borrowing costs remain structurally higher. For now, investors are betting they can.

Correction: An earlier version of this article misstated the week of the fund flow data. It was the week through September 30, not September 23.