AI’s rapid rise dominates CES 2026 keynote conversation
A common refrain echoed across the CES 2026 keynote circuit: AI is reshaping technology at a pace and scale unlike prior platform shifts. That theme carried into a live taping Tuesday of the All-In podcast, where co-host Jason Calacanis interviewed Bob Sternfels, Global Managing Partner of McKinsey & Company, and Hemant Taneja, CEO of General Catalyst.
The discussion centered on two questions increasingly defining boardroom agendas: how AI is changing investment outcomes and how it will reorder the workforce. Both guests argued that the current moment is not simply another tech cycle, but a structural reallocation of capital and labor driven by fast-improving models, falling deployment friction, and a growing belief that software “agents” will become ubiquitous across industries.
From $100 billion to “a couple hundred billion”: the new valuation tempo
Taneja described what he sees as unprecedented velocity in the scaling of leading AI companies. “The world has completely changed,” he said, pointing to the speed with which high-profile model builders have accumulated market value and investor interest.
As an illustration, he contrasted the timeline of Stripe—which he said took roughly 12 years to reach a $100 billion valuation—with Anthropic, a General Catalyst portfolio company. Taneja said Anthropic rose from about $60 billion last year to “a couple hundred billion dollars” this year, underscoring what he characterized as a new valuation tempo for frontier-model companies.
Looking ahead, Taneja argued that a fresh cohort of trillion-dollar enterprises could emerge from the current wave. “That’s not a pie-in-the-sky idea with Anthropic, OpenAI, and a couple of others,” he said, framing the moment as one where market size assumptions are being rewritten by the breadth of use cases and the speed of adoption among developers and early enterprise customers.
Why many CEOs hesitate: the CFO-versus-CIO dilemma
Calacanis pressed both guests on what is powering the apparent acceleration. Sternfels offered a more nuanced view from the vantage point of McKinsey & Company, noting that while experimentation is widespread, full-scale adoption remains uneven—especially outside the technology sector.
Many organizations are running pilots, testing copilots, and exploring productivity tools, he said. But for non-tech enterprises, the leap from trial to transformation often stalls at the same decision point: whether leadership can justify the spend, reorganize workflows, and accept short-term ambiguity in exchange for long-term competitiveness.
Sternfels said a recurring question he hears from CEOs is: “Do I listen to my CFO or my CIO right now?” In his telling, CFOs frequently argue that returns are not yet clear enough to warrant aggressive investment, especially amid budget scrutiny and competing priorities. CIOs, by contrast, warn that delaying adoption is “crazy” because “we’ll be disrupted,” he said.
The tension highlights a broader challenge: AI ROI can be difficult to measure early, particularly when benefits require process redesign, new data practices, and changes in how teams make decisions. Yet the perceived cost of waiting—ceding advantage to faster-moving rivals—can be just as hard to quantify.
Workforce anxiety and the future of entry-level work
The conversation also turned to labor-market disruption, a topic increasingly raised by employees and policymakers as AI tools become capable of handling tasks once reserved for junior staff.
“Some people are looking at AI and they’re scared,” Calacanis said, pointing to fears that automation could hollow out entry-level roles traditionally filled by recent graduates. He asked Sternfels and Taneja what guidance they would offer young people navigating an economy where software can draft, summarize, analyze, and execute a growing range of routine work.
Sternfels emphasized that while models may complete many tasks, humans still need to provide judgment and creativity. In his view, these are the capabilities that will remain differentiators as tools become more powerful and widely available. The implication for early-career workers: focus not only on technical fluency, but also on decision-making, problem framing, and the ability to synthesize context.
Taneja framed the shift as a permanent change in how careers will be built. He argued that “skilling and re-skilling” must become continuous. “This idea that we spend 22 years learning and then 40 years working is broken,” he said, suggesting that periodic reinvention will be required as tools and workflows evolve.
Calacanis agreed, arguing that the speed of building an AI agent may soon outpace the time it takes to train a new worker for certain tasks. In that environment, he said, standing out will require initiative and intensity—“chutzpah, drive, passion”—alongside adaptability.
McKinsey’s agent-heavy future—and a shifting headcount mix
Sternfels offered a concrete example of how large professional services firms are preparing for the change. He said he expects McKinsey & Company to have as many “personalized” AI agents as employees by the end of 2026.
However, he cautioned that this does not necessarily translate into a smaller workforce. Instead, he described a rebalancing of roles: increasing employees who work directly with clients by 25% while reducing back-office positions by the same percentage. The message was that AI may compress certain support functions while expanding the capacity for higher-touch, client-facing work—at least for firms able to redeploy talent rather than simply cut it.
For executives watching the sector, the exchange underscored a developing consensus: AI is not only a product category, but a reorganization force. The remaining question for many companies is whether they can move from experimentation to durable advantage before the next wave of AI-first competitors resets expectations again.
Related
