This week, thousands of founders, engineers, and venture capitalists—including leaders from Microsoft, Writer, Cohere, and more—zipped up their Arc’teryx and descended on Vancouver for Web Summit’s first run in the new city. Set against the mountainous backdrop of the waterfront convention centre, the debut marked not just a change of scenery, but a westward handoff from Collision, which had long anchored Canada’s tech scene in Toronto. Snow-dusted peaks loomed just beyond the harbor’s edge—a quiet reminder that nature, and the real world, is always just one pane of glass away, no matter how many panels on AI infrastructure are underway inside.

If opening night was any indication, the program has a decidedly different tone than the months of guardrails off, risk on, unbridled AI optimism we’ve seen. Front and center: the moral debate over AI—who gets harmed, who benefits, who is accountable, what kind of AI we should be building and who decides that code for massive swaths of the population. 

“If we are going to hook these things up to the entire universe, we should guarantee they can function properly,” NYU’s Gary Marcus told an at-capacity crowd in his opening keynote, noting that the idea these problems can be solved by just adding more data “hasn’t worked.” 

Marcus’ comments follow a week where Anthropic’s Claude model reportedly resorted to blackmail when faced with being shut down; OpenAI’s ChatGPT Model refused to turn off when instructed; and on the regulatory front, the California Assembly passed AB 316 (nearly unanimously) to prevent companies from escaping liability by blaming their algorithms, while the Trump administration took a very different approach with a plan to ban state regulation on AI technology. 

BlueSky’s CEO Jay Graber held the line on the need to build frameworks around user consent and data use, along with the ever-shifting need for platform managers to keep up with AI bots and moderation. “What’s going to happen with the open web and what is the future?” Graber queried Web Summit’s audience, suggesting that as users get wiser about how LLMs and other AI companies are training across their posts, there will be a change in behavior.

But techno-optimism about the opportunity for AI to deliver incredible breakthroughs and value to consumers and investors remained on display, particularly in Deep Tech, and there is growing enthusiasm for employing the technology in specialized fields like biotechnology, where highly skilled humans can use it to solve the biggest problems of our time. “Someday it will completely revolutionize science,” Marcus said, nodding to the opportunity and recent breakthroughs like Google’s DeepMind, which last year helped researchers win a Nobel Prize in Chemistry. 

It’s a palpable shift, as AI coding applications are accelerating the capabilities of these highly specialized founder teams, and 30 years of SaaS dominance is giving way to a new era of technology and physical application. 

In the healthcare space, the technology is seen as especially promising for solving entrenched problems that have eluded medical professionals for decades. 

Tom Biegala, founder at Bison Ventures, nodded to the opportunity for accelerating development of treatments: “We invest in AI for super high value applications, from developing new drugs to improving clinical workflows, and find the best companies are those that deeply understand and use AI in a rational way, then use traditional workflows where applicable.” 

But, he cautioned, companies must understand where their technology fits within the broader healthcare ecosystem—including clinical workflows, billing systems, and incentive structures—because the tech doesn’t operate in a vacuum.

Perhaps the top takeaway is that having an AI strategy is essential for every founder and VC—regardless of industry, the technology is fundamentally changing how companies are built and run, with implications for investors. 

“Change is happening in minutes—not days, weeks, years as it used to,” Maria Palma, Freestyle Capital Founder, told a packed room on the VC stage, on how AI is driving efficiency, adding that she “cannot think of a job or industry that isn’t impacted by AI.”