Five Front Lines, Not One: Otaviano Canuto on the U.S.-China Technology Race
By Salvator Mundi Times Salvator Mundi Café No. 3, Sunday 6 September 2026 Speaker: Dr. Otaviano Canuto, Chair, SMG Global Council on Global Economic Futures
At the third Salvator Mundi Café, the former World Bank Vice President argued that reading the U.S.-China rivalry through semiconductors alone misses most of the contest.
Dr. Canuto presented forty-nine slides drawing on the IMF, the World Bank, the International Energy Agency, Goldman Sachs Global Investment Research, Ember, Bloomberg, the International Federation of Robotics, the Stanford AI Index and the Financial Times.
Download the presentation 👇👇👇👇👇👇👇
The talk draws on two of Dr. Canuto's recent publications: "The New South as a Frontline of the U.S.-China Technological Rivalry", Policy Brief PB-07/26, Policy Center for the New South, February 2026; and "The Multiple Frontlines of the US-China Technological Rivalry", Capital Finance International (CFI.co).
__________________
Dr. Otaviano Canuto is a former Vice President and Executive Director of the World Bank, a former Executive Director at the International Monetary Fund and a former Vice President of the Inter-American Development Bank. He is a Senior Fellow at the Policy Center for the New South, an Affiliate Professor at Mohammed VI Polytechnic University and a Non-Resident Senior Fellow at Brookings.
He opened with a caveat that framed everything after it. Since retiring from the World Bank and the International Monetary Fund in 2019, he said, he holds no official position in any institution or government, which means he can say exactly what he thinks. Over the next two hours he did.
He joined the Café from Rabat, coffee cup in hand, and apologised in advance for bringing slides. The apology was unnecessary. The charts carried an argument that is difficult to make in words alone: that the technological rivalry between the United States and China runs along at least five separate front lines, and that whoever reads it as a single contest over advanced chips will misjudge where it is actually being decided.
The ladder, and the homework
Canuto began further back than the rivalry itself, with the question of how countries get rich.
His 2021 book, Climbing a High Ladder: Development in the Global Economy, was written against the argument that advanced economies pull the ladder up behind them. His finding was close to the opposite. The globalisation that began around 1990 laid the ladder down. Three shifts arrived at once: the fall of the Berlin Wall and the integration of Eastern Europe and the former Soviet Union into the market economy; Deng Xiaoping's reforms, which brought roughly a billion new workers into that same economy; and a leap in information technology and container shipping that allowed production chains to be broken into segments and located wherever each segment was cheapest to run.
The results, he argued, are in the data and are not seriously contested. Trade as a share of global output climbed steeply from the 1990s. Per capita income grew faster in emerging and developing economies than in advanced ones, and that remains true when China is excluded. Poverty rates fell across every World Bank region, with sub-Saharan Africa the laggard rather than the exception. Global inequality, measured across the whole world population, declined.
"Globalisation was good for reducing poverty and reducing inequality. Period. This is what the data show."
But the ladder being there is not the same as climbing it. That, he said, is the homework. Countries that combined the opportunity with sustained domestic investment in their own capabilities moved up the value chains. Countries that did not stalled in a middle income trap. South Korea did the homework. Brazil, his own country, did not.
He is careful about one thing the same charts show. Inequality within some advanced economies rose over the same period. That is the material behind the politics of the last decade, and he does not dismiss it. What he disputes is the diagnosis. The losses suffered by lower income workers in the United States, he argued, reflect technological change and a failure to retrain far more than they reflect China, and the promise that history can be rewound is a lie whether it is sold in Washington, in London or in Paris. The steel plants of Illinois are not coming back.
2019: climb the rest alone
Then the world turned restrictive. Trade restrictions on goods, services and investment rose sharply after the global financial crisis and again after 2020.
The moment Canuto marks as the opening salvo is 2019, when the first Trump administration imposed restrictions on market access and technology for Huawei, ZTE and their subsidiaries. He wrote at the time that this looked like the start of something durable.
The message, as he reads it, was this:
"China, you have made good use of the opportunities to climb the ladder of technology and income. But now you will have to climb the rest of the ladder alone."
China heard it. What followed was a sustained investment programme aimed at cutting dependence on external technological frontiers. Export controls on advanced chips and manufacturing tools slowed China at the frontier, he said, but did not stop it, and in several areas they accelerated the development of domestic alternatives.
The five front lines
Semiconductors and artificial intelligence. There is no single semiconductor supply chain, but a network of interconnected chains spanning the globe, with bottlenecks at advanced lithography, etching equipment and leading-edge manufacturing capacity. The United States and its allies hold a clear edge there, and replicating advanced lithography remains a genuine technical barrier for China. But leadership in artificial intelligence will not be settled by hardware alone. It depends on dissemination and adoption. China's bet on open and low-cost models has produced rapid real-world deployment even behind the hardware frontier, and by the start of 2026 the best Chinese open-source models had overtaken their American counterparts on published intelligence rankings. Chinese patent filings in artificial intelligence, semiconductors and quantum computing have grown steeply. The persistent lag sits in the high-end areas targeted by export controls, and in semiconductors China's success remains concentrated in lower value-added assembly and packaging.
Applications. This is where Canuto thinks the United States is quietly losing ground. The question is shifting from who builds the most sophisticated model to who embeds it most effectively in real systems. China installs roughly 470 industrial robots per ten thousand employees, ahead of Germany, Japan and the United States and behind only South Korea and Singapore. Electric vehicles, batteries, drones, robotics, smartphones and artificial intelligence form overlapping industrial ecosystems in which progress in one domain reinforces progress in the others. Survey evidence puts China far above the global average on enthusiasm for artificial intelligence and well below it on anxiety.
Clean energy. Over the past twelve years, Chinese investment in clean energy has been roughly double the combined investment of Europe and the United States. That has produced dominance across every stage of the solar supply chain, from polysilicon through wafers, cells and modules, a wind capacity curve that has pulled away from every other country, and nuclear reactor construction costs that have fallen while American costs rose. Green technology exports in batteries, solar and vehicles have climbed sharply, increasingly into developing countries rather than into Europe and the United States. Canuto notes a reversal that has not yet been absorbed politically: where clean energy once carried a cost penalty against fossil fuels, the technology has moved fast enough that, given minimum infrastructure, it is now the cheaper option, including in the United States, where many private investors have declined to follow the drilling agenda for straightforwardly commercial reasons.
Critical minerals and rare earths. The front line that almost everyone underestimated, and that China mapped out more than a decade ago. Reserves are scattered. Brazil holds the second largest rare earth reserves in the world. But value in this chain is not captured in the ground, it is captured in refining, which is complex and technology intensive. China holds up to ninety per cent of refining capacity in some segments, and around ninety-two per cent in rare earths refining and ninety-eight per cent in magnet production. Canuto is blunt about what this bought Beijing. When Washington restricted semiconductors, China restricted processed critical minerals, and a truce followed. It remains the only instance in that period of a country successfully pushing back. He is equally blunt about the enthusiasm now sweeping Africa and Latin America over mineral endowments: extraction is the low-value end, and the environmental and social record of cobalt production in the Democratic Republic of Congo is, in his words, nothing beautiful.
Energy, power and infrastructure. Power demand from artificial intelligence data centres is projected to quadruple within ten years. The United States holds around twenty-four per cent of global data centre capacity. But American power plant installation has stagnated while consumption has risen sharply since the release of the first mass-market chatbots, whereas China has expanded electrification capacity so far ahead of peak demand that the margin is unusually wide. Availability of power, Canuto argued, can no longer be taken for granted on either side, and it is now a constraint on the technology race itself.
The new normal
Behind all five front lines sits a change in how governments think.
For decades after the Washington Consensus, Canuto said, the institutions he worked in were reflexively sceptical of industrial policy. That era is over. China did not win through tariffs. Globalisation 2.0 was built on countries lowering trade barriers, and China lowered its own. What China used, at enormous scale, was subsidies and coordinated industrial policy. Washington has now followed, first through the semiconductor legislation and the Inflation Reduction Act under President Biden, which he describes as industrial policy under another name, and then through the tariff regime that replaced them.
The styles differ sharply. The American approach is hybrid and fragmented, mixing wartime-style mobilisation with market incentives. The Chinese approach is holistic and highly coordinated, and in scale, as he put it, humongous.
The deeper shift is one he heard stated directly. In 2023, at Brookings, the White House National Security Advisor addressed a room of economists and told them they had it wrong: they see efficiency and economic gains, but national security matters as much, if not more. That, Canuto said, is now the operating assumption, and it has displaced efficiency as the organising principle of policy in both capitals.
The consequence for everyone else is a choice they did not ask for. Standards diverge, and pressure will grow on other countries, developing countries above all, to select between Chinese and American platforms.
From the floor
The discussion that followed ran almost as long as the talk.
Mario Diel, President of Salvator Mundi Global, reflected that technology has displaced military size, territory and money as the axis of global power, that Europe has slept for at least fifty years, and that Russia barely appears in the picture at all. Canuto agreed on Europe. It has a currency but no fiscal union and no real integration of its banking and financial systems, which is why nothing resembling the American mechanism for pulling in global savings to fund artificial intelligence could happen there. Western Europe, he said, did not do its homework either.
Whatch the conversation
The full tvideo is on Salvator Mundi Global TV. It is worth your time, and if it is, please like the video and subscribe. The channel is new.
Dr. Canuto opened by noting that since retiring from the World Bank and the IMF he holds no position in any institution or government, "so I can say whatever I want with no concern." For two hours on a Sunday afternoon, he did. The Salvator Mundi Café meets every Sunday at 15:00 CET.
About the Author
Shlomi Selim "Sally" Benbasat
Editor in Chief, Salvator Mundi Times
Serial entrepreneur and investor in B2B e-commerce, carbon markets and emissions reduction, chemicals and raw materials trading, EPC, technology, telecommunications, digital media and real estate, with ventures built and exited across Europe, LATAM, the Middle East and Asia.
Comments