Hello reader,

By now, you’ve heard the news about DeepSeek’s latest breakthrough. It’s been making waves for weeks, and for good reason: training large-scale models on public data is getting cheaper, faster, and more efficient. 

But here’s the part people aren’t talking about: this moment is a turning point, not the finish line.

We’ve reached peak public data. Every major AI model has been trained on the same scraped datasets, and the marginal value of retraining on that data is approaching zero. Public data has fueled an era of incredible AI advancements, but it’s also a dead end.

The next frontier? Enterprise data—proprietary, high-fidelity, and immensely valuable, but also sensitive and highly regulated.

This is where the conversation needs to shift. AI’s next chapter isn’t about squeezing a little more efficiency out of public models; it’s about how we securely integrate enterprise data into AI workflows without compromising privacy, sovereignty, or compliance.

That’s the problem we solve at Opaque. 

Our confidential AI platform makes it possible to unlock the full power of enterprise data without exposing it to unnecessary risks. We enforce policies. We guarantee privacy with cryptographic verification. We provide the proof enterprises need to operate AI at scale in regulated environments. No masking. No shortcuts. Just verifiable, confidential AI.

Opaque eliminates the need for traditional data masking by enabling confidential AI processing on encrypted data, ensuring full data utility while maintaining verifiable privacy and sovereignty.

And this goes beyond analytics and machine learning. The rise of confidential agentic workflows—AI systems making decisions on sensitive data—demands a new level of trust. Enterprises need more than just security; they need attestation, compliance, and interoperability. Opaque ensures every AI-driven action is verifiable, every interaction is policy-compliant, and every deployment is secure.

So while DeepSeek marks an evolution in how models are built, it doesn’t solve the real challenge ahead: how do we make AI truly enterprise-ready? At Opaque, we’re not just watching this shift—we’re leading it.

Aaron Fulkerson
CEO, Opaque Systems

In the Lab

The Latest Happenings at Opaque Systems

Opaque’s Chester Leung Spoke with BigDataWire on Keeping Data Private and Secure with Agentic AI
Opaque Co-Founder and Head of Platform Architecture, Chester Leung, shared with BigDataWire how Opaque is solving AI’s biggest roadblock—data privacy—by making confidential computing effortless for enterprises. With encryption, access controls, and secure AI architectures, Opaque is turning bottlenecks into launchpads, helping businesses take AI from the lab to real-world impact.

We Also Spoke with GovTech About Confidential AI in Government
Additionally, Chester Leung shared insights with GovTech on the critical need for secure AI processing in the public sector, where agencies handle vast amounts of sensitive data. “Just as government employees undergo rigorous background checks and need security clearances, AI systems used in the public sector must also be verifiably secure before handling private information,” said Leung.

AI in Action Podcast: Empowering Enterprises to Leverage Data Through Collaboration with Opaque
Chester Leung joined AI in Action podcast to discuss how Opaque is accelerating AI adoption by solving data privacy challenges. He shares how Opaque emerged from Berkeley research on secure collaborative learning—highlighting the team’s expertise in distributed systems, security, and machine learning—and further explores Opaque’s commitment to enhancing data governance and supporting agentic AI. 

Whitepaper: Confidential AI: Securing Data, Models, and Agents in the Age of AI
Check out our latest whitepaper to dive deeper into how confidential AI emerges as a solution at the intersection of AI, cloud, and confidential computing—enabling organizations to balance the need for innovation with stringent requirements for data privacy and compliance.

Confidential Computing Summit: Manu Fontaine, Founder and CEO of Hushmesh Announced as Featured Speaker
Manu Fontaine is revolutionizing internet security with "the Mesh," a groundbreaking infrastructure that integrates automated cryptographic security and universal zero trust. His visionary teachings will inspire transformative discussions on the future of secure and trustworthy AI. Don’t miss these insights, and more—register today for the 2025 Confidential Computing Summit at the Early Bird rate. Interested in contributing to the conversation? Submit a speaker proposal by Feb. 17.

Code for Thought

Worthwhile Reads

➡️ A shift in business AI adoption. Enterprise AI has shifted from experimentation to widespread, revenue-generating adoption, with global AI spending expected to reach $632 billion by 2028. GenAI is accelerating this transformation, growing at nearly twice the rate of traditional AI applications, enabling businesses to scale AI-driven solutions faster than ever.

📕 The DeepSeek lesson. Enterprises know AI’s value depends on access to their internal data—but they don’t trust LLM vendors to handle it securely. The future of AI will be defined by who earns customer trust. Opaque is committed to secure, privacy-preserving AI solutions, providing enterprises with confidence in how their data is used.

📊 GenAI is solving its own data growth crisis. CIOs are rethinking data strategy as GenAI demands faster, smarter access to massive datasets. The solution? GenAI itself. Intelligent tiering, predictive pre-fetching, and retrieval-augmented generation (RAG) are turning data management from a bottleneck into a breakthrough.

👍 The critical role of responsible AI. AI is transforming businesses, but without strong governance, it can compromise privacy and erode trust. As governments debate AI regulation, businesses must take the lead in operationalizing AI ethics, ensuring transparency and accountability. Confidential AI solutions enable organizations to adopt responsible AI while driving innovation.