Aidan Gomez, the 28-year-old CEO of Cohere and co-author of the groundbreaking 2017 “Attention is All You Need” paper, has weighed in on DeepSeek’s disruptive entry into the AI market, offering a nuanced perspective on the Chinese startup’s capabilities and limitations for enterprise adoption.
While Gomez acknowledges that DeepSeek’s R1 model validated Cohere’s strategy that “spending billions of dollars a year isn’t necessary to produce top-tier tech that’s competitive,” he maintains that the model falls short of enterprise requirements. The former Google researcher called DeepSeek’s release “really impressive” but emphasized that businesses need customized AI solutions rather than off-the-shelf models, particularly when handling sensitive proprietary data.
Cohere, valued at $5.5 billion, focuses specifically on building AI for enterprises, giving Gomez unique insight into corporate AI adoption challenges. He explained that enterprises “don’t just want to buy a model” but require extensive customization with proprietary data to unlock real value. This process demands significant technical resources and time investment, making a simple DeepSeek deployment insufficient for most business needs.
Data privacy concerns loom large in Gomez’s assessment, particularly given DeepSeek’s Chinese origins and servers. With US lawmakers seeking to ban the startup’s software from government devices, enterprises remain “hesitant to build systems that touch sensitive data.” Gomez positioned Cohere as offering more secure alternatives, stating “our competitors treat it in a way that’s less secure.”
Looking beyond base models, Gomez identified agentic AI as the industry’s next frontier. Cohere recently launched North, an early-access agentic AI platform designed for specific enterprise workloads. This aligns with predictions from Nvidia CEO Jensen Huang, who declared 2025 “the year” for autonomous AI agents. Gomez sees agents as Cohere’s competitive edge, allowing enterprises to move beyond simple model deployment toward autonomous operations.
On infrastructure spending, Gomez challenged the necessity of massive capital investments like Sam Altman’s $500 billion Stargate project. He argued that focusing spending on training rather than inference represents a strategic mistake, with DeepSeek serving as “a big proof point” that capital-intensive approaches aren’t required for competitive AI.
Despite his reservations about enterprise readiness, Gomez views DeepSeek as a positive disruptive force, particularly for validating open-source approaches and demonstrating training efficiency. However, his core message remains clear: “It’s not just enough to download a model” for enterprises seeking meaningful AI integration.
Key Quotes
I think it validated Cohere’s strategy that we’ve been pursuing for a while now. Spending billions of dollars a year isn’t necessary to produce top-tier tech that’s competitive.
Aidan Gomez, Cohere’s CEO and co-author of the seminal transformer paper, explained how DeepSeek’s cost-efficient approach vindicated his company’s strategy of building competitive AI without massive capital expenditures.
We don’t see the enterprises that we sell to relying on R1 to power their systems. We don’t see it as a competitor on our side.
Gomez dismissed DeepSeek as a direct competitor for enterprise customers, emphasizing that businesses require customized solutions with robust data privacy rather than off-the-shelf models.
What we’re seeing from enterprises is that they don’t just want to buy a model. You’re going to have to build something with that model, you’re going to have to deploy a lot of technical resources to see value, and it will take time.
The Cohere CEO outlined the gap between impressive AI demos and enterprise-ready solutions, highlighting the extensive customization and integration work required for business deployment.
The fact that they published their training efficiency numbers let people see that it doesn’t need to be so capital-intensive to publish fantastic models.
Despite his concerns about enterprise readiness, Gomez praised DeepSeek’s disruptive impact on industry assumptions about the capital requirements for developing competitive AI models.
Our Take
Gomez’s perspective reveals a critical disconnect between AI model performance and enterprise value creation. While DeepSeek’s technical achievements are genuine, the enterprise AI market operates on different principles than consumer AI—security, customization, and integration trump raw model capabilities.
The emphasis on agentic AI represents a strategic pivot from the base model arms race toward application-layer differentiation. This shift favors companies with deep enterprise relationships and integration expertise over those simply offering powerful models.
Most significantly, Gomez’s infrastructure spending critique challenges the “bigger is better” paradigm dominating Silicon Valley AI strategy. If DeepSeek proves that efficiency matters more than scale, it could trigger a fundamental reassessment of AI economics, potentially benefiting nimble players like Cohere over capital-intensive competitors. The real test will be whether enterprises agree with Gomez’s assessment or embrace DeepSeek despite his warnings.
Why This Matters
This analysis from a key AI industry figure provides critical perspective on the DeepSeek phenomenon that rattled markets and sparked $1 trillion in stock value losses. Gomez’s insights matter because they come from someone who literally helped create the transformer architecture underlying modern AI, giving his technical assessments significant credibility.
The enterprise AI adoption gap Gomez identifies represents a crucial market reality often overlooked in hype cycles. While DeepSeek demonstrated impressive cost efficiency, the path from impressive demos to production enterprise systems involves customization, security, and integration challenges that favor established players like Cohere, OpenAI, and Anthropic.
The shift toward agentic AI that Gomez emphasizes signals where industry value creation is heading in 2025. Rather than competing solely on base model performance, companies are racing to build autonomous systems that can perform complex tasks—a capability requiring deep enterprise integration rather than simple model deployment.
Finally, Gomez’s questioning of massive infrastructure spending challenges the prevailing Silicon Valley narrative that AI leadership requires unlimited capital. If DeepSeek’s efficiency claims hold up, it could fundamentally reshape investment strategies and competitive dynamics across the AI industry, potentially democratizing access to cutting-edge AI capabilities.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- OpenAI’s Valuation Soars as AI Race Heats Up
- Mistral AI’s Consumer and Enterprise Chatbot Strategy
- The AI Hype Cycle: Reality Check and Future Expectations
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- Google’s Gemini: A Potential Game-Changer in the AI Race
Source: https://www.businessinsider.com/cohere-aidan-gomez-deepseek-ai-r1-businesses-2025-2