A history of Google's relationship with China from its first foray in 2006 to Project Dragonfly, and how Google hopes to reenter China in the era of AI
Matt Sheehan / MIT Technology Review :
Context & Ripple Effects
Matt Sheehan's retrospective lands after a year of escalating disclosures about Google's China ambitions: August reporting detailed how the company used subsidiary 265.com to compile a list of sites China blocks for its censored Dragonfly search engine, an October transcript showed search chief Ben Gomes telling staff things that appeared to contradict Google's own claims of limited progress, and November interviews documented open tension inside the company between its founding ideals and its growth appetite.
The piece reframes that year as one chapter in a longer arc — from the 2006 market entry through withdrawal to today — and argues the reentry vehicle this time is AI rather than search alone. That matches the broader picture of Alphabet's multi-front push into China spanning search, smartphones, and autonomous cars, where every product line faces its own regulatory bar.
First-order effects
- Google's leadership now has to reconcile two public positions at once: official claims that Dragonfly is barely underway versus Ben Gomes' own staff remarks suggesting otherwise, making internal credibility the immediate constraint on any China move.
- The 265.com arrangement means Google's compliance work for China is already operational — site-blocking lists are being compiled under a subsidiary banner — so reentry planning is further along than the company has publicly admitted.
Second-order effects
- Every additional front in Alphabet's China push raises the price of the whole strategy: opposition that crystallized around censored search can extend to smartphones and autonomous vehicles, forcing Google to weigh whether one compromised product endangers the others.
- Staff dissent documented in the November interviews gives competitors and critics a ready-made narrative — a company trading founding ideals for growth — that shapes how any future China announcement will be received before Google controls the framing.
Third-order effects
- If the pattern holds, the durable lesson is that Western AI companies seeking China market access will need purpose-built compliance structures — subsidiaries, filtered data pipelines, state-negotiated product scopes — rather than porting domestic products with modifications.
- The 2018 cycle also establishes a template for how such efforts fail or stall: leaks, internal transcripts, and employee revolt as the primary checks, ahead of any formal regulator or legislature weighing in.
The trend: US tech firms are treating AI as the new wedge for state-compatible reentry into China, repeating the search-era bargain under terms the next technology wave sets.