AI computing power theme is gaining momentum for a major counterattack! Non-farm payrolls only increased by 57,000, far below expectations. Has the hawkish narrative of the Federal Reserve been severely damaged?

Just as news emerged that Facebook's parent company Meta is preparing to lease/sell idle AI computing power infrastructure, triggering concerns about oversupply of AI computing power, causing a sharp drop in AI computing power investment themes and leading to a global technology stocks sell-off related to AI computing power infrastructure, a much weaker-than-expected US non-farm payroll data pushed back rate hike expectations, fueling a major rebound in the stock prices of these AI computing power industry leaders in pre-market trading. The latest US non-farm payrolls data for June showed a sharp slowdown in the US labor market hiring process, significantly restraining some of the momentum forming in this year's US non-farm payroll growth trend, even though the unemployment rate decreased. Data released by the US Labor Department on Thursday showed that after the significant downward revisions in the data for the previous two months, the number of non-farm payroll jobs increased by only 57,000 people last month, just about half of market expectations and far below the economists' widely expected range of about 110,000 to 115,000 people, and also lower than the revised 129,000 people in May (initially reported as 172,000). Due to an unexpected significant decline in the labor force participation rate, the unemployment rate fell from 4.3% in the previous statistical month to 4.2%. This report indicates that despite strong signs of hiring in recent months, the labor market still faces some challenges. Despite the resilience of consumer spending data related to energy shocks caused by the Iran war, American residents are pessimistic about the high prices and wage growth not keeping pace with inflation, which may also lead employers to be cautious in hiring. The combined downward revision of 74,000 jobs in April and May data shows that the strong employment trend in the previous months was significantly overstated. After the release of the non-farm payroll data, S&P 500 index futures turned sharply higher, with popular technology stocks associated with AI making a big rebound, including CoreWeave, which suffered a heavy blow yesterday, with its stock price rising by over 3% at one point, while US Treasury yields saw a significant drop. Investors and interest rate futures traders also significantly reduced their bets on Fed rate hikes this year. Betting on Fed rate hikes in the interest rate futures market has significantly contracted, with the expected number of rate hikes in 2026 reduced from three times to just once, and the initial expected timing of rate hikes pushed back from the previously most aggressive July to December or even early next year. The latest weak non-farm payrolls data, coinciding with the pessimistic tone of "Meta selling surplus and significant AI computing resources," has provided a macro-level narrative reversal catalyst for the AI computing power industry chain just hit hard. The weak non-farm payrolls reduce the tail risk of rate hikes, lower the discount rate of high-duration assets, and combine to distinguish between the two completely different narratives of "AI computing power oversupply panic" and "AI cloud service monetization capacity," significantly shaping the AI computing power industry chain. Meanwhile, Meta Platforms Inc. and Microsoft Corp. are among some large tech companies that are implementing layoffs, in part to offset massive spending on artificial intelligence. Employment in the information industry continues to decline, marking the 17th decline in the past 18 months. Employment in the financial activities industry remained virtually unchanged. This industry is a significant employer of white-collar workers, who are considered one of the groups most vulnerable to automation. Average hourly wages rose by 3.5% year-on-year. Economists are closely monitoring how labor supply and demand dynamics are affecting wages, especially as inflation begins to outpace wage growth in a range of industries. With peace talks between the US and Iran resuming and oil prices plummeting, global consumer confidence is on a significant path to recovery, which may encourage employers to accelerate hiring in the coming months. Another report released on Thursday showed that the number of initial jobless claims last week remained almost unchanged. Layoffs have been at low levels in recent years, contributing to what economists call a "low layoffs, low hiring" labor market. Economists from Bloomberg Economics remarked after the release of the non-farm payrolls report: "The June employment report sends mixed signals, but overall indicates a stable labor market. Although job growth was slightly below expectations and data for the past few months have been revised downward, the underlying trend remains strong and is still above most people's breakeven point expectations." "The June employment report is clearly disappointing, but this report should not shake anyone's view of the overall economic outlook," noted Neil Dutta, Chief Economist at Renaissance Macro Research, in a report. "The key significance of this report is that the labor market reflects the condition of the overall economy. Economic growth is uneven, so the labor market also has imbalances." For the hawkish monetary policy path of the Federal Reserve, not only did the non-farm payroll data affect the hawkish path, but the reopening of the Hormuz Strait also struck a blow at the hawkish path of the Federal Reserve. On Thursday, four super oil tankers, loading crude oil at Saudi Arabia's main oil and gas export hub, appeared in the Gulf of Oman. This is the largest number of tanker exits since the peace agreement took effect about two weeks ago. In addition, since Saudi Arabia resumed oil tanker loading in the Persian Gulf, its crude oil exports have surged to nearly pre-war levels. This further confirms that following the midterm peace agreement between the US and Iran, oil-producing countries in the region are in the midst of a recovery in supply. US CPI data for May showed that the energy index rose by 3.9% on a month-on-month basis, with gasoline prices rising by 7.0%, and the energy subcomponent accounting for over 60% of the overall CPI increase in that month. Over the past 12 months, energy prices have risen by 23.5%, and gasoline prices by 40.5%, indicating that previous oil price shocks have substantively pushed up inflation readings. The latest Energy Information Administration's Short-Term Energy Outlook also indicates that higher global oil prices are pushing up expectations for US refined oil prices; diesel and aviation coal wholesale prices have been significantly revised higher compared to the forecast from two months ago, and gasoline wholesale prices are also expected to rise significantly. Therefore, the unexpectedly weak non-farm payroll data, combined with the gradual recovery of the Hormuz Strait, has significantly reduced the urgency for the Federal Reserve to further hike rates, rather than immediately opening up a path for significant rate cuts. Only when further energy declines push down inflation expectations, and core services, wages, and housing inflation all cool down simultaneously, will the Federal Reserve have a more compelling reason to shift from a "highly restrictive rate" stance to a truly loose cycle. The narrative of the technology bulls regains control! Disappointing non-farm payroll data leads to a significant cooling of rate hike expectations and sounds the horn for a counterattack in the AI computing power industry chain Only 57,000 jobs were added in June, significantly below the latest survey's expected 115,000, and May's job additions were revised down from 172,000 to 129,000, indicating a significant cooling of labor demand from the previously strong state; although the unemployment rate fell to 4.2% and layoffs remained low, indicating that it is not yet a "recession confirmation," it is enough to weaken the urgency of the Federal Reserve to continue hiking rates; in addition, hourly wages were up 0.3% month-on-month and 3.5% year-on-year, not yet signaling full relief from inflation for the Federal Reserve, but enough to take away a key pillar for rate hike trades. Pre-market trading saw a collective rebound in the AI computing power industry chain, reflecting the market trading on a clearer logic: the growth trajectory of the US economy and labor market has not collapsed, employment has significantly cooled, energy risk premiums have fallen, the probability of further rate hikes by the Federal Reserve has been lowered, DCF discount rate pressures on local and global technology risk assets led by the AI computing power industry chain are being marginally released. As the weakening-than-expected employment report eases market concerns about further rate hikes by the Fed this year, for stock strategies, this means that short-term capital flows may shift from "defending against inflation, betting on rate hike cycles" to "betting on high-quality growth-oriented tech stocks, betting on the leading forces in the AI computing power industry chain that have real AI cash flows and are sensitive to rate expectations." The leverage and crowding of AI semiconductor trading themes, along with increased pressure on rising prices for consumer electronics leaders like Apple, highlighted by a single-day 7.9% drop in the Philadelphia Semiconductor Index and multiple violent fluctuations of over 5% within a month, demonstrate that the AI computing power industry chain associated with semiconductors has entered a high-volatility, leveraged and crowded long position, high expectation cashing pressure stage. Combining Meta's shift towards selling computing power and the entry of SoftBank's SB Neo into the US AI cloud sector, this also explains why institutional investors have recently begun to emphasize a shift to overly pessimistic bearish narratives such as "the end of the AI semiconductor trading boom" and "the gradual bursting of the AI bubble." However, the well-known Wall Street investment firm Nomura released a latest research report on Wednesday, refuting the "peak semiconductor" theory. The key to Nomura's refutation of the "peak semiconductor" theory lies not simply in saying that AI chips will continue to rise, but in pointing out that the demand for AI cloud infrastructure is shifting from a shortage of single-point GPUs to systemic component mismatches. According to Nomura's research framework, AI server revenue is expected to grow by 78% and 76% in 2026 and 2027, respectively, with the number of global data center projects increasing from 240 to 280, including around 50 gigawatt-level projects, with an expected 32GW of new computing power deployments in 2027 and 23GW already in visibility for 2028; but the real bottleneck is moving from GPU capacity and TSMC's advanced CoWoS packaging to wafer-level substrates, AI PCBs, copper-clad laminates (CCL), electronic fabrics, MLCCs, glass substrates/ABF substrates, IC carriers, high-end capacitors, power management chips, and data center optical high-speed interconnect components. McKinsey's medium- to long-term calculations also support Nomura's emphasis on this direction: by 2030, the global value chain of computing power will require an investment of about $52 trillion to meet AI-related demand, corresponding to a demand for approximately 156 gigawatts of AI-related data center capacity. This means that the main theme of semiconductor trading is not "peaking," but rather "rotating the shortage positions": from GPUs to HBM, from advanced packaging to substrate materials, semiconductor equipment and raw materials, data center CPUs and optical interconnect systems, then to power, liquid cooling, networks, and cloud scheduling software, profit hikes and price hike expectations may still be the strongest catalyst for the core hardware chain of the AI computing power industry. On a deeper level, Meta and SoftBank Group founded by Masayoshi Son may be betting on a long-term scarcity of computing resources in the AI workload shifting from training centers to inference centers. McKinsey estimates that by 2030, global data centers will require about $6.7 trillion in investment to meet computing demands, with capital spending on AI inference end processing loads related to data centers accounting for about $5.2 trillion; the International Energy Agency forecasts that global data center electricity consumption will double to about 945 terawatt-hours by 2030, with annual electricity growth driven by AI acceleration servers expected to be around 30%; Goldman Sachs also expects US data center power demand to rise from 31 gigawatts in 2025 to 66 gigawatts in 2027.