Samer Choucair: AI’s Competitive Advantage Is Shifting From the Model to Energy, Chips and Data
Investment leader Samer Choucair said U.S. allegations involving six Chinese artificial intelligence companies, including DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun and Z.AI, are forcing markets to reassess where competitive advantage truly resides in the AI industry and prompting institutional investors to reconsider both the sources of value and the risks attached to AI investments as technological competition between the United States and China intensifies.
Choucair said U.S. authorities have accused the companies of using distillation techniques to extract capabilities from advanced models developed by American companies including Anthropic, OpenAI, Google and xAI.
He noted that model distillation is itself a legitimate technique for developing smaller, more efficient and less expensive models. The significance of the allegations, however, is that they move the debate beyond technical methodology and into the increasingly sensitive areas of intellectual property, national security, export controls and cross-border investment.
“Institutional capital does not buy performance convergence; it buys ownership of knowledge,” Samer Choucair said. “If access to advanced model capabilities can be shortened through distillation techniques, markets will have to reassess the difference between the cost of developing intelligence and the cost of accessing its capabilities.”
For investors, that distinction could become increasingly important as advanced AI capabilities become more widely distributed and the performance gap between competing models narrows.
The Competitive Moat Is Moving
Samer Choucair said competition between the United States and China is no longer primarily about which model performs best on benchmarks.
The strategic contest increasingly encompasses advanced semiconductors, access to training capital, electricity supply, data centers, supply chains, proprietary data, and the ability to transform an advanced model into a scalable commercial product.
That shift has important consequences for investment analysis.
If model capabilities become easier to replicate, approximate or access at lower cost, the economic moat surrounding the model itself could weaken. Value may increasingly migrate toward the scarce resources required to develop, deploy and operate AI systems at scale.
“The investable competitive advantage is no longer simply having an advanced model,” Choucair said. “The advantage is owning an innovation cycle that can be protected, data that can be controlled, computing capacity that can be secured, and APIs over which the company can maintain commercial control.”
In that environment, access to electricity and computing infrastructure can become as strategically important as the underlying algorithm.
Investors Will Separate the AI Stack
Choucair expects investors to increasingly distinguish between three major layers of the AI economy: the model itself, the application layer built on top of it, and the computing infrastructure required to operate it.
For Samer Choucair, that separation will become increasingly important in institutional portfolios during 2026 and 2027, particularly as investors confront greater uncertainty around sanctions, restrictions on advanced semiconductor exports, licensing conditions and cross-border technology access.
The risk profile of a company building a foundation model is fundamentally different from that of a business developing enterprise applications, while both differ from the economics of companies owning data centers, power infrastructure, networking equipment or advanced computing capacity.
Institutional investors will therefore need to determine not only which AI companies are growing fastest, but which layer of the technology stack gives them the strongest combination of pricing power, capital efficiency, regulatory resilience and defensible intellectual property.
Choucair also expects due diligence in venture capital and private equity transactions to become significantly more demanding.
Evaluating an AI company will no longer be limited to assessing model quality or benchmark performance. Investors will increasingly examine the provenance of training data, ownership of intellectual property, model-update mechanisms, licensing conditions, computing dependencies and the company’s ability to distribute its products across different jurisdictions.
Saudi Arabia’s Infrastructure Opportunity
Turning to Saudi Arabia and the Gulf, Samer Choucair said the growing fragmentation between the American and Chinese technology ecosystems creates an opportunity for the Kingdom to establish itself as a regional computing hub rather than entering a direct contest over which country or company ultimately produces the dominant AI model.
Saudi Arabia possesses several assets that could become increasingly valuable as AI infrastructure expands: energy, land, sovereign capital, growing data-center capacity and international technology partnerships.
Together, Choucair said, those resources could allow the Kingdom to position itself as a trusted operational platform within the global AI ecosystem.
“Saudi Arabia does not need to enter a direct race between the American and Chinese models,” Choucair said. “The larger opportunity is to build the intermediate layer that provides energy, computing, data centers, data, financing and governance, making the Kingdom a trusted operating hub for investors and companies.”
That strategy would shift the investment thesis away from attempting to predict a single technological winner and toward owning the infrastructure required by multiple potential winners.
Beyond the AI Model
Choucair said investment opportunities consequently extend far beyond foundation models.
Data centers, electricity generation and transmission, cooling infrastructure, networks, semiconductors, specialized technical training and cybersecurity could all benefit from the continued expansion of AI computing demand.
The opportunity also extends to localized AI applications across healthcare, financial services, energy, government services and other sectors where access to domestic computing capacity and data could become strategically important.
For Saudi Arabia, this could create a broader investment ecosystem in which AI demand supports physical infrastructure while infrastructure investment, in turn, strengthens the Kingdom’s ability to attract technology companies and advanced digital industries.
Choucair cautioned, however, against constructing portfolios around a single geopolitical assumption.
Investors should neither assume that China will inevitably dominate by reducing AI development costs nor that the United States will succeed in completely restricting Chinese access to advanced technologies.
The more probable outcome, he argued, is sustained competition accompanied by an increasingly fragmented global technology architecture.
That environment would place a premium on companies and jurisdictions capable of operating across uncertainty, securing critical infrastructure and adapting to changing technology restrictions.
The New Economics of AI
For Samer Choucair, the investment implications extend beyond the current dispute over distillation.
The deeper question is whether AI models themselves will remain the industry's most defensible source of economic value as capabilities become more widely available.
If model performance increasingly converges, scarcity may shift elsewhere.
Advanced chips remain difficult to manufacture. Large-scale computing requires enormous capital expenditure. AI data centers require reliable electricity and sophisticated cooling. Proprietary datasets can be difficult to replicate. Regulatory permission to operate across major markets can itself become a competitive asset.
These constraints could determine which companies ultimately capture sustainable returns from artificial intelligence.
The investment thesis therefore moves from asking simply, “Who has the best model?” to asking, “Who controls the resources required to build, operate, distribute and monetize AI at scale?”
“2026 will not be decided by a single model,” Samer Choucair concluded. “It will be decided by who controls the conditions under which artificial intelligence is produced: energy, chips, data and legal trust. For investors in Saudi Arabia and the region, capital allocation is no longer a late-stage technology decision; it is an early sovereign decision.”