Beijing, April 8, 2024 -- MistralAI has suddenly risen to prominence, with this European AI startup being dubbed as the "arch-rival" to OpenAI. In less than a year since its inception, MistralAI has unveiled its "compact yet high-performance" hybrid expert model (MoE model) MT8x7B. Recently, the partnership between Microsoft and MistralAI caused quite a stir. Just a few days later, IBM announced the provision of Mistral AI models on its WatsonX platform. However, this information didn't garner much attention, as IBM is currently not in the spotlight.
IBM has become a "familiar stranger" to us. At the recent IBM Spring Strategy Summit, Chen Xudong, Chairman and CEO of IBM Greater China, remarked, "Today's IBM is both familiar and unfamiliar. We are familiar with IBM's leadership in numerous industry transformations in the IT sector, but unfamiliar with IBM selling off consumer businesses like PCs and X86, fading from the view of ordinary consumers, and focusing on enterprise-level services."
Sebastian Krause, Senior Vice President of IBM, candidly admitted in a media interview, "Today's IBM is fundamentally different from what it was a few years ago."
Indeed, missing out on public cloud and doubling down on AI ventures in "weather" and "healthcare" have left IBM fatigued from multi-front battles over the past decade, losing its competitive edge. Through reforms such as acquiring Red Hat, splitting infrastructure management services, and selling off The Weather Company, IBM is refocusing on hybrid cloud and AI, showing signs of revival. In addition to maintaining growth in 2023, IBM's stock price has risen from $120 at the beginning of 2022 to nearly $190 recently, approaching its historical peak in 2013.
As generative AI sparks a new industrial revolution, how will IBM participate in the upcoming transformation? Is partnering with MistralAI an attempt to replicate Microsoft's AI strategy? This year marks IBM's 40th year in China, and in this fertile ground for AI, a challenging and perhaps harsh self-transformation is just beginning. Focusing on enterprise AI and targeting private enterprises and the automotive industry, can this strategy lead IBM to find new growth opportunities in China and overcome its predicament of being "greener on the other side"?
"Steam engine + crankshaft": Enterprise AI in focus
With the advent of AI's "steam engine," how can one seize the opportunity in this era of great change? Wang Yuquan, Founding Partner of Silver Sea Capital, recently stated that large, general-purpose models are not the best choice for startups, considering factors such as computational cost and the ability to solve specific problems. He shared the story of Watt and his employee William Murdoch: Watt invented the steam engine, but it was Murdoch who invented the crankshaft, enabling the steam engine to power various industries such as trains and ships. "In times of great change, while the steam engine is important, inventing various supporting technologies to solve problems across industries is even more crucial," Wang Yuquan said.
Zhou Hongyi, Founder and Chairman of 360 Group, also believes that large, general-purpose models are not directly applicable in enterprise scenarios. He argues that developing vertical and enterprise-specific large models can better meet the personalized needs of enterprises, improve productivity and service quality, and promote industrial upgrading.
It's understandable why IBM chose to launch enterprise basic models based on watsonx instead of competing directly with general-purpose models from OpenAI, Google, and others. By leveraging watsonx and Red Hat OpenShift, along with a more comprehensive product portfolio, integrating end-to-end enterprise AI and hybrid cloud capabilities, IBM's choice in this era of upheaval is akin to "steam engine + crankshaft." While general-purpose models and the consumer market may seem enticing, the enterprise market is the real gold mine. Even OpenAI recently set its sights on enterprise AI. IBM's strength has always been in the enterprise market; it released the enterprise AI Watson in 2011, which has already served over 40,000 enterprise users.
With the arrival of generative AI, IBM announced the launch of the enterprise data and AI platform, watsonx, in May last year. Consisting of the AI development platform watsonx.ai, the data warehouse solution watsonx.data, and the governance module watsonx.governance, these components address concerns of enterprise users regarding AI model adoption, data governance, and regulatory compliance. Watsonx seamlessly integrates with the existing Watson products, providing enterprises with comprehensive solutions from machine learning to generative AI.
From the perspective of models, IBM offers both proprietary basic models and open-source models, as well as large models from third-party vendors. For example, MistralAI's MT8x7B hybrid expert model, Meta Llama's 13 billion and 70 billion parameter models, are aimed at enriching the model library's options and not replicating Microsoft's AI business strategy. Of course, IBM's self-developed basic model, Granite, is the heavyweight contender. With different technical architectures and parameter scales, Granite caters to various enterprise tasks, including content generation, summarization, classification, insight extraction, and retrieval-augmented generation (RAG) for common use cases. IBM has already released four models in the Granite series, along with detailed information on model training datasets, demonstrating its commitment to transparent and responsible AI.
IBM provides more diverse choices and flexibility in both models and deployment environments. The watsonx architecture, built on the Red Hat OpenShift open hybrid cloud platform, can be deployed on public clouds or on-premises.
For enterprise users, besides deploying computing environments and selecting suitable models, the biggest challenge in adopting generative AI lies in data governance. According to the "2023 Global AI Adoption Index" commissioned by IBM and conducted by Morning Consult, enterprises that have not deployed AI are concerned about data privacy and security, while those that are deploying or have deployed AI encounter the biggest problems with data complexity and project integration difficulties. Therefore, data governance is the key to unlocking the landing of enterprise AI. It should be noted that in recent years, IBM's focus on the enterprise market has leveraged its greatest advantage: understanding enterprises and their data. Different from traditional AI, generative AI poses new challenges in data governance for enterprises. Recently, both IBM and NVIDIA have taken similar actions, with their venture capital arms participating in the Series B funding of Unstructured, a startup founded in 2022 that primarily addresses the challenge of processing unstructured data in enterprises to make it usable for foundational model training. The purpose of this funding round is to expand the team and accelerate the development of the LLM data preprocessing tool.
In fact, amidst the frenzy of IT giants racing to stake claims on AI unicorns in 2023, IBM quietly conducted acquisitions of nine companies, most of which are data-related. Acquiring StepZen aims to help enterprises find the data necessary for deploying AI; purchasing Polar Security aims to tackle shadow data issues; acquiring the data lineage platform Manta Software targets enhancing WatsonX's data and AI governance capabilities, resolving data source issues; acquiring two enterprise technology platforms from Software AG aims to address data access and delivery consistency issues.
It's evident that IBM is solidifying its capabilities across the entire lifecycle of enterprise generative AI through both self-research and acquisitions, and its open and inclusive strategy is enriching its product portfolio, building end-to-end enterprise AI capabilities. Regarding this series of capabilities, IBM hopes to highlight transparency, trustworthiness, and professionalism.
Huang Yadan, General Manager of Echo Power (Suzhou) Information Technology Co., Ltd., stated, "We chose IBM's WatsonX series products based on four aspects: Firstly, on-premises deployment ensures data security; secondly, the deployment flexibility of hybrid cloud makes enterprise migration more convenient; thirdly, the integrated platform capabilities make the product easy to use; fourthly, enterprise-level technical support capabilities also reflect IBM's professionalism."
IBM's "Steam Engine + Crankshaft" model is in line with the demands of the enterprise market. The latest data shows that thousands of enterprise clients are currently experimenting with WatsonX, resulting in over 700 pilot projects, more than 500 signed contracts, and over 150 outstanding case studies. In the fourth quarter of last year, the business volume of WatsonX and other IBM generative AI solutions nearly doubled compared to the previous quarter.
Accelerating Diffusion with "Conveners" and "Co-Creators"
The greatest challenge for any technology, especially enterprise-level, has always been diffusion - the ability to spread determines the technology's viability. While many attribute NVIDIA's current success to its technological prowess, the reality is that its ability to diffuse and implement its technology through software has created a formidable ecosystem, serving as an unassailable moat.
A few years ago, when IBM positioned itself as an ecosystem convener and a trustworthy co-creator, the industry was skeptical, with some media dubbing the once-leading IT powerhouse as today's "pace car" of ecosystems. Yet, IBM has never lacked cutting-edge technology. For instance, in the forefront of quantum computing, IBM recently unveiled Condor, the industry's first quantum processor with over 1000 qubits, continuing to lead a new wave of IT technological revolution. Moreover, in the semiconductor field, IBM's analog AI chip introduced last August can simulate up to 17 million parameters on a 14nm chip, with energy efficiency 40 to 140 times that of a GPU. Since the 1960s, more than half of the global IT industry's technologies originated from IBM. Looking back, IBM's greatest challenge has always been strategic direction and business model decisions, determining priorities and diffusion strategies.
Former IBM Chairman and CEO Ginni Rometty once stated in her book "Who Says Elephants Can't Dance," "Lack of focus is the reason for mediocrity in a company." When IBM announced its focus on hybrid cloud and artificial intelligence, it finally concentrated its efforts.
In today's era of accelerating technological change, not only are the ways of technological diffusion changing, but the modes of technological innovation are also evolving. Traditional closed innovation is being replaced by open innovation, emphasizing the importance of collaborative creation. Therefore, IBM no longer emphasizes technological superiority but pragmatically proposes to be the "convener of an open ecosystem" and the "trustworthy co-creator of technology." Miao Keyan, General Manager of IBM Greater China Technology Division and IBM China, said, "An ecosystem is a strategy, not just a part of a strategy." In fact, companies that continue to succeed in the market today are those that can quickly combine technologies to meet user needs. Pragmatism and innovation are essential for becoming a great company.
The value realization of generative AI requires top-level design, business alignment, platform technology, integration of various application technologies, and innovation capabilities in various scenarios. The paradigm shift brought about by generative AI triggers reconstruction in every dimension, bringing new value spaces for vendors in each dimension. In this context, co-creation becomes crucial and indispensable for the implementation of generative AI.
In 2023, IBM launched a new partner program, IBM Partner Plus, significantly improving incentive programs and collaboration models. IBM also opened three major co-creation platforms (IBM Lu Ban Platform, IBM Garage Innovation Experience Center, IBM Innovation Experience Center) to partners, empowering them from various dimensions such as technology platforms, innovation methodologies, and customer experiences. In addition to increasing investment in partners, IBM has also established teams such as the Garage Innovation Team and technical sales team, spanning pre-sales and after-sales. These teams are not solely tied to sales performance but are intended to better deliver IBM's technology and innovation transformation methodologies to customers, with partners playing a crucial role in co-creation.
In September 2023, industrial IoT and big data solutions provider Zhoubang Technology announced its collaboration with IBM to provide AI-empowered solutions for China's manufacturing industry using IBM's intelligent automation software. In November 2023, IBM Consulting signed a cooperation agreement with Alibaba Cloud, becoming a service provider in Alibaba Cloud's consulting system, with cooperation covering planning consulting, cloud service sales, solution development, technical cooperation and exchange, operations management, and delivery services. In December 2023, Tencent Cloud reached a comprehensive cooperation with IBM, deeply integrating advantageous products and services, and jointly creating the "Tencent-IBM Hybrid Cloud and Artificial Intelligence Solution."
Whether embedding solutions from others, incorporating solutions from others into IBM, or jointly creating new solutions, open co-creation is accelerating the further diffusion of IBM's capabilities.
Chairman and CEO Arvind Krishna recently stated in an open letter to investors, "In 2023, we executed a validated strategy, optimized our product portfolio, expanded our partner ecosystem, and improved our productivity. Today, IBM has more comprehensive capabilities and higher productivity, as well as a strong product portfolio and a solid foundation to support sustainable growth. I have never been more confident in IBM's development direction than I am now." It should be noted that the key information here is the "validated strategy," which means that IBM's technology and capability diffusion and implementation are accelerating through its role as an ecosystem convener and technology co-creator. Indeed, from a global perspective, in 2023, revenue generated from strategic partnerships accounted for over 40% of IBM Consulting's total revenue, with double-digit growth in both signings and revenue. China's Market Transition: Racing Against Time
On January 10, 2024, in Baoding, Hebei Province, Great Wall Motors and IBM Consulting signed a cooperation agreement titled "Process and Digital Transformation." This event has garnered significant attention from the automotive, consulting, and IT sectors.
From the automotive perspective, it marks the first initiative by an independent automotive brand to kickstart a transformation in its management system, potentially accelerating the industry's shift from "brand and price competition" to "management-oriented competition." From the viewpoint of consulting and IT, it represents a significant contract. While specific figures have not been disclosed, the project encompasses multiple business areas within Great Wall Motors, including supply chain, marketing, strategy, finance, human resources, and quality. Moreover, it's a long-term collaboration, and judging by Great Wall Motors' corporate stature, the contract is substantial.
The greater significance of this project lies in its potential to bring about a "leapfrog" transformation for Great Wall Motors, propelling it to new heights.
Cai Yingzhen, Vice President of Process and Transformation at Great Wall Motors, responded to why the company embarked on this endeavor by stating two reasons. Firstly, the automotive industry is inevitably transitioning from brand and price competition towards competition centered on management systems. Secondly, Great Wall Motors is accelerating its globalization efforts. Given these factors, Great Wall Motors aims to drive a transformative leap in management on a global scale. The cost of self-discovery in such a management transformation would be prohibitive in terms of time. Starting the initial phase of collaboration with research and development, supply chain, and marketing systems, Cai emphasized the significance, stating, "Just considering the supply chain, if we can reduce costs by 1%, it means profits of tens of billions."
Cai also revealed an important piece of information: the reason Great Wall Motors chose IBM Consulting is because Cai had previously worked at Huawei and understood IBM Consulting's capabilities and approach to work.
The collaboration between Huawei and IBM Consulting is a classic case, often cited by Huawei's founder Ren Zhengfei. In 1998, Ren invested heavily in hiring IBM to provide management consulting to Huawei, involving comprehensive restructuring of business processes, internal management, finance, research and development, and supply chain. After the initial five-year consultation period, the contract was extended for another decade, with Ren Zhengfei noting, "The tuition fee was hefty, but the results were very clear." Taking the supply chain as an example, before December 1998, Huawei's on-time delivery rate was 30%, inventory turnover rate was 3.6 times per year, and order fulfillment cycle was 20-25 days. By 2003, within just five years, Huawei's on-time delivery rate reached 65%, inventory turnover rate increased to 5.7 times per year, and the order fulfillment cycle was shortened to 17 days.
Today, the collaboration with Great Wall Motors holds significant importance for IBM in China. Great Wall Motors bears three key labels: "automotive," "private enterprise," and "going global," signifying the shift in focus for IBM's Greater China market.
During the first quarter media and analyst briefing in 2024, Chen Xudong stated, "IBM's strategy in the Greater China region for 2024 revolves around three key pillars: tapping into key large clients, breaking into new markets, and aggressively expanding channels. Breaking into new markets entails leveraging China's domestic private enterprise market as a primary growth driver, bringing IBM's leading technology and industry expertise to private enterprises to support their digital transformation and business development, including overseas expansion. Multinational corporations and the automotive industry, with their immense growth potential, will be our primary targets as we aim to cultivate new key clients."
IBM is facing challenges in transitioning focus markets in China. Previously, IBM's main battleground in the Chinese market was in the financial industry and state-owned enterprises. However, with changes in policies, markets, and customer demands, IBM must find new areas of growth in this regional market. It should be noted that the three focus markets chosen by Chen Xudong share similar characteristics: industries experiencing rapid growth, with an urgent need for digital empowerment, where IBM's capabilities and strengths can be fully showcased.
Another example is the collaboration between IBM and Suzhou-based Source Photonics Technology Co., Ltd. (hereinafter referred to as Source Photonics). Source Photonics is a global supplier of digital photolithography production equipment and process solutions. Recently, they adopted IBM's generative AI solution, building an intelligent enterprise knowledge base in just two months, solving the pains of knowledge sharing and talent training within the company.
Source Photonics' founder and president, Zhang Lei, admitted: "In the past 3 to 4 years, the semiconductor optical equipment field in China has been very hot, but the downside of the industry's heat is intense talent competition, making it difficult to recruit suitable personnel, and salaries are also high." In order to reduce costs and increase efficiency, Source Photonics' solution is to combine "leading figures + emerging forces" to quickly form combat effectiveness. Such a "combination" requires strong support from an intelligent enterprise knowledge base. Zhang Lei's goal in building the knowledge base is also very clear: to find an international company with profound heritage for cooperation. On the one hand, Source Photonics has aspirations for internationalization, and on the other hand, the longer a company has been developed, the more it can accumulate superior management science. Therefore, they chose IBM.
It can be said that both Source Photonics and the industry to which Great Wall Motors belongs are among the fastest-growing fields in China, and their demands are representative. IBM's capabilities are highly compatible with these demands. The challenge lies in the fact that these companies' digital purchases cannot form purchasing power equivalent to that of finance and state-owned enterprises in a short period of time. It is for such reasons that IBM's business revenue in China has been declining at a double-digit rate for several consecutive years. It should be noted that IBM's team in China has been working hard. In the two years since Chen Xudong became the Chairman and General Manager of IBM Greater China, a large amount of work has been done, tilting more initiatives and resources towards these new fields, including establishing an automotive business group in Greater China, and expanding more partner channels. In the past year, IBM Greater China's partner ecosystem business achieved double-digit growth.
Recently, Chen Xudong and Miao Keyan led a team of 100 IBM product and technical experts to Suzhou, where more than 100 executives from over 80 core partners gathered, discussing how to build a "borderless team" and "empower Chinese enterprises to scale AI applications." Among them, "scaling AI applications" is the key phrase. If we consider the current AI application cases implemented as "points," then how to connect these "points" into "lines" and turn "coincidence" into more "inevitability" is a common topic for IBM and its partners.
The diffusion and landing of enterprise-level market technology follow certain rules and also require time. Can IBM globally give IBM Greater China more time? After all, IBM is growing in markets outside of China and the Asia-Pacific region. Can IBM Greater China achieve a rapid transition in focus markets, and can IBM gain more opportunities in China, testing the wisdom of the IBM team and market resonance.
This year marks the 40th year since IBM entered the Chinese market. The past forty years have been a period during which IBM has served China's IT information construction and achieved its own glorious growth. Now, the era of intelligence has begun, and we look forward to IBM continuing to write a new legend in the Chinese market. (This article was originally published in "Tech with Attitude" by Li Jiashi. Reprinted with permission.)
About IBM IBM is a leading global provider of hybrid cloud, artificial intelligence, and enterprise services, helping customers in over 175 countries and regions to gain business insights from their data, simplify business processes, reduce costs, and gain industry competitive advantages. Over 4,000 government and enterprise entities in critical infrastructure areas such as financial services, telecommunications, and healthcare rely on IBM's hybrid cloud platform and Red Hat OpenShift for rapid, efficient, and secure digital transformation. IBM's breakthrough innovations in artificial intelligence, quantum computing, industry cloud solutions, and enterprise services provide our customers with open and flexible choices. IBM's long-term commitment to enterprise integrity, transparent governance, social responsibility, inclusive culture, and service spirit is the cornerstone of IBM's business development.
