Two things are true at the same moment as you walk past the job boards at each university career center in the nation: students are nervous about the job market, and the job market is nervous about them. It’s not that there isn’t any work—in fact, there is more than most hiring managers can handle—but rather that the skills needed to execute it have changed more quickly than any training pipeline has been able to keep up. Automation and displacement are expected to make headlines in 2026. For a significant portion of the economy, the reality is more akin to the opposite: businesses are hiring heavily but failing to fill positions.
Even if the word “restructuring” has been awkwardly frequently appearing with job advertising, the major internet companies are still in expansion mode. The positions that Google and Meta are finding difficult to fill—Forward-Deployed Engineers, MLOps Engineers, and AI Governance Specialists—barely existed as job categories five years ago. Both companies are actively developing AI infrastructure.
Similar to this, AWS and Microsoft are looking for DevOps engineers that can incorporate large-scale AI models into business operations that were planned long before any of this was a possibility. It’s possible that hiring and restructuring are taking place concurrently in the same organizations, but with separate teams and skill sets. From the outside, that image appears confused, yet it’s a real one.
Healthcare, on the other hand, has a more straightforward explanation. The workforce has been under pressure for years due to an aging population, and the shortage of medical imaging technicians, registered nurses, and healthcare data analysts has gotten so bad that hospitals are vying fiercely for applicants, much like tech companies do for engineers.
Ironically, one of the industries least affected by AI automation is healthcare; the scarcity there is essentially human, stemming from years of inadequate financing for training pipelines and a fast-paced workforce. Speaking with anyone who oversees staffing in a large health network gives the impression that this issue will worsen before a structural solution is found.
Because the nature of the skills gap has changed, it is worthwhile to examine it closely. Employers complained about candidates’ inadequate coding skills a few years ago. People comprehend AI intellectually, but they are unable to apply it in complicated, real-world settings. This is a more specific and challenging problem. When you witness a team spend six months attempting to incorporate a language model into a financial compliance workflow and consistently making mistakes, prompt engineering may seem like a soft skill.
Another sector where supply and demand have completely diverged is cloud security, where cybersecurity engineers with knowledge of identity access controls are engaged in bidding wars across industries that would have appeared unthinkable ten years ago.
How soon these holes will be filled by the training ecosystem, which includes business programs, boot camps, and colleges, is still unknown. Instead of relying for the market to supply what they require, some businesses are beginning to construct internal pipelines. Others are paying more than the going rate for a dwindling number of individuals who already possess the necessary set of abilities. Both strategies are difficult to scale. There are jobs available. Where people will come from is an issue that has not yet been properly addressed.
