Ouyang Xiaohong/Text

One

-26%,+80% - On July 1st, Ms. Yang, a small and medium-sized investor, checked the returns on two assets in her account.

The former is her gold investment income; The latter is a storage concept stock that she has been trying out for a while, with a promising rise. And some of the WeChat group members she knew were already "crazy" between consecutive board meetings and risk warnings: listed companies urgently announced risk warnings, institutions began to panic, but investors became increasingly excited as prices rose.

This is not an isolated case. Whether the AI (artificial intelligence) bull market or the AI foam has reached this point, it has become a serious problem.

On June 28, 2026, the Bank for International Settlements (BIS) warned in its annual economic report that AI investment itself may improve productivity, but the scale, speed and market optimism of the current boom are close to the early form of the infrastructure foam in history. Once the return realization is slower than expected, overinvestment, valuation drawdown, and financing chain contraction may amplify each other through stock indices, private credit, and non bank systems.

Three days later, at the European Central Bank Central Bank Forum held in Sintra, Portugal, Federal Reserve President Walsh once again refused to provide forward guidance on interest rates, emphasizing that the next Federal Open Market Committee will wait for officials to "close their doors" before fully debating. The market wants a roadmap, but he offers new game rules: fewer previews, more data; Less appeasement, more disagreement.

More importantly, Walsh did not treat AI solely as a technology stock story. He acknowledges that AI may expand the economic supply capacity and profoundly change the way monetary policy is calculated; But whether it will lower inflation or drive up demand and costs due to investment frenzy still needs to be answered by data.

The voices on Wall Street are equally noisy. Ev Randle, a partner at Silicon Valley's veteran venture capital benchmark, recently warned that AI is causing the old coordinates of venture capital and growth stock investments to fail. In the past, the bigger the enterprise grew, the lower the risk should be: first verify the product market fit, then run through unit economics, and finally confirm the market ceiling. But in the era of AI, this old path has been disrupted. A company can quickly achieve $1 billion in revenue, but it has not yet proven whether the profit model is stable or whether product differentiation can be sustained in the long run.

Scale no longer automatically means certainty, and growth is no longer naturally equal to a moat.

Perhaps those investors who follow with real money and silver are hearing the roar of the era train. The ground is shaking, the platform is swaying, and the economy is racing against time to pave the way to the future.

The ancient gold has also been heard. However, as a naturally interest free asset, it cannot withstand the knife of interest rates. The conflicts in the Middle East and the restructuring of the Federal Reserve by Walsh have collectively pulled the market back to a harsh reality - money is no longer cheap.

The international gold price fell from its pedestal as a result, dropping from a high of nearly $5600 per ounce at the beginning of the year to around $4000 per ounce. Mainstream institutions are starting to lower their target prices, and the most steadfast bulls are no longer shouting so loudly.

On July 1st, the World Gold Council released its "Mid Year Outlook for the Global Gold Market in 2026", placing gold investment at a critical juncture: in the second half of the year, gold prices will be simultaneously driven by geopolitical factors, interest rate environment, and investor sentiment; If the price continues to fall below $4000 per ounce, it may trigger further selling, but deeper declines may also trigger bargain hunting for long-term funds.

However, at the Jingshun Mid Year Investment Outlook Media Conference on June 17th, Zhao Yaoting, Jingshun's global market strategist for the Asia Pacific region, was not pessimistic in the face of a gold pullback. He believes that this is more like a health reset. The reason why gold did not continue to rise with the conflicts in the Middle East is not difficult to understand: when high-yield assets regain their attractiveness, gold that does not generate dividends or interest will have a decrease in short-term attractiveness. But central banks around the world are still buying gold, and the long-term reserve logic of gold has not disappeared.

Ms. Yang also vaguely felt that the story of gold was far from over. Gold will always shine, but this time it has to pay for the actual interest rate first.

Two

As of June 30, 2026, several mid year performance reports put together will create a "K-shaped" differentiation picture of the global financial market.

According to the A-share performance report compiled by CITIC Securities, the core stock market index has almost gone up across the board: 9 out of 10 stock market indices have risen, Sci Tech Innovation 50 has risen by over 60%, ChiNext Index has risen by over 30%, and CSI 500 has risen by over 20%. However, upon closer inspection of the industry and individual stock levels, less than one-third of the industries have risen, with electronics and communications experiencing their best first half market performance in nearly a decade, while consumer services, retail trade, food and beverage, automotive, and transportation have experienced their worst first half market performance in nearly a decade; More than 300 stocks doubled, reaching a new high in the same period of 10 years, but the median of individual stocks fell by more than 15%, with only about 30% of stocks rising.

In the same first half of the year, it feels like a bull market in the index and a missed opportunity in many people's securities accounts.

Looking at other financial markets, it also points in a similar direction. On the stock index list, the Korea Composite Index (KOSPI) rose 101.13% in the first half of the year, the Philadelphia Semiconductor Index (SOX) rose 93.55%, the Nikkei 225 Index rose 39.18%, and the Nasdaq and S&P 500 indices rose much lower than markets with higher exposure to AI hardware.

This transcript looks like a global stock index ranking, but in reality it is like an AI exposure ranking table. The closer AI is to hardware, the higher the market growth; The more AI stays within the platform, application, or traditional weight, the gentler the market growth.

Three

Take a closer look at several sets of slices showing how the "K-shaped" gravity develops.

The first set of slices is gold. According to data from the World Gold Council, central banks around the world increased their net holdings of gold by approximately 244 tons in the first quarter. As of the end of May, the People's Bank of China has increased its holdings for 19 consecutive months. The central bank's buying has not stopped along the way of the gold price retracement. On the other hand, Bitcoin, known as the "digital gold," has experienced a strong leverage flattening, and long-term believers have also begun to let go. Also known as gold, facing real interest rate revaluation, falling from high places; The difference is that under real gold lies the central bank, while under "digital gold" lies mainly narrative.

The second set of slices is in US dollars. According to Jing Shun's judgment, the US dollar index may further weaken in the next six months. This is contrary to the market sentiment in the first half of the year. At that time, the rebound of the US dollar index and the strong return of the US dollar almost became an explicit narrative.

But Zhao Yaoting's reason is that the valuation of the US dollar is still expensive, and there is room for recovery in emerging market currencies; If the risk of conflict in the Middle East eases and oil prices fall back to a more controllable range, inflationary pressures will also ease. At that time, Walsh's hawkish stance would lose one pivot.

The third group of slices is AI. Jing Shun regards AI as one of the most important investment themes in the second half of the year, and also suggests that one of its biggest risks is concentration. Zhao Yaoting mentioned that the current profits in the US market are highly concentrated in a few cloud factories, and the profits in Asia, except for the Japanese market, are also highly dependent on a few semiconductor companies. If the profits or capital expenditures of one or two leading companies fall short of expectations, it may affect the risk appetite of the entire market. However, there is no foam in the AI sector, nor can it be simply compared with the Internet foam in 2000.

This is not an ordinary industry rotation, but a concentration at the index level.

Wu Ge, Chief Economist of Changjiang Securities, pointed out in the research report "The Other Side of Technology" that technology financing is basically synchronized with stock prices and leads investment by about a year. As long as external financing continues to follow, even if free cash flow falls, valuation and capital expenditures may still be supported. Whether financing has peaked depends on two variables: technology expectations and macro interest rates.

The light of the first half of the year came from the steep upward trend of storage chains and computing power chains in South Korea and China; The thunder of the second half of the year may be hidden in the capital expenditure guidelines of the seven major cloud factories.

Four

It is not difficult to see that when several sets of slices are put together, the financial market in the first half of 2026 is not a few isolated news stories.

The decline of gold is the scar left by real interest rates on interest free assets. Bitcoin's decline is the floor exposed when narrative assets are liquidated. The rebound of the US dollar is a reflection of the expectation gap in short-term prices. The sharp rise in AI hardware is supported by capital expenditures and financing cycles.

On the surface, everyone goes their own way; Looking deeper, several paths lead to the same issue: the cost of money and the credibility of the future.

The downward trend of real interest rates is direct. Houses, cars, gold, consumption, and long-term assets are all sensitive to interest rates. The more expensive the money, the heavier the discount of future cash flows, and the higher the opportunity cost of interest free assets.

The pressure on the upper edge of real interest rates comes slower. AI and hard technology are still hanging high at the moment, as they have another self circulation: financing supports stock prices, stock prices support valuations, and valuations support the next round of capital expenditures. As long as the capital expenditure of cloud factories continues to rise and the order and revenue curves continue to steep upward, the market may be willing to temporarily ignore the drag of high interest rates.

But this sling also has a cost. Some analysts believe that when real interest rates remain high enough, the financing chain will be tested; When the guidance for AI capital expenditure becomes loose, the market will ask again: whether income can be deposited into cash flow, whether scale really reduces risk, and whether growth equals a moat.

A few months ago, what Ms. Yang saw was still a faintly visible "amphibious" smile curve of funds.

At one end, there are gold, treasury bond and other assets that can give people a sense of security; On the other end, there are waves of AI, technology, and computing power that make people hesitant to get off the road.

At that time, the market curve was still U-shaped. In the middle of the year, a piece of the left end collapsed first, and the international gold price fell below $4000 per ounce. Safe haven assets began to pay for real interest rates. Although there is some oscillation on the right end, it is still rising - AI hardware, semiconductors, and storage chains continue to absorb liquidity. The original "smile" curve has been cleaved by the knife of interest rates into a more difficult to place K-shaped curve.

Some analysts suggest that in the second half of 2026, we need to focus on three checkpoints.

The first is oil prices and inflation. The receding geopolitical risks and falling oil prices should have cooled inflation; As long as the stickiness of core inflation persists, the biased eagle response function will be confirmed.

The second is Microsoft, Amazon, Google Meta、 Oracle, as well as major cloud companies such as Alibaba Cloud and Tencent Cloud, will submit their Q2 capital expenditure guidelines in July and August. If the guidance remains stable, the 'siphon effect' continues; If the guidance is significantly lowered, the market will doubt whether the upper edge "sling" can continue to hang high.

On the morning of July 2nd in the Asia Pacific market, AI concept stocks practiced another adjustment. The news that Meta plans to sell excess AI computing power has led the market to re evaluate the narrative of "computing power scarcity", putting pressure on South Korean storage leaders and A-share AI concept stocks simultaneously. It does not necessarily mean that AI demand has been falsified, but it reminds the market that high-level crowded AI trading, every time a new variable appears, will re-examine the relationship between computing power, capital expenditure, and valuation.

The third is gold. The World Gold Council believes that under benchmark conditions, international gold prices may fluctuate around $4100; If geopolitical risks escalate, the economy weakens, or expectations of interest rate cuts advance, gold may regain its upward trend; If the US dollar strengthens and the Federal Reserve raises interest rates beyond expectations, falling below $4000 may trigger further selling.

According to the Research on Global Sovereign Asset Management by Jingshun, 61% of the central banks believe that the US debt level is weakening the long-term reserve position of the US dollar, up from 20% in 2024; Under the constraint of the lack of sufficient large-scale alternative assets, gold remains the most realistic beneficiary of reserve diversification, with over one-third of central banks expecting to continue increasing their gold allocation in the next three years.

The market performance in the second half of 2026 is not only with the Federal Reserve or Nvidia, but also hidden in the central bank's purchase orders, the capital expenditure table of cloud factories, and the long-term allocation of sovereign funds.