McKinsey: Companies Adopting AI Unprofitable, Lee Jae-myung's '400 Trillion Won National Growth'
McKinsey & Company released a report on June 13th stating that large US companies adopting AI are "unprofitable." On August 13th, the Lee Jae-myung administration's National Policy Planning Committee announced a "400 Trillion Won National Growth Market" for AI powerhouses.
The New York Times reported on the 13th that despite the near-total lack of profitability in AI, large tech companies are fiercely competing to invest billions of dollars to keep up with the technological competition.
McKinsey's report, "Securing the Agent AI Advantage," stated on the 13th that "almost eight out of ten companies report using generative AI, but just as many report that it has had little impact on their profits," and suggested that "consider this the 'generative AI paradox.'"
The report said, “At the heart of this paradox is the imbalance between ‘horizontal’ (enterprise-wide) co-pilots and chatbots, which have scaled rapidly but deliver widespread and difficult-to-measure benefits, and more innovative ‘vertical’ (function-specific) use cases, of which about 90% remain in pilot mode.” “But the bigger challenge is not technological. It will be human to build trust, drive adoption, manage agent autonomy, and establish the right governance to prevent uncontrolled proliferation.”
The report, while offering some positivity, stated, "AI agents offer a way out of the generative AI paradox. By combining autonomy, planning, memory, and integration, agents have the potential to automate complex business processes, transforming generative AI from a reactive tool to a proactive, goal-driven virtual collaborator. This shift enables more than just efficiency; agents can increase operational agility and create new revenue opportunities."
However, it also highlighted the limitations of the technology, stating, "However, unlocking the full potential of agent AI requires more than simply plugging agents into existing workflows. This requires reimagining these workflows from the ground up, with agents at their core."
The report stated, "Managing the rapidly evolving organizational AI landscape and enabling teams to blend custom and off-the-shelf agents while managing increasing technical debt and new risk levels requires a new AI architectural paradigm—the Agent AI Mesh.
However, the bigger challenge is not technical." It added, "To scale impact in the agent era, organizations must reconfigure their approach to AI innovation from a distributed initiative to a strategic program."
The report continued, "From use cases to business processes; from siloed AI teams to cross-functional innovation teams; In "From Experimentation to Industrialized and Scalable Delivery," it stated, "Building trust, fostering adoption, managing agent autonomy, and establishing proper governance to prevent uncontrolled proliferation will be humane."
The report stated that "organizations must build a foundation for effective operation in the agent era," and that "they must enhance workforce skills, align technology infrastructure, accelerate data productization, and deploy agent-specific governance mechanisms." It concluded by suggesting that "the moment has come when the generative AI experimentation ground, a turning point that only a CEO can achieve, is complete."
The New York Times reported that “nearly 40 years ago, during the height of the personal computer boom, a phenomenon known as the ‘productivity paradox’ has resurfaced with the AI boom,” referring to the fact that despite companies’ massive investments in new technologies, there was little evidence of corresponding improvements in worker efficiency.
The New York Times reported that “AI has raised expectations that it will revolutionize everything, but the rewards for companies outside the tech sector are lagging behind, due to problems including the annoying tendency of chatbots to make things up.” “This means companies will have to continue investing billions of dollars to keep up, but it could take years for the technology to pay off across the economy as companies gradually figure out what works best,” and that “there will be no short-term returns.”
Technology research firm IDC, regarding the “generation of AI paradox,” told the New York Times that “corporate investment in generative AI is expected to increase 94% this year to $61.9 billion.”
On the other hand, data and analytics The firm S&P Global reported in a survey of more than 1,000 technology and business managers that “the percentage of companies abandoning most AI pilot projects has surged from 17% the previous year to 42% by the end of 2024.”
Alexander Johnston, a senior analyst at S&P Global, told the Times that projects failed not only because of technical hurdles but also because of “human factors,” such as employee and customer resistance or a lack of skills.
Gartner, a research and advisory firm that charts technological “hype cycles,” predicts that AI is sliding into what it calls the “trough of disillusionment,” with John-David Lovelock, a senior forecaster at Gartner, predicting a potential downturn, saying, “We expect the technology to bottom out next year before it becomes a proven productivity tool.”
AI follows the pattern of past technologies like personal computers and the internet: an initial boom, followed by the arduous effort to master the technology, and then a shift in industry and work.
The winners in AI so far have been the suppliers of AI technology and advice. This is limited.
This includes Microsoft, Amazon, and Google, which provide AI software, while Nvidia is the runaway leader in AI chips.
Executives at these companies have boasted about how AI is reshaping their workforces, eliminating the need for some entry-level coding tasks and making others more efficient. South Korea's Lee Jae-myung administration has shifted its government to a stock-boosting policy with a 100 trillion won fund to "build an AI powerhouse."
The AI powerhouse strategy has included shifting the administration to AI-controlled management, basing its politics on the hypothesis that AI will eventually replace entire human employees. Companies investing heavily in AI have established a system that is widely accepted by the mainstream and drives political resonance.
"Raw technological horsepower is wonderful, but it won't determine how quickly AI transforms the economy," Professor Andrew McAfee, chief research scientist and co-director of the Initiative on the Digital Economy at MIT, told the New York Times.
The New York Times reported, "Two years ago, the largest bank in the United States, JPMorgan Chase blocked access to ChatGPT on its computers, citing potential security risks, and stated that only a few hundred data scientists and engineers were currently able to experiment with the AI.
Currently, approximately 200,000 bank employees have access to a general-purpose AI assistant (essentially a business chatbot) on their work computers for tasks such as data retrieval, business question answering, and report writing.
This assistant, tailored to JPMorgan's specific needs, leverages ChatGPT and other AI tools while ensuring data security for confidential bank and client information.
The New York Times reported that "about half of workers use it regularly, and it reports spending up to four hours less a week on basic office tasks." "Wealth advisors at the bank are hiring more specialized AI assistants that use bank, market, and client data to provide investment research and advice to wealthy clients.
The bank says this has boosted revenue by helping them retrieve information and make investment recommendations nearly twice as fast as before," according to Lori, JPMorgan's global chief information officer. Beer oversees 60,000 technical staff worldwide.
The New York Times asked, "Did she kill the AI project? “It’s probably a few hundred in total,” she said, adding that “many of the shelved prototypes have been developed with concepts and code to move on to other ongoing projects.”
“It’s not surprising that early AI efforts are lacking,” Professor McAfee, an MIT research scientist and founder of the AI consulting firm Workhelix, told the Times. “Innovation is a process that fails quite regularly.”
One company where AI flaws have surfaced is USAA, which provides insurance and banking services to military personnel and their families.
The Times reported that “after several pilot projects (some of which have since ended), the company has implemented AI assistants to help its 16,000 customer service agents provide the right answers to specific questions. USAA is tracking its AI investments, but there’s no financial calculation yet on the return (if any) on the call center software.”
The company said “employee feedback has been overwhelmingly positive,” but added that despite the presence of software apps that can now answer customer questions online, its staffed call centers still receive an average of 200,000 calls a day, a significant loss for human labor. It was described as an "operating system."
"Those moments are crucial," said Ramnik Bajaj, USAA's chief data analytics and AI officer, explaining why they maintain human-powered call centers.
This is similar to an AI app developed a little over a year ago for field workers by Johnson Controls, a major supplier of construction equipment, software, and services.
The company fed machine operation and service manuals into an AI program trained to generate problem summaries, suggest repairs, and relay everything to the technician's tablet computer.
In tests, the app reduced repair calls that took more than an hour by 10 to 15 minutes, a useful efficiency gain but not a workplace innovation in itself.
Fewer than 2,000 of the company's 25,000 field service workers have access to the AI assistant, but the company plans to expand.
"It's still early days, but we think over time everyone will use it," said Vijay Sankaran, Johnson Controls' chief digital and information officer. "The long-term vision is that it will be used by everyone."
He said, “Companies are using AI to improve multiple systems, including sales, procurement, manufacturing, customer service, and finance, and this is a game changer,” but he told the NYT, “I predict that the change will take at least five years.”
The McKinsey report stated, “The key issue is the imbalance between ‘horizontal’ and ‘vertical’ use cases,” and that “while the former, such as employee co-pilots and chatbots, have been widely deployed but offer diffuse benefits, vertical or function-specific use cases with greater impact rarely move beyond the pilot stage due to technological, organizational, data, and cultural barriers.” The report concluded, “Unless companies address these barriers, the innovative potential of generative AI will remain largely untapped.”
On the 13th, the Presidential Committee for State Affairs Planning under the Lee Jae-myung administration announced the launch of the “AI 100 trillion won fund” in the “National Growth Fund,” stating, “The key is policies to increase public access to AI,” and “Infrastructure investment in AI data centers, energy highways, power generation, and water supply through project financing (PF).”
The AI highway and energy highway are large-scale investment construction funds that are being used for "future renewable energy that has no commercial viability."
In particular, the three major AI powerhouses announced a 3% economic growth system. They stated that they would "install AI highways, secure over 50,000 GPUs, secure high-quality data, and cultivate core technologies and talent."
They also proposed "expanding AI-related education, utilizing AI to prevent disasters such as floods and wildfires, and innovating public services through AI" as top national priorities.
The AI highway and energy highway are large-scale investment construction funds that are being used for "future renewable energy that has no commercial viability."
Regarding the growth strategy, the State Affairs Commission stated, "We will create 'K-culture' by expanding the arts and culture industry market to 300 trillion won and increasing exports of culture-related industries to 50 trillion won."
To this end, the government announced plans to invest 10 trillion won in policy financing and tax incentives, expand performance-type arenas, and invest 10 trillion won in the national budget to create a 300 trillion won market. This transformed "national growth" into a "national contribution project to create a 300 trillion won market," with the 10 trillion won loaned to the public, with the principal and interest on government bonds becoming a national burden.
The State Affairs Commission also proposed "package exports" for the defense industry. This shift to a "stock price boosting economic system by increasing national debt burden" through "fund raising by allocating funds to the national growth fund" and a 1.5 trillion won fiscal investment, coupled with a "crisis of system collapse in the event of a stock price decline," has led to a shift to a "stock price boosting economic system by increasing national debt burden."
See <Trump's debt increased by $1 trillion in 5 months, $37 trillion; Lee Jae-myung borrows $100 trillion from AI fund PF, August 13, 2025>