China AI capital markets open $28 trillion funding front

China is steering its $28 trillion capital markets toward AI and chip companies as CXMT’s rapid listing tests Beijing’s new financing model.

Mei Lin ·

China AI capital markets open $28 trillion funding front

China AI capital markets are being steered toward a $28 trillion funding pool as Beijing tries to finance its AI rivalry with the US.

CXMT tests the new channel

CXMT Corp. became the clearest test case when its Shanghai debut turned the memory-chip maker into mainland China’s most valuable listed company. The shares rose more than 500% within hours of trading and passed Industrial and Commercial Bank of China Ltd. by market value.

The company raised about $9.8 billion after moving from filing to trading in less than eight months, a pace described in the source as unusually fast for a process that can take years. By the end of its first session, CXMT was up 466% from the offer price, showing both investor demand and the limits of conservative IPO pricing.

A person with knowledge of the matter, who was not authorized to discuss private deliberations, said CXMT’s launch was one factor behind a rapid official response when technology shares sold off in July. The person said the action was not designed only for CXMT, but it helped preserve conditions for a strategically important listing.

Markets replace some subsidies

Beijing’s approach marks a shift from the older playbook of subsidies, tax breaks and direct state investment. The new channel seeks to pull household savings, estimated in the source at about $26 trillion, toward companies in chips, AI models and advanced manufacturing.

The comparison with the US remains wide. Market data for the two years through August 7, 2026 showed Chinese technology firms raised about $217 billion from IPOs and bond sales, compared with about $1.4 trillion for US peers over the same period.

Chris Miller, a Tufts University professor and author of Chip War: The Fight for the World’s Most Critical Technology , said, "Over the past few years, US firms have had greater access to capital, but financing costs are rising." He also said China’s domestic computing remains more expensive because Chinese AI chips are lower quality.

Cheap debt widens Beijing’s toolkit

The debt market is becoming part of the same industrial policy. Chinese technology companies sold at least $38 billion of onshore and offshore bonds this year, the highest comparable total since 2016, but still about 7% of the $578 billion raised by US counterparts.

Funding costs give Chinese issuers a counterweight to that smaller scale. Major Chinese technology companies borrowed this year at an average coupon of 1.9%, more than 300 basis points below US peers and the widest gap cited since at least 2015.

Contemporary Amperex Technology Co. issued five-year yuan notes with a 1.58% coupon, compared with 5.25% on a similar-tenor dollar bond from South Korea’s LG Energy Solution Ltd. The spread reflects China’s lower interest rates and inflation as well as official pressure to keep credit flowing to strategic sectors.

Banks are still cautious where cash flows are thin. People familiar with regulatory discussions said lenders have been urged to provide stable financing to technology companies, while People’s Bank of China data showed tech loans accounted for 22% of new corporate lending in the second quarter.

Savings move into chip risk

Equity investors have moved with the policy signal, shifting money from property, consumer names and other older growth sectors into chipmakers and advanced manufacturers. The chip-heavy STAR 50 Index reached a record in June and was up 30% this year, compared with a 1.4% gain for the CSI 300.

The risk is that policy-backed demand outruns earnings. Gary Tan, a portfolio manager at Allspring Global Investments, warned that CXMT trades at a premium to global memory peers, suggesting scarcity and policy expectations are shaping near-term pricing.

Hongxu Wei, a senior economist at Anbound, framed the constraint more directly: "Capital is a necessary condition, but it is not a sufficient one." Talent, technology and commercialization will determine whether cheaper funding turns into durable AI capacity.

Scenarios turn on investor demand

If investor appetite holds, China can channel savings into AI infrastructure without relying as heavily on public balance sheets; that would support domestic demand, give CXMT a larger financing base and push more chip and model companies toward listings. If valuations instead detach from profits, the mechanism could transmit losses back to households, narrow CXMT’s room for follow-on fundraising and leave the sector with excess capacity.

A third path is slower but more durable: capital markets stay open, but regulators tolerate fewer speculative surges. In that case, the macro effect would be smaller, CXMT would need to prove production and margin gains after its IPO, and China’s AI industry would compete less on funding volume than on cost control and manufacturing scale.

More stories