Headline: IBM, surrounded by the aura of AI and quantum computing, failed to deliver impressive performance, the market returns to rationality from "technological frenzy".

The American old-tech giant IBM (IBM.US), which is surrounded by the cutting-edge technologies of artificial intelligence and quantum computing, recently announced disappointing earnings that have undoubtedly left investors disappointed. Benefiting from the unprecedented AI and quantum computing boom since 2024, IBM's stock price has skyrocketed. So far this year, it has risen by more than 30%, significantly outperforming the S&P 500 index and repeatedly hitting new highs. This also means that market expectations for IBM's performance are constantly being raised, and IBM needs to deliver much stronger-than-expected performance to sustain its strong stock price momentum. However, the second-quarter earnings announced on the morning of July 24th showed that IBM's core business software and consulting services performed poorly, significantly dampening market enthusiasm and causing the stock price to fall by more than 6% after the earnings announcement. IBM's post-market sell-off highlights investors' fervent bullish sentiment towards cutting-edge technologies such as artificial intelligence, quantum computing, room-temperature superconductivity, and controlled nuclear fusion. If IBM does not respond with performance growth, it will face significant downward pressure on valuations. IBM's much-anticipated software business department unexpectedly fell short of market expectations, disappointing investors who were increasingly optimistic about IBM's revenue growth prospects. Financial data shows that IBM's software department's revenue in the second quarter grew by 10% year-on-year to $7.39 billion, slightly lower than the average Wall Street analyst estimate of $7.49 billion. The company's other major business, consulting services, had already experienced a significant slowdown in growth in the previous quarter, with second-quarter revenue increasing by only 3% to $5.31 billion. While the software performance was lackluster, the infrastructure business unexpectedly surged. Investors have been optimistic about IBM's software business and the strong growth potential expected from artificial intelligence tools and quantum computing. Since the 1990s, the company's management has regarded software and consulting services as the core pillars of its revival strategy. However, it was only after CEO Arvind Krishna took office that this strategy began to show real results. Today, software accounts for over 40% of the company's annual revenue, while consulting services are second only to software. Headquartered in Armonk, New York, the century-old IBM stated in its financial report that since mid-2023, the overall bookings associated with artificial intelligence have exceeded $7.5 billion, up from $6 billion reported in the previous quarter. Of this, about 80% comes from the IBM consulting business, while the rest mainly comes from software. IBM's stock price, after closing near its all-time high of $282.01 on the New York Stock Exchange, fell by over 6% in after-hours trading. Year-to-date, IBM's stock price has risen by over 30%, outperforming many tech peers and the S&P 500 index, and even significantly outperforming the Nasdaq 100 index, which includes high-weighted tech giants such as NVIDIA Corporation, Microsoft Corporation, and Alphabet Inc. Class C. Overall, IBM's total revenue in the second quarter increased by 8% year-on-year to $17 billion, exceeding the average Wall Street analyst estimate of around $16.6 billion. This performance was largely driven by IBM's infrastructure business department, rather than the software and consulting business departments that IBM has focused on for a long time. The infrastructure business department saw a 14% increase in revenue in the second quarter to $4.14 billion, surpassing Wall Street analysts' expectations of $3.66 billion. IBM's CFO Jim Kavanaugh stated in an interview that this was IBM's strongest launch of a large high-performance mainframe product in history. The demand mainly came from large American enterprises, especially in the financial and retail sectors, with high requirements for localized security and private storage and processing of data for large customers. Excluding certain items, IBM's adjusted earnings per share for the second quarter were $2.80, higher than the average Wall Street estimate of $2.62. Kavanaugh stated that the company is targeting more aggressive cost savings to significantly improve profit margins. Tech suppliers selling products or consulting services to the U.S. government have long been concerned that spending cuts by the Trump administration would have a negative impact. Kavanaugh stated that some previously suspended federal contracts have been reinstated, and there have been no new contracts affected. IBM's management continues to maintain its expectation of at least 5% sales growth at constant exchange rates for the full year, which is in line with Wall Street's expectations. Free cash flow is expected to exceed $13.5 billion, also in line with Wall Street estimates. It is understood that IBM's new generation of large mainframe z17 has become popular worldwide, driving the growth of the infrastructure business department. With the increasing trend of localized AI deployment, the century-old American tech giant IBM released its latest version of the mainframe system in the first half of this year. IBM emphasized that a large amount of enterprise data will always be stored on customers' proprietary on-premise server systems, rather than being hosted in cloud computing models that, although efficient for reading and processing, carry the risk of data leaks. The performance of the new generation mainframe z17 is sufficient to handle AI-related workloads of large models and collaborate with localized storage of data. The tech giant further added that the recently acquired HashiCorp company will assist in deploying new security features for these large machines. The z17 is IBM's first new mainframe model since 2022, and it follows IBM's approximately three-year update cycle for mainframe systems. With high-efficiency generative AI models like DeepSeek continuing to gain popularity, a new wave of large AI models with the core tag of "low-cost + high-efficiency" is accelerating the trend of AI localization deployment. This has led to a renewed market interest in traditional mainframes (such as IBM's z-series mainframes) due to their localized data storage and powerful localized AI inference capabilities, supporting high-intensity AI workloads. Unlike the cloud computing loads and storage provided by cloud computing service providers such as AWS under Amazon.com, Inc., and cloud platforms like Azure under Microsoft Corporation and Alphabet Inc. Class C, the mainframe consists of customers' proprietary on-premise physical hardware infrastructure. For many years, many industry observers have expected almost all enterprise data to eventually migrate to public or private cloud service providers like AWS, Microsoft Corporation Azure, and Alphabet Inc. Class C. But IBM's infrastructure head, Rick Lewis, pointed out that as the AI trend sweeps across the globe, some enterprise customers are increasingly realizing that some data should be retained on their own on-premise servers, and that the practical benefits of localized deployment of large AI models are stronger in certain application scenarios than cloud deployment. "You'll find the entire industry gradually moving towards this hybrid deployment model," Lewis said. "Some enterprise customers are keen on fully controlling their data, unlocking all the value it brings, and ensuring that it never leaks and is absolutely accurate." The market will continue to focus on how IBM monetizes the two cutting-edge technologies of "artificial intelligence" and "quantum computing." "While overall revenue performance remains robust, the core software and consulting businesses are weak, and investors may still question the sustainability of IBM's software business growth in the second half of the year," wrote Amit Daryanani, an analyst from Evercore ISI, after IBM's earnings announcement. During the analyst earnings conference call after the financial report was released, Kavanaugh stated that IBM's Red Hat software business revenue will continue to grow at a mid-teens pace and contribute to overall software performance growth. The group, acquired by IBM in 2019, achieved a 16% revenue growth in the second quarter. However, compared to traditional software sales, the market is more focused on how IBM, crowned with the technology halos of artificial intelligence and quantum computing, leverages these two cutting-edge technologies to drive its software and consulting businesses into a hypergrowth trajectory. These two cutting-edge technologies are directly tied to IBM's software (Data & AI, Red Hat, hybrid cloud platform) and consulting (AI & Analytics, Quantum Services) businesses, making them the most critical variables for Wall Street analysts in assessing the company's growth prospects. AI and quantum computing are narratives with the potential for revaluation, becoming the decisive increment assumptions in analysts' models. These two cutting-edge technologies not only bring recurring revenue with high growth and profitability, but also provide large-order drive for projects, thereby amplifying the profit leverage of software and consulting businesses. IBM's deep layout in the AI developer ecosystem and cloud-based AI inference and computing platforms in recent years has been a key factor in its stock price performance significantly outperforming the S&P 500 index. Although IBM's IaaS+SaaS cloud service market share is not as high as Microsoft Corporation Azure and Amazon.com, Inc. AWS, it has shown unique competitive advantages in enterprise-level AI solutions. IBM's core AI developer platform, watsonx, provides a complete set of AI tools, supporting the entire process from AI model development, training to inference, deployment, providing a strong IBM AI application developer ecosystem to help them build, train, and ultimately deploy AI large models and cloud AI computing power support. The platform supports a variety of AI large models, including IBM's self-developed Granite series large models and open-source models from Meta and Mistral, allowing users to customize their AI development pattern with the most mainstream frameworks such as PyTorch and TensorFlow. In addition, watsonx provides AutoAI, SPSS Modeler, and deep learning services, supporting mainstream frameworks including PyTorch and TensorFlow to meet the needs of developers at different levels. In terms of quantum computing, IBM announced that its cumulative orders for quantum business have exceeded $1 billion and it will launch the "Quantum System Two" in New York by 2025, updating its roadmap for a thousand quantum bits, aiming to enter the "quantum practical" era by the end of this decade. IBM has been deeply involved in the underlying theory and quantum hardware systems of quantum computing for many years, focusing on gate-model quantum computing, quantum annealing, quantum simulation, superconducting quantum bits, quantum entanglement, and topological quantum, among others, hoping to integrate quantum bits with binary computing power infrastructure. It is understood that the Quantum System Two has been launched at the Poughkeepsie data center in New York, using a multi-Heron super interconnect architecture tailored for "quantum central supercomputing." In addition, IBM's "Quantum-as-a-Service" provides efficient access through IBM Cloud subscriptions, and IBM has signed long-term quantum computing cooperation agreements with large enterprises such as Boehringer Ingelheim and ExxonMobil, striving to create commercially valuable quantum computing application systems based on quantum computing technology.