"Step up efforts to advance AI innovation breakthroughs, accelerate key technology R&D and ultra-large-scale intelligent compute cluster construction, deeply implement the 'AI+' initiative, and firmly hold the safety bottom line."

On June 29, Premier Li Qiang chaired a State Council executive meeting dedicated to hearing a report on AI development. That passage was the core AI statement in the day's official release. Fifty-two characters, but every word carries weight.

This isn't the first time the State Council has mentioned AI, but the timing and context are unusual. Anthropic's Mythos model hit US export controls within 72 hours of release; the H20 chip "indefinite" ban is still reverberating; DeepSeek just closed a record 51 billion RMB funding round; the global compute gap is roughly 3x -- at this juncture, the highest-level dedicated briefing sends a far more powerful signal than ordinary documents.

Intelligent Compute Clusters: From "Encourage Building" to "Accelerate Building"

The most notable wording shift is around "ultra-large-scale intelligent compute clusters."

Previous policy documents used language like "coordinated layout" and "encourage construction" for compute infrastructure; this time it went straight to "accelerate construction." Behind the upgraded language: the supply-demand contradiction in compute has reached a point where national-level action is necessary.

According to earlier Tom's Hardware reporting citing sources, China is drafting a national AI data center grid plan worth approximately $295 billion (~2.1 trillion RMB), targeting 80% domestic chips. On June 29 the first Global Space Compute Conference opened in Beijing's Zhongguancun, inaugurating the Beijing Space Compute Innovation Center with a 5 billion RMB three-year industrial fund. The National Data Administration previously released a plan proposing exploration of new trading models including Token trading -- the first time any nation has proposed building a data value system on a Token basis.

Hygon Information Technology's June 29 Hong Kong IPO provides an observation window: it opened up 74.56% at HK$199, with a market cap of approximately HK$10.2 billion. Domestic compute chip companies' red-hot reception in capital markets clearly resonates with the policy direction.

"AI+": From Technology Breakthroughs to Industrial Deployment

Another keyword is "deeply implement the 'AI+' initiative."

"AI+" isn't a new concept, but adding "deeply implement" means policy focus is shifting from "issuing documents" to "seeing results." The meeting specifically noted the need to "leverage our advantages of a complete industrial system and rich application scenarios to accelerate the scaled application of intelligent products and services."

In plain English: stop benchmark-chasing in the lab; go solve problems in real industrial scenarios.

This orientation aligns with recent industry data. A Goldman Sachs report on June 29 estimated that AI infrastructure stocks are expected to contribute approximately 60% of S&P 500 EPS growth in Q2, with Micron and Nvidia together accounting for over 40%. But China's AI deployment has its own path advantages: rich manufacturing scenarios, a massive consumer internet user base, and strong government digitalization foundations. Bloomberg Economics estimates AI could contribute 0.2-0.7 percentage points annually to Germany's economy over the next decade, corresponding to cumulative output up to $368 billion -- China's application scenario scale is multiples of Germany's.

Safety Bottom Line and Development Initiative

Another easily overlooked but critical line in the release: "Improve support policies and governance systems, and firmly grasp the initiative in development."

"Development initiative" -- those five characters carry deep meaning against the backdrop of US-China AI competition. It points to compute autonomy (domestic chip substitution), model safety (content governance and alignment), and data sovereignty (Token trading and data factor-ization). This isn't a conservative security view; it's finding balance between openness and autonomy -- using global resources well, but ensuring that critical chokepoints can't be cut off.

The A-share market's reaction on June 30 already illustrates how the market reads the policy: the Shanghai Composite approached 4,000 points, the ChiNext index kept making new highs, and AI sectors led gains. But a cool head is needed: favorable policy winds don't mean every AI company gets an equal share. The State Council's signal is clear: reward those genuinely doing technology breakthroughs and industrial deployment; weed out the storytellers and concept-pumpers.

The second half has begun.