Accelerators, whether YC, Plug in Play, or Techstars, have always been the golden key to startups in the 21st century. They provided everything from funding to guidance to help a startup scale their product to millions of customers.
However, in this new age, things are beginning to change.
Accelerators in the past
Accelerators worked in the 2000s because that was the age of B2C companies following the initial expansion and development of internet and computing hard tech in the 1980s. Companies like Microsoft, Facebook, or Google thrived during the 2000s as the internet slowly began to trickle down from a professional equipment to a general consumer product.
Accelerators also worked in the 2010s because that was the decade defined by B2B companies. B2B Saas startups thrived because as software became more sophisticated, new CRM or marketing platforms can be built to optimize business operations and increase efficiency.
It is important to observe that general B2C companies saw a decline during this age as the market for consumer softwares became generally monopolized by a few major players: Meta(Facebook), Snap, and Amazon. Even when a new B2C startup through a lucky hand catches hype such as BeReal, they only see a very brief period of success before eventually declining or being bought by one of those larger players.
Deeptech
Now in the 2020s, a new pattern is beginning to emerge. As B2B competition becomes fiercer as the market is increasingly monopolized, investors are becoming more reluctant by the day to fund these companies. After all, there are only so much demand for commercial efficiency and so much technology available to startup founders. However, as successful B2B startups become more scarce while B2C startups are buried in the water, a new type of successful startup is beginning to emerge: deeptech.
Deep tech or hard tech startups are those that focus on developing cutting edge technology to solves some of the greatest problems the world faces. Some examples are SpaceX, Cerebras, and Lightmatter. They have a few general characteristics: they tend to be hardware, they take a long time and lots of funding to develop, and they create new markets. In other words, the risk in these startups are not that they can't find product-market fit, but instead whether the product can be developed.
I was recently at a pitch competition, and this new wave of deeptech was definitely apparent in the amount of medical startups that were pitching and successful. Furthermore, even accelerators like Y Combinator quoted themselves that they are 10X more likely to accept a deeptech company regardless of their progress or feasibility.
This then begs the question of how useful accelerators are in this modern age. While in the past they are often there to guide founders on how to find product-market fit, this new wave of startups no longer needs to find product market fit. What they need to focus on is developing the product and making sure it works! While there are some industry specific accelerators that could potentially help by providing equipments or labs, I still can't imagine them being too helpful through the analogy of "if they could develop it they would have done it."
The future?
So, what is it for accelerators then. Could this be the end for accelerators? Would they die down in scale? Would they be forced to adapt/
I personally see this in two ways. First, accelerators can definitely adapt to become more domain specific. In fact, there are already certain medical startup accelerators being developed.
In a more important way, while the deeptech age has come to define this new decade, it is not here to stay! Just like how the 1980s saw a focus on developing general hardware technologies, the 2020s is focused on developing this new age of technologies that would be eventually, like the 2000s, translated into the B2C field.
In fact, this can already be seen in Artificial Intelligence in many ways. While the 2020s is seeing new papers uploaded to arXiv every single day, the AI available to consumers are either very outdated or nonexistent. This is due to a few problems: first, prompt engineering and AI alignment remains as a big issue hindering AI from moving into the consumer field. Also, computing power also remains as a problem, as GPUs are both very inefficient and costly. Finally, just like the internet, it will take artificial intelligence some time to move away from the 'sci-fi robot' technology to an everyday tool.
Once these issues are solved, I am confident that these new technologies would see an explosion in consumer applications as it begins to replace everything we know of physically or digitally- just like the internet.
These are just some of my general thoughts.