Last week I attended the Emerging Technology Conference at MIT, which brought together researchers and industry leaders across AI, Quantum, Robotics, Energy, Biotech, and Material Sciences. It was a week full of learning, networking, and broadening my horizons as to what’s currently technologically possible and what could be achievable in the next 3–5 years. At one point, I even held a small piece of uranium (photo below), supposedly enough to power a human body for a day—once our energy systems can support that kind of density.
Here are four ideas that stood out to me.
1. The Way We Interact and Transact on the Internet Is Quietly Changing
The way we interact with information online is changing faster than most businesses realize. More people now prefer to use chatbots or AI summaries that give them answers instantly instead of visiting individual websites. This shift is rewriting how internet traffic, discovery, and monetization work—and raises the possibility of information gatekeeping by a handful of firms. Both consumer and enterprise organizations should pay attention, as these changes will disrupt current business models.
For almost three decades, Google’s model relied on indexing and monetizing as much of the web as possible. Ten years ago, Google sent about one visitor for every ten pages it scraped. For large language models like ChatGPT or Claude, that ratio is estimated at 35,000–44,000 pages per query. Matthew Prince, CEO of Cloudflare, has spoken extensively about this transformation (for example, in his recent interview with Wired).
The economics of information discovery are shifting. The next evolution in “consumerizing” enterprise applications may be to make them accessible through these new conversational or summary-based interfaces. The main driver is user experience—just as the iPhone was a leap forward in usability. But this change will upend business models built on search traffic and discovery. Media organizations are the early test case: Reddit now licenses its data to AI firms at roughly seven times the rate of The New York Times.
We’re seeing a move from interacting with individual creators to engaging with collective intelligence models. That shift opens major questions about how internet-era businesses will adapt—or fail to.
2. Quantum Computing May Be Hitting an Inflection Point
Quantum computing has long lived in the “five years away” zone, but optimism at MIT this year felt different and more technically grounded. MIT Technology Review named Quantum the technology to watch for 2026 and beyond. Advances in logical qubits and error correction are reducing the number of physical qubits needed for meaningful computation (see this recent MIT Review article).
This progress is driving renewed confidence that useful quantum advantage may soon be achievable. Improvements in error correction aren’t incremental—they’re reshaping the scaling curve. Some early (and debated) results show quantum advantage in niche use cases, like quantum-enhanced feature engineering in finance. MIT and USC researchers remain cautious, but promising data is emerging.
Hardware is also progressing. Annealing systems are nearing 5,000 qubits, and neutral atom machines from companies like Pascal and Infleqtion may exceed 10,000 next year. NVIDIA’s Krysta Svore described their “qubit-agnostic” approach—combining classical GPU-based ML with quantum processors.
If current trends continue, utility-scale quantum computers could be commercially relevant before the end of the decade.
3. Enterprise AI: Moving From Experiments to Infrastructure
In 2023–2024, most enterprises treated AI as a productivity experiment. In 2025 and beyond, the focus is shifting to governance, security, and integration. Organizations are building a bridge from productivity gains toward standardized AI infrastructure that embeds trust, compliance, and data sovereignty.
Last year’s pilots focused on coding assistants, marketing content, and customer support. The next phase is about scale: identifying the right use cases, integrating agents into enterprise systems, and building guardrails that protect data integrity.
Several speakers emphasized that building agents from scratch isn’t sustainable. The better approach is partnering with trusted providers, enabling broad experimentation, and maintaining a secure data corpus that supports safe agentic behavior.
CIOs will need to evolve their ERP, CRM, HR, and analytics platforms to support agentic operations rather than replacing them outright. The key question for 2026 isn’t “how many copilots can we run?”—it’s “how well are they governed, measured, integrated, and delivering ROI?”
4. Skills and Mindsets for the Next Wave
The conference was rewarding in many ways, especially the conversations with industry visionaries. But the most important takeaway in today’s uncertain environment was about the skills needed to thrive in an AI-native world. Those skills aren’t prompt engineering—they’re curiosity, adaptability, and continuous learning.
The people who will thrive are those who can absorb new tools quickly and apply them meaningfully within their domain. Many still treat AI systems as smarter search engines, but the real value emerges when teams redesign workflows around collaboration with these systems.
Final Thought
EmTech MIT 2025 offered a grounded, realistic glimpse into the near future of technology. The internet’s rearchitecture, the emergence of usable quantum machines, AI’s shift into enterprise infrastructure, and the growing importance of lifelong learning—along with rising energy demands and geopolitical dynamics—will define 2026 and beyond.
Conversations about AI have moved from speculation to implementation. The sci-fi elements can wait; what’s emerging now is real, measurable, and already reshaping how we work and innovate.
“I’ll be back!”
