Gemini delayed: low morale at DeepMind holds Google back

What happened?

Google DeepMind, Alphabet's artificial intelligence unit, is experiencing a significant delay in the launch of its flagship model Gemini 3.5 Pro. According to reports from Axios and Bloomberg, the model, which was supposed to directly compete with OpenAI's GPT-4 and Anthropic's Claude, is months behind schedule. But the problem goes beyond the technical: six current and former DeepMind employees revealed to Axios that low morale is the main factor holding back the lab's releases.

The news was picked up by The Next Web (reliability 78/100), which highlighted that the delay exposes a deeper problem: the demotivation of talent. Google has tried to counter the brain drain with million-dollar compensation packages, but the internal environment continues to deteriorate. The Axios report notes that employees cite excessive bureaucracy, lack of autonomy, and a culture that prioritizes safety over speed as causes of low morale. Additionally, Bloomberg indicates that Gemini 3.5 Pro is not only delayed but its performance on key benchmarks is below expectations, forcing Google to postpone its launch until at least late 2025.

Why is this important?

Google was for years the undisputed leader in artificial intelligence. However, since the arrival of ChatGPT in November 2022, the company has been losing ground to OpenAI and Anthropic. The delay of Gemini 3.5 Pro is not an isolated incident: it reflects a cultural and organizational crisis that could have long-term strategic consequences. Low morale at DeepMind not only affects development speed but also the ability to retain the world's best talent. In recent months, several key researchers have left the company to join startups like Mistral AI, Cohere, or even OpenAI. According to LinkedIn data compiled by The Information, DeepMind lost at least 15 senior researchers in the past year, many of whom cited lack of impact and bureaucracy as reasons.

This talent exodus is particularly concerning because DeepMind has always been a magnet for the world's best minds. The merger with Google Brain in 2023, rather than generating synergies, seems to have exacerbated cultural problems. Meanwhile, OpenAI has doubled its research staff and Anthropic has attracted key figures from Google. If the trend continues, Google could lose its competitive edge in a market where innovation is the currency. Historical context shows that Google already missed the boat on smart assistants against Alexa and Siri, and now risks repeating the same mistake in generative AI.

What consequences will it have?

• Product roadmap delays: The absence of a competitive model against GPT-4 and Claude could erode Google's market share in smart assistants, search, and cloud services. For example, Google Assistant has already lost ground to ChatGPT and Copilot, and without a cutting-edge language model, integration into products like Google Workspace or Android could be compromised. According to Gartner analysts, Google could lose up to 5% of the AI assistant market share by 2026 if it doesn't launch a competitive model soon.
• Accelerated talent drain: Demotivation and aggressive competition from startups and large tech companies could leave DeepMind without its brightest minds. The case of Mistral AI, which raised 600 million euros in 2024, is an example of how investors are betting on small, agile teams. Additionally, OpenAI has offered compensation packages including stock and signing bonuses that double those of Google, according to Bloomberg sources.
• Impact on brand perception: Google, once synonymous with AI innovation, risks being seen as a slow, bureaucratic giant. This is already reflected in developer surveys: in Stack Overflow 2024, only 12% of developers considered Google the AI leader, compared to 45% for OpenAI. Brand perception affects not only hiring but also business partnerships and investor confidence.
• Opportunity for competitors: OpenAI, Anthropic, and others can capitalize on Google's stumble to consolidate their leadership and attract disgruntled talent. Anthropic, for instance, has launched Claude 3.5 Sonnet, which outperforms Gemini 1.5 Pro on several reasoning tests. Meanwhile, startups like Cohere and Mistral are gaining traction in the enterprise market. If Google doesn't react, it could fall behind in the next wave of innovation, as it already did with social search or mobile operating systems.

What should readers know?

The case of Google DeepMind is a reminder that technological innovation depends not only on money or infrastructure but on people and organizational culture. Low morale is a symptom of deeper problems: lack of autonomy, excessive bureaucracy, and leadership that prioritizes safety over speed. Comparison with other companies is illustrative: while OpenAI fosters a culture of 'move fast and break things' (with the risks that entails), Google has adopted a more conservative approach after the LaMDA scandal and criticism over biases. However, this balance seems to be tipping too far toward caution, stifling creativity.

For industry professionals, this underscores the importance of a work environment that fosters creativity and calculated risk. Companies that want to lead in AI must prioritize culture over process. For investors, it's a sign that even giants can stumble if they don't care for their human capital. And for users, it means the next big AI innovation may not come from Google but from a more agile competitor. In fact, we are already seeing startups like Perplexity AI or ElevenLabs innovating in niches that Google has neglected.

"The delay of Gemini is just the tip of the iceberg. The real battle for artificial intelligence is being fought in the corridors of DeepMind, where morale is undermining Google's ability to compete." — Analyst at TheVortiq

In summary, Google's problem is not technical; it's human. And until it solves it, it will continue losing the AI race. The lingering question is whether Sundar Pichai and Demis Hassabis can turn things around before the damage becomes irreversible. History shows that companies like IBM and Xerox also had elite research labs but lost their edge due to cultural problems. Google faces the same risk.