China’s advances could boost AI’s impact on global GDP

DeepSeek and a handful of other Chinese companies have reportedly developed sophisticated generative artificial intelligence (AI) models at a lower cost than existing offerings. The development may spur faster adoption of AI and help the technology have a larger impact on global economic growth, according to Goldman Sachs Research.

The breakthrough challenges the view that prohibitive investment costs are a barrier to entry for the largest, most powerful AI models. While it’s still being determined how Chinese researchers developed their AI technology and the full cost of it, a lower cost structure could help AI to develop and spread around the world more quickly, Joseph Briggs, co-leader of the Global Economics team, writes in the team’s report.

The “breakthrough could raise macroeconomic upside over the medium-term if its cost reductions help increase competition around the development of platforms and applications,” Briggs writes. “Limited adoption is still the main bottleneck to unlocking AI-related productivity gains, and adoption would benefit from competition-induced acceleration in the buildout of AI platforms and applications.”

“That said, the near-term adoption impact is probably limited since cost itself is not currently the main barrier to adoption,” he adds.

So far, the biggest barriers to near-term adoption reported by companies are a lack of knowledge about AI capabilities and privacy concerns, according to Census Bureau data. Just 6% of US companies report using AI for regular production, up from 4% in late 2023. 

How much will AI increase GDP?

Previously, the economics team’s baseline estimate was that widespread adoption of generative AI could raise US labor productivity by 15% over roughly 10 years, mainly by automating work tasks. That would unlock about $4.5 trillion of annual US GDP (in today’s dollars). The economic benefits are expected to accrue to hardware and infrastructure providers in the early phases, extend to the developers of platforms and applications in a later phase, and ultimately show up productivity and efficiency gains across industry more broadly.

The team has also forecast an AI investment cycle that peaks in the US at 2% of GDP before fading as the compute costs of training AI models and running AI queries fades. AI software investment was anticipated to increase steadily over time as end-user adoption increases.

While the Chinese developments in AI raise questions about investment and technology leadership by a handful of incumbents, the team’s perspective on the macroeconomic impact of AI is unchanged: The main macroeconomic boost is expected to come from increased productivity as companies incorporate AI-driven automation into their businesses.

“The emergence of a credible competitor to US-based AI leaders could provide an uplift to global adoption and productivity,” Briggs writes. The emergence of a solid non-US competitor could prompt some governments to raise the importance of developing domestic AI capabilities. Increased global competition may spur cross-border cooperation or lower regulatory barriers to encourage AI development and adoption.

At the same time, the potential automation and productivity gains from generative AI are generally similar for economies across the globe, Briggs writes. “While we still expect that the US will adopt AI more quickly than other countries, given its leadership in AI model development, the emergence of non-US based platforms and applications could accelerate the adoption timeline elsewhere.”

How AI will boost productivity

The team’s forecasts assume that US adoption of generative AI technology will start to show up in productivity figures in 2027, with peak impact expected in the early 2030s. Other developed markets and key emerging market countries lag behind the US timeline by a few years in these projections. “The recent DeepSeek reports suggest adoption could happen sooner,” Briggs writes.

Goldman Sachs Research still expects AI adoption will climb in the medium-term, and Briggs points out that the types of work tasks automatable by generative AI would result in several thousands of dollars of cost savings per worker per year. “Given that potential cost savings from generative AI are large and the marginal cost of deployment once applications are developed will likely be very small, we see adoption of generative AI as more of question of ‘when’ rather than ‘if,’” he writes.

There are valid questions about how lower-cost AI models could impact stakeholders in the AI ecosystem, Briggs writes. How any profits are distributed will depend on things like market concentration, intellectual property rights, scalability, and ultimately the competitive landscape. While it’s still too early to know the new models’ effect, if expensive hardware and computing power are going to be less essential to realizing the economic benefits, companies that build the physical infrastructure may garner less of the overall gains.

Briggs points out that questions about the distribution of growth are less relevant, however, to the overall macroeconomic story. The outlook doesn’t depend on who specifically benefits, and the overall implications of the breakthrough in China are most likely net positive.  

Will China’s AI developments reduce investment?

One of the big questions is whether more efficient AI models could result in lower capital expenditures for AI — equity analysts forecast, based on consensus estimates, that AI-related capex will rise to $325 billion by the fourth quarter of 2025 — and whether that could filter through to slower GDP growth.

Should cheaper models result in lower capex on AI, two things are likely to limit the economic impact in this scenario, according to Goldman Sachs Research. While companies have been reporting that they are boosting their AI-related investments, this has had limited impact on official GDP data so far. And companies are unlikely to significantly adjust their capital allocations simply because of the recent news in China, according to Goldman Sachs Research’s equity analysts.

In the meantime, while there’s some risk that lower-cost AI models could result in less building of AI infrastructure than expected, there is also the possibility that these advances will push AI incumbents to invest more to maintain their lead. Fundamentally, if these new developments spur competition and lower costs, they could catalyze a faster buildout of AI platforms and applications. 

 

This article is being provided for educational purposes only. The information contained in this article does not constitute a recommendation from any Goldman Sachs entity to the recipient, and Goldman Sachs is not providing any financial, economic, legal, investment, accounting, or tax advice through this article or to its recipient. Neither Goldman Sachs nor any of its affiliates makes any representation or warranty, express or implied, as to the accuracy or completeness of the statements or any information contained in this article and any liability therefore (including in respect of direct, indirect, or consequential loss or damage) is expressly disclaimed.