Mobile has made internet usage [and instant on-demand access to data] into a ubiquitous utility -- access to data is literally more prevalently available to the world's population than even electricity or indoor plumbing. Maybe not always, but it is often easier to either find wifi access or get a cell data signal than it is to find a bathroom or an outlet to recharge devices. Increasingly people do not even think about this -- they might not stream movies, but otherwise simply just count on having an acceptable data connection everywhere they go. It's not a luxury anymore, but rather it has become a need. Most people use a data connection for over almost all (i.e. well over 90%) of the tasks that require data or information for an informed choice in their work and daily lives. We are past the point of exuberant hype, past the point the trough of disillusionment ... we are not steadily, sustainably finding more enlightened implementations and just finding better ways to add productivity and take out costs and uncertainties.
Client-side front-end programming for cross-platform native mobile apps is rapidly maturing beyond the janky OS-specific toolchains that were needed for in the early pioneering years of iOS and Android app development ... and on the back-end, the next generation of server-based enterprise architectures for data resources, messaging, machine learning, security and other applications that run on commodity hardware in the cloud is also steadily getting more advanced, secure, robust, extensible and generally more mature. All of that means that cost of developing a working, web-scale proof-of-concept revenue-generating SaaS is so low that many startups are self-funded at the early stages. Tech funding by venture capitalists is increasingly only used for the final capital-intensive stages of technology-enabled blitzscaling or polishing the edges of a proven concept before IPO or acquisition.