Demis Hassabis stepped down as Google DeepMind CEO on August 5. Here is what the leadership shakeup and Gemini 4 pivot mean for your Google Ads, AI Overviews, and local search.
Ido Cohen · Published 2026-08-07 · AI News
Google just made the biggest leadership move in its AI history, and it affects every service business running ads, depending on local search, or showing up in AI Overviews. On August 5, 2026, Demis Hassabis — the co-founder and CEO of Google DeepMind — stepped down from day-to-day operations, handing control to Koray Kavukcuoglu as the new Senior Vice President of Google DeepMind. Alphabet shares fell roughly 4% on the news. If you're a plumber, dentist, real estate agent, or HVAC company that lives and dies by Google, you need to understand why this happened and where it's heading.
This is not a quiet internal shuffle. Three things happened simultaneously on August 5, and the combination is significant.
First, Hassabis transitioned from CEO to Chair of Google DeepMind and Chief Scientist of Alphabet. According to Fortune's reporting, the move formalizes "a shift away from day-to-day management and toward AGI strategy and scientific applications of AI." He is also dedicating more time to Isomorphic Labs, the AI drug discovery company he co-founded, which Google has backed. In his note to staff, Hassabis said he believes artificial general intelligence is "close at hand" — a remarkable statement from the man who built the lab.
Second, Jeff Dean — a 27-year Google veteran and one of the most respected engineers in tech history — announced he was leaving to co-found a new AI startup called Discovery Loop alongside three other departing Google researchers. According to Tech Journal, Discovery Loop is a public benefit corporation that aims to make AI capable of recursive self-improvement to speed up scientific discovery, with Google itself as a founding investor and cloud partner. That detail matters: even Google's own departing talent is building on top of Google's infrastructure.
Third, Koray Kavukcuoglu — previously DeepMind's Chief Technology Officer and Google's Chief AI Architect — now owns the entire Gemini roadmap as SVP reporting directly to Alphabet CEO Sundar Pichai. According to The Decoder, his remit includes Gemini model development, frontier AI research, the Gemini app, and all of Google's AI developer platforms.
The question for service businesses is not "who is Koray Kavukcuoglu?" It's what problem this restructure is trying to fix — and how fast.
Google did not restructure for the fun of it. Geeky Gadgets reported that DeepMind "faced setbacks in 2026, including delays and underperformance of Gemini 3.5 Pro, 3.5 Flash and 3.6 Flash models, highlighting inefficiencies in decision-making and organizational agility." Meanwhile, OpenAI and Anthropic have been shipping faster, grabbing developer mindshare, and — most importantly for this audience — eating into the search queries that used to default to Google.
Sundar Pichai's memo to staff pointed to strong AI momentum at earnings across Search, YouTube, and Cloud, and noted that the Gemini app has reached 950 million monthly users. That sounds impressive, but 950 million users of the Gemini app is still well behind where Google wants to be in the race against ChatGPT and Claude.
According to Bloomberg's coverage, the restructuring is Google "concentrating its artificial intelligence leadership at its Mountain View, California headquarters to gain momentum in the accelerating race against Anthropic and OpenAI to build the world's dominant AI models." The phrase "gain momentum" in an official Bloomberg piece is a polite way of saying Google is behind.
For service businesses, the practical read is this: the tools you rely on — AI Overviews, Google Business Profile AI features, Gemini-powered ad bidding, and local search answers — all run on Gemini. A lab that admits it underperformed its last two model generations, and is now under new operational leadership, is a lab that is about to ship aggressively.
The clearest signal from this restructure is the pivot toward Gemini 4. According to Geeky Gadgets, "the company is pivoting its focus to the development of Gemini 4, a next-generation AI model designed to redefine its capabilities and regain its competitive edge." Kavukcuoglu now fully owns that development.
Here is why Gemini 4 is not an abstract developer story:
When Gemini 4 ships, expect all of these to shift. Some shifts will help service businesses (better local intent recognition, more accurate smart bidding). Some will create new disruption (AI Overviews expanding into even more commercial queries, reducing direct clicks). The businesses that are paying attention will adapt first.
Let's be honest about a real risk embedded in this story. This is not just a leadership reshuffle at the top. Time magazine reported that the restructure "comes as Google attempts to catch up to industry leaders OpenAI and Anthropic" — and that Google has been losing top talent at a rapid pace. Jeff Dean's departure is the highest-profile example, but he left alongside three other senior Google researchers (Quoc Le, Oriol Vinyals, and Sanjay Ghemawat). John Jumper, a DeepMind scientist who shared a Nobel Prize with Hassabis for AI protein structure prediction, left earlier in 2026 to join Anthropic.
When a lab loses its CEO, its chief scientist, and several Nobel-level researchers in the same quarter, that is a brain drain, not a routine restructure. It does not mean Google is finished — Alphabet still has world-class compute, trillion-dollar distribution, and the search monopoly. But it does mean there is a real possibility that Gemini 4 takes longer to ship than the new leadership wants, or that it underperforms against whatever OpenAI and Anthropic ship in Q4 2026.
For service businesses, this matters in one specific way: do not build your entire marketing strategy around one platform's AI features. The businesses that will win in 2026 and 2027 are the ones treating AI search presence across Google, ChatGPT, and Perplexity as parallel channels, not assuming Google will always dominate the answer layer.
Here is a quick comparison of where the three major AI labs stand as of this week — framed around what matters for service businesses.
The key takeaway from this comparison: Google's AI stack is in a transition period. Meta's ad AI (which runs your Facebook and Instagram campaigns) is actually in a more stable operational state right now. That is not a reason to abandon Google — it is a reason to make sure your Meta presence and your AI search presence are not being neglected while everyone stares at Google's org chart.
You do not need to panic, but you do need to act. Here are five concrete moves for service business owners:
1. Audit your AI Overviews presence before Gemini 4 ships. Search five to ten of your most important service + location queries on Google right now. Screenshot the AI Overviews results. Are you cited? Is a competitor? This is your baseline. When Gemini 4 rolls out — likely Q4 2026 based on the leadership's stated urgency — the citation patterns will change and you will want to know what shifted.
2. Check your Google Business Profile for AI-generated content errors. Gemini currently pulls from your GBP to populate AI Overviews answers about hours, services, and reviews. With a model transition coming, review your GBP now to make sure the data is accurate, complete, and up to date. Bad data in a weaker Gemini can become cited bad data in a stronger Gemini 4.
3. Review your Google Ads bidding targets this week — not next week. As noted in the existing August 17 tCPA/tROAS change (a separate Google Ads update taking effect August 17), your bid targets are about to be enforced more strictly. The combination of a model transition at DeepMind and a bidding mechanics change in Ads is a double reason to log into your Google Ads account now and confirm your Target CPA and Target ROAS reflect your actual cost-per-lead, not optimistic guesses.
4. Do not consolidate all AI visibility efforts onto Google. Spend 30 minutes this week asking ChatGPT, Perplexity, and Claude a few of the questions your best customers ask before hiring you. If you are a dentist in Austin, search "best dentist for dental implants in Austin." If you are a plumber in Seattle, search "emergency plumber Seattle." Note whether your business is mentioned. This is the beginning of your AI citation diversification strategy.
5. Watch the Gemini 4 launch announcement. Sign up for Google's developer blog updates and Search Central Blog. When Gemini 4 drops — likely announced at a Google event — that is the single most important moment to update your website's structured data, content, and GBP to align with how the new model categorizes service businesses.
---
What does Demis Hassabis stepping down actually mean for everyday Google products?
Hassabis built DeepMind into the world's leading AI research lab, but he had been stepping back from Gemini's day-to-day operations for months before the formal announcement. Koray Kavukcuoglu, who now owns the Gemini roadmap, has been at DeepMind for 13 years and already led the team building Gemini. For most Google products, the day-to-day work is likely to continue without immediate disruption — but the medium-term bet is a more aggressively shipped Gemini 4 under a leader with something to prove.
Will Gemini 4 change how Google AI Overviews work for local service searches?
Almost certainly. Each Gemini upgrade has expanded AI Overviews into more query types and changed which sources get cited. Gemini 3.5 and 3.6 underperformed expectations, which may explain why AI Overviews have felt inconsistent in local search over the past few months. Gemini 4 is positioned as a major capability jump, which means AI Overviews for "near me" and service-category queries will likely become more expansive and more authoritative — good news if you are cited, bad news if you are not.
Should service businesses be worried about Google losing AI talent to OpenAI and Anthropic?
It is worth watching but not worth panicking over. Google still has by far the most distribution of any AI company — Search, YouTube, Android, Chrome, and Gmail give Gemini access to billions of users that OpenAI and Anthropic cannot match. The talent drain is a real headwind on model development speed, but it does not threaten Google's position in search or advertising in the near term. The bigger risk for service businesses is assuming Google's AI Overviews and ad products will stay stable during a model transition period — they will not.
How does the Google DeepMind restructure affect Google Ads performance?
Google Ads Smart Bidding — the system that manages Target CPA and Target ROAS campaigns — runs on Gemini-class models. Leadership changes do not flip a switch on active ad accounts overnight, but model upgrades (and the disruption that precedes them) historically create periods of bidding volatility. Combined with the separate August 17 tCPA/tROAS enforcement change, this is a bad time to leave your Google Ads account on autopilot. Review your targets, confirm your conversion tracking is accurate, and plan for two to four weeks of performance volatility around any Gemini 4 launch.
Is this a good time to shift budget from Google to Meta ads?
Not necessarily a shift, but a rebalance is worth considering. Meta's Advantage+ ad system is currently in a more operationally stable state than Google's, which is mid-restructure. If your Google Ads performance has been inconsistent in recent months, diversifying some spend into Meta's AI-driven ad formats is a reasonable hedge — not an abandonment. The businesses that win are rarely all-in on one platform; they run Google for high-intent search and Meta for awareness and retargeting simultaneously.
---
Sources: