Japanese stocks rise early as chip shares lead, Nikkei up 0.4%

Semiconductor shares continued to recover from a sharp July selloff as investors added back exposure to chipmakers tied to artificial intelligence and cloud infrastructure spending, even after one of the most volatile months for the sector in years. The Philadelphia Semiconductor Index had plunged more than 20% in July, wiping out roughly $2.2 trillion in market value, but recent earnings from major cloud companies and channel checks cited by investor Gavin Baker pointed to continued strength in core demand indicators including GPU rental prices, DRAM spot prices and token usage. The rebound was visible across markets. The Direxion Daily Semiconductor Bull 3X Shares ETF jumped nearly 18% on Tuesday to around $137.61 by 11:20 a.m. Eastern Time from Monday's close of $116.71, while the Philadelphia Semiconductor Index rose about 4% and major U.S. benchmarks including the S&P 500 and Dow Jones Industrial Average hit intraday records. Micron Technology rose 7.5%, Intel gained 9.4%, Advanced Micro Devices advanced 7.1% and Nvidia traded about 1.7% higher. Baker, the former manager of Fidelity's flagship technology fund, said the July selloff reflected a chain of market fears rather than a collapse in measurable fundamentals. He said concerns over Meta leasing some compute capacity, the rise of open-source models, and worries tied to DUV lithography equipment were interpreted as signs of oversupply or weaker AI infrastructure demand, but argued those conclusions were not supported by data. In his view, open-source and closed-source models consume the same underlying compute to generate tokens, meaning shifts in model market share may redistribute profits within the AI stack without reducing infrastructure demand. He pointed to sharply higher Blackwell GPU rental prices as one of the clearest signs of persistent tightness. Baker said one startup that rented a cluster about seven months ago at roughly $2 per GPU per hour expects to pay nearly $4 on renewal, while an inference cloud company has said it plans to pay 100% higher Blackwell prices after current contracts expire. He also said changes in credit markets deserve attention, with real rates rising and credit spreads widening, but argued that faster growth in hyperscaler operating cash flow could significantly reduce the sector's need for external financing if compute is monetized at current pricing levels. That argument aligned with strong results from cloud providers. Microsoft reported $90 billion in second-quarter revenue, up 18% year over year, with Azure and other cloud services revenue up 43% and commercial remaining performance obligations at $678 billion. Amazon said AWS revenue rose 37% to $42.2 billion and operating profit reached $16.6 billion, while Alphabet reported Google Cloud revenue of $24.8 billion, up 82%, and cloud operating profit of $8.8 billion. Microsoft, Amazon and Alphabet also continued to raise or maintain aggressive capital spending plans, reinforcing expectations that demand for GPUs, memory, networking and data-center equipment remains strong. The July swings were amplified by leverage and forced positioning, especially in Asia. South Korea's KOSPI fell 22.19% during the month, while China's ChiNext and STAR 50 dropped 23.00% and 25.90%, respectively. Analysts said the region's heavy use of leveraged products and sharp moves in the dollar added pressure as investors cut risk. Hedge fund positioning later turned more constructive, with Goldman Sachs Prime Book data showing the week ending July 30 brought the strongest net buying of U.S. tech stocks since December 2022 and the third-largest weekly buying spree in at least five years. Baker also highlighted High Bandwidth Memory as the most important bottleneck in AI compute and argued long-term supply agreements have become strategically harder to break because losing memory allocation could damage a hyperscaler's market position in future shortage cycles. He said regulation remains a key tail risk, particularly if local governments curb data-center development over electricity, water or employment concerns. For now, however, easing inflation pressure, stronger cloud earnings and signs of stabilization in leveraged positioning have helped support the market view that AI capital spending is still growing and remains the sector's main fundamental anchor.