Rates Are Rising and Big Tech Is Holding. Here Is Why That Is Not a Contradiction
The 10-year yield tells you the price of money. It does not tell you how much money is left. Put the two together and the confusing headlines of the past three years line up into one picture.
Two questions that explain most of the stock market
First: what is the price of money? Second: how much money is left in the system?
The headline number for the first question is the yield on the 10-year Treasury note. For the second, you look at the Fed's balance sheet, the Treasury General Account, the overnight reverse repo facility and the shape of Treasury issuance.
The common mistake is to watch only one of them. "Rates are up, so tech is finished." Or the reverse: "There is plenty of liquidity, so everything goes up." Both forces are running at the same time, all the time.
The 10-year is the rent table for the whole market
Suppose the US government will borrow for ten years at close to 5 percent with essentially no credit risk. An investor asks the obvious question: how much more do I need to earn before it is worth taking equity risk instead?
That question is the starting point of the equity risk premium. Stocks cannot be equally attractive when Treasuries pay 2 percent and when they pay 5 percent. The higher the risk-free yield, the more a stock has to prove, either with faster growth or with a cheaper price.
That is why a rising 10-year usually hits high-multiple growth names first. The further out a company's profits sit, the more a higher discount rate shaves off their present value.
And here is the exception that matters most. If yields are rising because the economy is strong and productivity is improving, corporate earnings can rise along with them. The discount rate goes up, but earnings growth outruns the drag.
Not every 5 percent is the same 5 percent
The good kind. AI investment, productivity gains, resilient consumption and employment lift growth expectations. Treasury yields rise, but revenue and earnings forecasts are revised up alongside them. Strong companies offset the rate burden with profit growth.
The bad kind. Inflation picks back up, deficits push more long-dated supply into the market, or investors demand a bigger term premium for holding duration. Earnings estimates stay where they are while the discount rate climbs.
In practice the 10-year does not decompose cleanly into "growth plus inflation." It contains the market's expected path of future short-term rates plus a term premium, the compensation for holding long bonds. The New York Fed splits long-term Treasury yields exactly that way.
So the investor's question cannot stop at "where is the 10-year?" It has to go one step further: is this 30 basis point move about growth, or about inflation, fiscal supply and term premium?
Why higher rates can push money toward big tech
When Treasuries pay 4 or 5 percent, capital stops being available to just any business. A company losing money today, with an uncertain path to profitability years out, has to answer a hard question: why buy this instead of a government bond paying 5 percent safely?
Money compresses. Firms with weak cash flow, firms that need to keep raising capital, firms whose value sits mostly in distant expectations all get squeezed. Firms already generating enormous free cash flow, with no dependence on external funding, get relatively stronger.
This is the flight to quality. High rates are not uniformly bad for technology stocks. They widen the gap inside the sector between the companies that make money and the companies that burn it.
That alone does not explain 2023 to 2025
The Fed raised rates and simultaneously ran quantitative tightening, shrinking its holdings. By the textbook, market liquidity should have drained fast. Instead bank reserves held up far longer than expected, and tech stocks stayed strong.
This is where the Treasury enters the story.
Money market funds were sitting on an enormous pile of cash parked at the Fed's overnight reverse repo facility. When the Treasury leaned heavily on T-bill issuance, those funds rotated out of RRP and into short-dated government paper that paid more. Fed research describes much of the post-2023 decline in RRP balances as money market funds shifting cash into Treasury bills and private repo.
The key point: even with QT running, bank reserves did not fall one for one. Part of the shock was absorbed first by RRP, a cash reservoir sitting outside the banking system.
So "the Fed is draining liquidity" was never a sufficient description of that period. The central bank's tap was tightening, but water was moving between reservoirs inside the financial system.
The TGA is the Treasury's checking account
The Treasury General Account is the government's cash account at the Fed. Nothing about it needs to be complicated.
When the government collects taxes or issues debt and pulls cash into that account, money leaves private financial markets. That is generally a drag on liquidity.
When the government spends, TGA balances move out into the accounts of companies and households. Cash flows back into the system.
Which means the question is not only what the Fed did. It is also how much cash the Treasury is hoarding and how much it is releasing.
In 2026 the rules changed
Applying the old playbook to the current market produces errors. The Fed ended balance sheet runoff on December 1, 2025. Through 2026 it has been buying Treasury bills again to keep reserves ample. As of early July, reserves stood at roughly $3.1 trillion, and ON RRP was effectively near zero on most days.
The giant shock absorber that cushioned the market in 2023 and 2024 is gone.
From here, the question is no longer how much more cash can come out of RRP. It is whether bank reserves are adequate, how large the TGA grows, how much long-dated paper the Treasury issues, and how much the Fed buys to manage reserves.
The buffer used to be there. It has been spent. The same size of Treasury issuance, or the same build in the TGA, can now hit financial markets more directly than it did three years ago.
What to watch with the 10-year near 5 percent
In September 2026 the US 10-year yield approached 5 percent again. That number alone is not a sell signal.
If oil prices, inflation worries, fiscal strain and long-end supply concerns are driving the move together, the rise skews toward the bad kind. Earnings estimates are not climbing at the same speed.
If instead AI-driven productivity and strong growth keep pushing earnings estimates higher while long rates stabilize, the largest and most cash-generative tech companies can hold up even in a high-rate environment. The Fed itself assessed that the S&P 500's gains in the first half of 2026 owed a great deal to strong corporate results and AI optimism.
The call an investor has to get right is not simply whether rates go up or down. It is the combination: the character of the rate move, the pace of earnings, and how much spare cash is left in the financial system.
