Hi friends,

Confidential computing has tremendous potential, and we’re seeing new use cases emerge every day to unlock that potential even further. The technology is relatively new, and, as with any nascent technology, companies need concrete examples of how confidential computing can solve real-world problems and deliver value. 

The Internet had a similar trajectory. It began as a government project in the late 1960s, primarily for researchers and academics. But in the 1990s, with the development of the World Wide Web, browsers like Netscape, and the rise of ISPs, the Internet became accessible to the general public. Not only was it intuitive and easy to use, but its value also became evident. People began to see its transformative potential as it gradually revolutionized communication, commerce, and entertainment. Today, the Internet is virtually everywhere and life without it is unimaginable.

Confidential computing will eventually be everywhere, too. Our mission at Opaque is to offer an intuitive platform that delivers the technology seamlessly to meet a specific, timely need: accelerating AI into production, without adding new barriers to the flow of data or operations. 

Still, there’s an appetite for more success stories, more pilot projects, and more collaboration to showcase its tangible benefits and value. 

Consider the financial sector, where banks can collaborate and leverage trustworthy AI models to detect fraud without exposing sensitive customer information. Or the high-tech space, where confidential computing can enable companies to securely process sensitive data on cloud-based AI platforms, ensuring data remains encrypted during computation. These are the kinds of use cases that need to be highlighted and communicated more effectively.

In this issue, we’re bringing you an article inspired by the Confidential Computing Summit that dives into today’s top confidential computing use cases, as well as a roadmap for more applications. 

Let’s keep paving the way for a more secure digital future.

— Aaron Fulkerson, CEO at Opaque Systems

AI Applications Will Expand the Potential of Confidential Computing

In time, as the technology becomes increasingly essential and seamless, all computing will become confidential. And once it is, it will unlock a new world of secure data sharing, enable cloud-powered innovation, and accelerate AI models into production. But more awareness is needed around the full potential of confidential computing and the outcomes it can drive. Below is an excerpt from an article written by technology reporter Marshall Kirkpatrick, based on conversations he had at Confidential Computing Summit, where he discusses existing use cases and why more applications are essential.

While technical innovation moves quickly, most experts at the second annual Confidential Computing Summit agreed that the industry’s biggest challenge is building awareness of what’s already possible. Nearly everyone shared a vision that, in time, all computing will become confidential.

“Confidential computing is an enabling technology that allows customers to move to the cloud more than ever before,” said NVIDIA Chief Security Officer and Head of Product Security Dave Reber. “Companies that traditionally had to be on-prem or couldn't take advantage of the cloud for very sensitive workloads can now take advantage of everything that the cloud service providers have to offer, while ensuring their data is protected.” 

In short, confidential computing can unlock a whole world of cloud-powered innovation. But the problem today is that many businesses still don’t realize that confidential computing offers the protection they require to innovate. 

There are three main families of use cases for confidential computing today, explained Giuseppe Giordano, R&D Principal Director at Accenture Labs. The first is data sharing and collaboration. There is risk involved with bringing sensitive workloads to the public cloud, potentially giving exposure to users who aren’t authorized to access that cloud. Confidential computing addresses this by enabling computations on encrypted data, ensuring it remains secure throughout its entire lifecycle.

The second use case is multi-party collaboration. Confidential computing enables secure multi-party collaboration by allowing multiple parties to compute on shared data without exposing the underlying data to any of the involved parties. This is particularly useful in scenarios where data privacy and security are paramount, such as in financial services, healthcare, and collaborative research. 

”Confidential computing looks to be the solution to enable this collaboration and help those companies extract the value that they couldn't create on their own with the data and the models that they have,” Giordano said.  

The third and final use case is IP and model protection. A typical example is predictive maintenance models running in the cloud and having to send the data from an on-prem solution to the cloud to access those predictive maintenance use cases. Depending on the data, it can be very difficult to extract from a data center. Confidential computing moves the model instead of the data. “We move the model towards the edge so that the inference is done locally at the edge and the data stays local,” Giordano explained. 

But these use cases, while exciting, shouldn’t be the only stories we tell, said Anand Pashupathy, Vice President and General Manager Security Software and Services Division, at Intel. Awareness around the full potential of confidential computing—and the outcomes it can drive—simply isn't there yet. 

“We need more use cases and production solutions to give confidence to general purpose industry that they can bring their solution and deliver it in a confidential computing manner,” he said. 

Read the full article here.

In the Lab

The latest happenings at Opaque Systems

Securing the Enterprise

As generative AI gains traction in the enterprise, so too do concerns about privacy. Already, companies are feeling the impact of AI on the security of their data and are increasingly seeking ways to protect themselves without missing out on the tremendous potential of AI. In the video below, the first in a series, Aaron Fulkerson, CEO of Opaque Systems, and Ion Stoica, Professor UC Berkeley, ​Executive Chairman Databricks & Anyscale,​ and Co-founder & Board Member of Opaque Systems, discuss the trends that have led to this moment in the industry, and how confidential computing offers a promising path forward. For more, download our latest whitepaper, Securing Generative AI in the Enterprise.

ICYMI: Confidential Computing Summit 2024

Missed the Confidential Computing Summit? All videos and presentations from the event are now accessible on the conference website. Visit the Past Events and Agenda pages to watch detailed sessions and discussions from the event. 

Code for Thought

Worthwhile reads

📈 Confidential computing market set to reach $53 billion by 2029. Investment in confidential computing is on an upward trajectory. Knowledge Sourcing Intelligence forecasts that the market will grow at a compound annual growth rate (CAGR) of 43.85% through 2029 to reach a total of $53.214 billion. This spotlights the increasing significance of confidential computing as a viable solution to protect sensitive information and help organizations meet compliance requirements. 

🔐 OpenAI breach spotlights the challenges of keeping AI models secure. An incident where a hacker gained access to OpenAI’s internal messaging systems is raising new concerns about the security of and sentiment around AI systems. A former OpenAI researcher recently discussed the matter on a podcast, where he said the company isn’t implementing enough measures to prevent foreign adversaries from stealing its secrets. But consider this: if companies deployed their models in secure, confidential AI platforms from the start, additional guardrails wouldn’t be necessary. 

🛠️ AI emerges as a vital tool for hackers—and cybersecurity defenders. Amid a growing wave of cyber attacks, it's no secret that hackers are weaponizing AI to access sensitive information. At the same time, AI presents a valuable tool for defenders to identify and stop threats. Radware’s new Global Threat Analysis Report finds that 46% of organizations are already using AI in cybersecurity, while 43% are planning to use it in the future. Specifically, they’re using AI to analyze vast datasets, identify real threats, and deploy natural language-based query builders to gather relevant data and produce personalized security awareness training.

🔍 Research finds executives are optimistic on AI, cautious on data security. The majority of technology executives are excited about the potential of AI capabilities, but concerned about the security risks that are also posed by the technology, according to a survey conducted by identity and access management provider Okta. When asked about their POVs on AI and security, 70% said they plan to use AI to improve security and threat detection—while nearly the same amount (74%) said they were worried about AI’s potential impact on data privacy.