Wall Street veteran analyst David Woo: AI trading and high interest rates—one of them must "give in" in the second half of the year.

```

In the first half of 2026, the global financial markets staged a brutal asset contest: traditional defensive safe-haven assets collectively plunged, while global real yields soared. Former top Wall Street macro strategist David Woo points out that the current macro market is stuck in an irreconcilable tug-of-war: AI trading has driven up real interest rates, yet stubbornly high rates are now biting back at AI trading. The market in the second half of the year has reached the tipping point of either/or, and the conflict between the two must have one side "admit defeat" to the market.

Key points summarized by Wall Street Insights:

Core market conflict: A decisive turning point will arrive in the second half. Either the AI bubble bursts under its own weight, giving defensive assets a chance to breathe; or real yields continue to climb, propelled by strong economic data, forcing the halt and destruction of AI and other risk asset trades.
The real driver behind high rates: The cause of this year’s exceptionally stubborn real yields is not the term premium, nor inflation panic sparked by high oil prices, but the far-surpassing economic activity data since April.
“Good news is bad news” regime intensifies: The market is now deeply embedded in this logic. Stronger economic data and hiring mean higher real yields, which directly translate into heavy downward pressure on equities and credit assets.
The Achilles’ heel of AI trading is “homogenization”: The greatest risk facing the AI concept is not liquidity, but "commoditization (homogenized low-price competition)." The rapid rise of new models in China and Japan is swiftly closing in on Western top standards, seriously weakening tech giants’ supernormal profits.

Real yields soar, defensive assets “slaughtered”

The first half of 2026 has been extremely brutal for defensive assets. Bitcoin became the worst-performing major asset, followed closely by depressed gold, while the yen/dollar exchange rate fell to its lowest level in sixty years.

David Woo believes that this weakness is not due to revived risk appetite (VIX and rate volatility remain high) but is driven by the mad re-pricing of real yields.

Since the start of the year, the U.S. 10-year real yield has surged by 33 basis points. Even as oil prices recently collapsed, real yields remain stubbornly elevated. This indicates that either real yields are about to face a dramatic downward correction, or stocks, credit assets, and emerging market currencies will face even heavier downward pressure.

Economy beats expectations, high rates are not supported by “tax refunds”

Many economists attribute the surprising resilience of the U.S. economy to the massive tax cuts and refunds from the “Big Beautiful Bill.” However, IRS data show the total amount of refunds issued in the 2026 filing season increased by only about $50 billion year-on-year—a drop in the ocean compared to a nearly $30 trillion economy.

David Woo says the real drivers are on the corporate side. Since January, U.S. core capital goods orders have doubled, and job growth has almost tripled. This is primarily driven by two policy forces:

Business tax provisions: The bill allows companies to immediately expense 100% of compliant equipment investments and domestic R&D, vastly reducing post-tax investment costs.
Policy uncertainty dissipates: Entering early 2026, as political uncertainty fades, businesses began to push forward previously shelved and delayed projects, directly igniting capital expenditures and later driving large-scale hiring.

This pent-up momentum will continue to support the economy in Q3, and upward inflation pressures will force the Fed to remain hawkish, driving real yields higher until it finally blows up the AI bubble.

Second half ultimate choice: The life-and-death tug-of-war between the AI bubble and high rates

Before real yields rise further, the AI bubble may well burst first under its own weight.

Currently, AI trading faces its toughest test since inception. OpenAI is reportedly considering delaying its IPO, while new AI large models emerging in China and Japan are rapidly approaching top Western standards. “Commoditization (homogenized low-price competition)” has become the greatest downward risk faced by all AI trading. As the major hyperscale cloud giants are about to report earnings, the market will face a severe test of real AI demand in the “post token maxing” era for the first time.

In addition, geopolitics (the Iran agreement is hard to maintain) may push up oil prices and interest rates again after July. Overall, in the second half either the AI bubble bursts first due to homogenization competition or demand collapse, or persistently high real yields point toward higher unknowns, forcing the halt of all risk assets. Until the AI bubble truly bursts and drags down real rates, it is still premature to blindly bottom-fish gold or short the dollar.

The following is the full speech, translated with AI assistance:

AI trading previously pushed up real yields, but now, higher real yields are threatening AI trading in turn. If the AI bubble bursts first, defensive assets will get a respite; but if real yields keep climbing, risk assets will be in big trouble. What is the market overlooking? Let’s make some predictions now.

The first half of 2026 is behind us, and it has been extremely brutal for defensive assets (safe assets). Bitcoin has become the worst-performing major asset, closely followed by gold. Meanwhile, the yen/dollar exchange rate has plunged to its lowest level in sixty years. Usually, poor performance of defensive assets indicates a sharp cooling of risk aversion and reduced demand for safe havens. However, the real situation in the first half of 2026 is different. The panic index has retreated from its highs, but remains above the level at the start of the year; both rate volatility and oil volatility are still high.

In fact, the weakness of defensive assets is mainly driven by the re-pricing of risk-free rates—especially real yields. Since the start of the year, the U.S. 10-year real yield surged by 33 basis points, nearing the highest point in this expansion. This repricing is not only happening in the U.S. Since the first U.S.-Israel attack on Iran at the end of February, real yields in nearly every major developed country except Canada have been climbing.

During the war and the initial ceasefire period, real yields and oil prices moved closely together because the market expected high energy prices to force central banks into tighter policy to prevent second-round inflation effects. But this logic broke in the past few weeks—oil prices have collapsed, yet real yields stubbornly remain high. This leads to two possibilities: either real yields are about to undergo a correction (in which case defensive assets will rebound significantly in the second half), or a new force is strongly supporting them. If real yields keep drifting higher, consequences for risk assets will be severe—stocks, credit assets, and emerging market currencies will be pressured ever lower.

Real yields depend on expectations of actual economic growth, monetary policy outlook, and term premium. Given that term premium has declined this year, we can rule it out as the main driver behind high real yields. What about growth expectations? Consensus forecasts for 2026 US GDP growth have been lowered from 2.4% at the start of the year to 2.1% now. However, since April, actual economic activity data has repeatedly exceeded expectations. In fact, our "weighted US economic activity surprise index" is near its highest level in the past five years.

My analysis shows that when inflation is above the Fed’s 2% target (as now), economic activity beating expectations is strongly linked with rising real yields. This is unsurprising—when inflation is high, the Fed will react more aggressively to good growth data. Thus I believe, though real yields are high now, it is not without reason. This matches the market pricing for “at least one Fed rate hike before year-end.”

At this moment, the market is stuck in a “good news is bad news” regime. Stronger economic data means higher real yields, and higher real yields mean lower stocks. So the multimillion-dollar question for the second half is: Where will real yields go? This will depend on whether growth keeps beating expectations and how inflation evolves. Let’s make some predictions.

Like many economists, I am surprised at the U.S. economy’s tremendous resilience in the past three months. A popular explanation says consumers survived the oil shock because of the huge tax cuts and refunds from the Big Beautiful Bill. However, the numbers don’t support this story. According to IRS data, by May 8, the total refunds issued in the 2026 filing season only increased by about $50 billion versus 2025. Against a nearly $30 trillion U.S. economy and roughly $20 trillion annual consumption, this is a drop in the ocean.

So what has driven the surge in business capex and hiring since January? Since the start of the year, core capital goods order growth has doubled, and the pace of job growth has almost tripled. AI is clearly the main driver of capex, but it’s hard to say it explains the job boom (tech sector is still laying off). I believe, aside from AI, two forces play key roles:

First is the business tax provision in the Big Beautiful Bill, which allows businesses to immediately expense 100% of equipment investments, domestic R&D, and certain manufacturing buildings, vastly lowering post-tax investment costs and optimizing capex economics. These provisions should have erupted in late 2025, but amid trade wars, prolonged government shutdowns, and extreme Fed uncertainty, companies postponed investment and hiring.

Second is the fading of policy uncertainty. As the policy fog dissipated in early 2026, companies began pushing forward shelved projects, directly igniting capex and, with a lag, hiring. With other conditions unchanged, these two forces will keep supporting the economy into Q3 as backlog investment and hiring ramp up. (This doesn’t even count most unissued tariff rebates.) Against this backdrop, inflation faces upward pressure, forcing the Fed to keep hiking, meaning real yields could climb further, until they finally burst the AI bubble.

However, “all else unchanged” is a strong assumption; the AI bubble may well burst under its own weight before real yields rise further. There are reports that OpenAI is considering delaying its IPO, while new AI models in China and Japan approach Mistral’s standards. This reinforces my core view: commoditization (homogenization and low-price competition) is the biggest downside risk for AI trading. Soon, we’ll get Q3 AI demand guidance from the major cloud giants in the post-token maxing era, but these earnings releases are still about 3 weeks off.

On geopolitics, I remain convinced the current agreement cannot survive the planned 60 days of negotiation. If I’m right, oil prices will likely rise again (possibly just after July 4th), pushing real yields higher. My overall judgment: either the AI bubble bursts itself, or real yields point toward higher unknowns. That’s why until the AI bubble truly bursts, I’m not ready to buy gold or short the dollar. But we may soon reach a tipping point where “something has to give”—which will undoubtedly benefit overall market volatility. Thank you for listening.

Key term explanations:

Real yield: The bond return after deducting inflation expectations. It is the core indicator measuring true borrowing cost in society. When real yields rise sharply, the opportunity cost of holding assets that don’t pay interest increases, putting huge pressure on them.

AI trading: Refers to capital flooding and over-concentrating in the AI industry chain, triggering stock surges.

Panic Index (VIX): Compiled by the Chicago Board Options Exchange, reflecting the expected implied volatility of the S&P 500 over the next 30 days. The higher the index, the greater the expected market volatility and fear among investors.

Term premium: The extra interest demanded by investors for holding a bond long-term (e.g., future uncertain inflation, rate changes, liquidity risk).

“Good news is bad news” regime: A unique macro market mode. In this mode, strong economic data makes markets worry central banks will keep rates high or hike again, thus causing stock market selloffs.

Capital expenditure (Capex): Funds invested by companies in purchase, upgrade, or maintenance of fixed assets. It is a barometer of an economic entity's confidence and expansion intensity toward future development.

Big Beautiful Bill: The major fiscal bill mentioned in the video background. Its core is aggressive business tax cuts and fiscal stimulus, especially immediate expensing of equipment and R&D investment, greatly reducing corporate tax burden in the short term.

Commoditization: When a formerly high-tech, high-margin product or technology loses its uniqueness due to competitor influx and tech diffusion, and becomes a standardized, low-margin commodity where only price matters.

Hyperscalers: Specifically refers to tech giants providing ultra-large-scale cloud computing, data centers and network infrastructure (Microsoft, Amazon, Google, etc). They are currently the main buyers of AI computing power and chips.

Token maxing: In the context of large language models, a token is the smallest text unit the model processes. This term refers to the industry’s explosive phase of maximizing compute calls and generating massive token volumes, sometimes in a rather reckless manner.

Risk warning and disclaimerThe market carries risk, invest prudently. This article does not constitute personal investment advice, nor does it take into account the individual investment objectives, financial situation, or needs of specific users. Users should consider whether any opinions, views, or conclusions in this article fit their particular circumstances. Invest accordingly at your own risk. ```