Core Viewpoint - SoftBank has posted its first annual profit in four years, with a net income of 1.15 trillion yen ($7.8 billion) for the year ended March, overcoming a previous loss of $1.5 billion, signaling a potential turnaround for the company as it invests heavily in AI [1][2][3]. Financial Performance - The fourth quarter saw a 124% year-on-year increase in quarterly profit, driven by a significant rise in Alibaba shares, which have increased over 55% this year, and profits from its telecom unit, including T-Mobile [2]. - The overall profit recovery provides relief for CEO Masayoshi Son, who has faced criticism for past losses, including a disastrous investment in WeWork and significant losses in the Vision Fund [3]. AI Investment Strategy - Masayoshi Son is heavily investing in AI, believing it will surpass human intelligence by 2035, and is making high-risk moves to achieve this vision [4]. - SoftBank has become a key backer of OpenAI, leading a $40 billion funding round and investing a total of $2.2 billion, viewing OpenAI as the closest partner to achieving artificial general intelligence [5]. - The company is also focusing on the necessary computing power for AI development, with Son serving as chairman of Stargate, a $500 billion infrastructure project in the US [6]. Broader AI Initiatives - In addition to OpenAI, SoftBank holds a majority stake in AI chip firm Arm and has established a new holding company, Robo HD, for its robotics investments [7]. - However, these investments come with risks, including legal challenges faced by OpenAI and uncertainties surrounding the Stargate project due to external factors like tariffs [8]. Market Concerns - There are concerns regarding the long-term demand for AI chips, as evidenced by Arm's decision not to provide full-year revenue guidance, leading to an 11% drop in its shares after earnings [9].

Investment Rating - The report maintains a cautious view on the technology sector, particularly on software and IT services, while highlighting structural growth opportunities in AI and semiconductors [26][31][44]. Core Insights - CIOs' IT budget growth expectations for 2025 have decreased to 5.8%, down 140 basis points from 2H24, with significant downward revisions anticipated in software and hardware spending [8][44]. - AI/ML/PA remains the top priority for CIOs, with 71% expecting AI/LLM projects to enhance IT investments in 2025, reflecting a 19 percentage point increase from 2H24 [53][70]. - The share of AI/LLM in total IT spending is projected to rise to 11.2% in 2025, up from 10% in 2H24, indicating a growing focus on AI-related investments [65][70]. Summary by Sections IT Budget Expectations - CIOs' average IT budget growth expectations fell to 5.8% for 2025, with a notable decline post-US tariff announcements, where expectations dropped from 6% to 2% [8][11][44]. - The survey indicates that 43% of CIOs foresee further downward revisions to their budgets following the tariffs, compared to 31% prior to the announcement [15][44]. AI and Technology Trends - AI/ML/PA has overtaken digital transformation as the most defensive area of IT spending, with significant increases in spending expectations [64][69]. - The report highlights that 34% of companies have initiated AI/LLM projects, with 39% expecting to have projects in production by 2025 [54][61]. Sector-Specific Insights - Software and IT services are expected to see the largest budget cuts, while semiconductors are anticipated to experience structural growth driven by AI, despite a delayed cyclical recovery [26][31][44]. - The hardware sector is expected to face reduced spending, particularly in PCs, while AI server hardware is projected to benefit from increased demand [33][37][70]. Preferred Companies - The report identifies several preferred companies within the AI and semiconductor sectors, including TSMC, MediaTek, and Beisen, which are expected to benefit from the ongoing AI trend [31][70][72]. - Caution is advised for companies with significant exposure to traditional tech and enterprise software, as macroeconomic uncertainties may hinder recovery [26][31][44].