US AI Leaders Face Revenue and Market Share Challenges
Jefferies estimates that half of the $2.3 trillion contracted backlog at Microsoft, Alphabet, Amazon, and Oracle is owned by OpenAI and Anthropic, raising concerns about future profitability due to competition from cheap Chinese and open source large language models.
Market Share and Revenue Concerns for US AI Leaders
In a recent research note focused on Asian markets, Jefferies estimates that about half of the roughly $2.3 trillion contracted backlog at Microsoft, Alphabet, Amazon and Oracle is owned by OpenAI and Anthropic. However, the emergence of cheap Chinese and open source large language models (LLMs), including small language models and weight-efficient models, is raising concerns about the future profitability of these US-based AI leaders. Jefferies questions where the hyperscalers' future revenue streams will come from, noting that the ongoing collapse in token pricing challenges the assumption that large language models will be profitable. The research firm suggests that the AI industry may be due for a significant reset in the not-so-distant future, as the cost advantages of alternative models, such as Kimi K3, could undercut the pricing strategies of US-based AI firms. This shift may force the hyperscalers to rethink their massive capital expenditures, as highlighted by the question: Are the four hyperscalers really going to spend $993 billion on capex in 2027?
Chinese AI Models and Their Impact on Token Processing
Chinese AI models and open source alternatives are rapidly reshaping the landscape of token processing, according to research from OpenRouter cited by Jefferies. In the week ended 27 September, Chinese models processed 76.12 trillion tokens, a significant jump from 4.37 trillion in late April. This surge highlights the growing influence of Chinese firms in the global AI market. Meanwhile, Moonshot AI’s open source Kimi K3 is said to be competing at the frontier, with performance estimates suggesting it reaches 95 percent of Anthropic’s Claude Fable 5. These developments could challenge the dominance of established hyperscalers and disrupt the trillion-dollar AI infrastructure sector. Jefferies also notes that hyperscalers’ reported profits may be heavily influenced by their investments in AI model builders, with “other non-operating income/loss” accounting for all of the increase in earnings over the four quarters to 2Q26. This raises important questions about the financial dependencies and revenue streams of major cloud providers.
Infrastructure Funding and Revenue Shortfalls in the AI Industry
The AI industry faces a significant infrastructure funding challenge, as highlighted by Bain & Company's 2026 Global Technology Report. To meet anticipated demand, the industry must generate $6 trillion in annual revenue by 2031, yet existing AI applications are projected to contribute only $1.8 trillion, leaving a $4.2 trillion shortfall. Even with the potential addition of $1.5 trillion from emerging applications like AI-powered autonomous vehicles and simulations, a $2.7 trillion gap remains. This gap may be filled by yet-to-be-invented products and uses. Meanwhile, the largest hyperscalers are expected to spend between $600 and $800 billion this year on datacenter construction, increasingly relying on debt rather than cash. The rise of cheap Chinese and open source large language models (LLMs) could further disrupt the industry, potentially triggering a major reset in AI infrastructure investment and revenue models.
The Role of Open Source in Linux's Global Influence
The role of open source in Linux's global influence is a testament to the power of collaboration and shared innovation. As highlighted in the 35 Years of Linux report, the Linux kernel has not only shaped the technological landscape but also inspired a culture of open contribution across software, hardware, and data domains. Insights from Linus Torvalds and other contributors reveal how the kernel has become a foundational element for industries relying on open source for operational and differentiated technologies. The report emphasizes that the collaborative ethos of the Linux kernel has extended beyond software, influencing open hardware and open data initiatives. According to the report, the impact of open source software on the global economy is significant: if open source were eliminated, global GDP for the average nation would be over 2% lower in the long run. This underscores the critical role of open source in driving innovation and sustaining economic growth.
Linux's Adaptation to AI Workloads and Community Growth
Linux's adaptation to AI workloads and its continued community growth highlight its resilience and relevance in the evolving technology landscape. As noted in the report, Linux powers an extensive range of devices, from 4 billion Android devices to 200 million-plus servers and all of the top 500 supercomputers. This ubiquity underscores its critical role in modern computing infrastructure. The Linux community remains vibrant, with strong contributor diversity and retention, as evidenced by over 100 contributors making up 50% of contributions and over 50% of contributors remaining actively engaged over quarters. Additionally, the mainlining of built-in real-time support (PREEMPT_RT) in kernel 6.12 marks a significant milestone, achieved through a twenty-year community effort. This advancement opens new possibilities for industrial robotics, automotive systems, and latency-sensitive edge environments, further solidifying Linux's position as a foundational element in AI and beyond.
Sources and reporting notes
Based on published source material. No independent hands-on testing is claimed.
