On Thursday, July 16, Alphabet lost about $200 billion in market value — a roughly 4.4% drop — after Bloomberg and CNBC reported that Gemini 3.5 Pro is delayed again, this time due to coding performance falling short of the company's own internal targets. The model was unveiled in May as Google's most capable yet, with a broad launch promised for the following month — a launch that never happened.
What Changed
According to Bloomberg and CNBC, the delay stems from coding performance below Google's own internal bar, and the company hasn't given a new launch date. The market's reaction was immediate and disproportionate to the event itself: a single unreleased model moved hundreds of billions of dollars in market value within hours.
Why It Matters
This isn't an isolated case of poor execution — it's a reminder of how risky it is to depend on a promised future AI capability, even from the best-capitalized lab in the world. If Google, with all its resources, can't reliably predict when — or whether — a promised capability will actually ship, no company should be planning its operations around a model roadmap that doesn't exist yet.
The Impact for Brazil
Brazilian companies delaying AI adoption decisions while "waiting for the next model" — Gemini, GPT, or anything else — are effectively betting their own operational timeline on a promise the vendor itself can't deliver predictably. This case shows the safer strategy is to build on what's already available and stable today, with an architecture that absorbs model upgrades without needing to rebuild the entire process every time a new version ships.
Entercast's Take
We made this same point in "From Pilot to Scale": the real bottleneck in AI adoption was never which model to wait for — it's having the process, data, and governance ready to make the most of the model that's already available. The Gemini episode is just more proof that betting your company's roadmap on the next model generation is a fragile strategy. Companies that build their operation on today's capability come out ahead when the next generation finally arrives — instead of losing time waiting for it.