At Morgan Stanley's TMT Conference, Sam Altman and Jensen Huang Warned AI Will 'Shock' Investors
Executives from leading AI labs told attendees that a non-linear capability leap is coming by mid-2026, while a Morgan Stanley survey found AI already caused a 4% average net workforce reduction globally.
The biggest names in AI and technology gathered in San Francisco last week for Morgan Stanley's annual Technology, Media and Telecom Conference, where the prevailing message was that artificial intelligence's economic disruption is arriving faster than markets have priced in. Executives from major AI laboratories, including OpenAI CEO Sam Altman and Nvidia CEO Jensen Huang, each warned in their own terms that the coming months would deliver capability advances that would surprise and potentially shock investors.
Huang argued that compute has become the defining economic input of the AI era, stating 'compute equals revenue' and noting that demand for AI computing power is 'higher than incredibly high.' Altman outlined a vision in which the near-term transition to agentic AI will enable companies of one to five people to outcompete much larger incumbents, a shift he said would compress into the next few years rather than play out over a decade. Morgan Stanley's own analysts predicted a non-linear jump in model capabilities will become evident between April and June of 2026.
The conference also surfaced sobering data on AI's workforce impact. A Morgan Stanley survey of roughly 1,000 executives across five countries found an average net workforce reduction of 4 percent over the past 12 months that was directly attributable to AI adoption.
University of Chicago economist Alex Imas told the conference he was 'amazed and alarmed' by upward revisions in aggregate productivity data confirming AI gains are now visible in macroeconomic measurements, not just micro studies — a shift that previous research had struggled to demonstrate.
The most revealing moment may have come from xAI co-founder Jimmy Ba, who announced his retirement at the conference and stated publicly that recursive self-improvement loops in AI are 'likely' within the next 12 months, adding that '2026 is going to be insane and likely the busiest and most consequential year for the future of our species.' The remarks, delivered by a researcher with direct visibility into frontier model development, underscored the sense among AI insiders that the pace of progress is accelerating beyond what public benchmarks and product releases have so far communicated.
Morgan Stanley's analysts also flagged a deepening infrastructure constraint: the United States faces a projected power shortfall of 9 to 18 gigawatts through 2028, representing a 12 to 25 percent deficit in the electricity needed to support planned AI data center expansion. The energy bottleneck, analysts said, could determine which companies can sustain the compute buildout required to participate in the next wave of AI capability gains.
Read the original reporting at Fortune.
