Zhipu AI Rejects ASI by 2030 - GLM-4.6 and Paid Coding Plan Chase Enterprise Revenue

Zhipu AI (Private)'s boss put a cold bucket on the "ASI by 2030" hype this week. Zhang Peng said a machine that beats humans across the board by 2030 is unlikely - models may outdo people in pockets, but will still lag in many other areas.

That runs counter to some loud forecasts floating around Silicon Valley. OpenAI (Private)'s CEO has suggested the decade could deliver artificial superintelligence, and Masayoshi Son of SoftBank Group (TYO: 9984) has offered an even later date. Zhang's take is more measured: progress will be uneven and domain-specific, not a single overnight leap.

Timing aside, Zhipu used the moment to push out a new model: GLM-4.6. The upgrade - billed as stronger on coding, reasoning, writing and agent-style tasks - is clearly aimed at winning enterprise contracts, not just impressing headline writers. Zhipu, born from Tsinghua University in 2019, has been positioning for a mainland listing and flagged that overseas revenue is starting to matter.

There are some practical details buried in Zhang's comments that traders should note (note: this is reporting, not advice). Chinese consumers have been slow to pay for chat-style services compared with the U.S., so Zhipu's near-term push is toward business customers and developer subs - hence a paid coding plan aimed at devs. That's where real revenue can be shown on a spreadsheet, not in vaporous ASI predictions.

What this means for the market: the debate about "when will machines rule?" is moving toward a debate about monetization, model quality and enterprise adoption. Hardware names and cloud suppliers remain on the front lines - think companies supplying GPUs and hosting - so the economic story is about contracts and margins, not sci-fi timelines. For example, AI chip and cloud providers like NVIDIA (NASDAQ: NVDA) still factor heavily into the plumbing of every big model rollout.

Risks are obvious. Chinese firms face tougher domestic regulation and different consumer payment behavior. Competing U.S. models still set the bar in many consumer markets. Price competition will intensify as more players ship capable models, which could pressure ARPUs for subscription products. Zhipu's message - partial superiority in some areas, material gaps in others - suggests investors and analysts who parse revenue growth and enterprise traction will get a clearer read than those fixated on end-of-decade prophecies.

So the soundbite is simple: Zhang pushed back on the boldest ASI timelines and pointed the conversation toward product upgrades and selling to businesses. The market's next move will likely prize measurable contracts and sustainable revenue streams over doomsday or moonshot forecasts. Still curious which companies will turn GLM-4.6-style advances into predictable top-line growth?