An open letter to those who are about to start a technology accelerator.



The accelerator business is about hastening the arrival of the future.


The goal of accelerators is not to make money, to increase survival rates for technology businesses, or to build communities.  Those are outcomes of successful accelerators.  The goal is to hasten the arrival of the future.


Before you start an accelerator, you need two things-

1.     a concrete, deterministic, optimistic view of what the future will be.

2.      assembled resources you will deploy to help hasten the arrival of that future.  Those can be people, products, knowledge and money, but a great accelerator will have all these. 

Few accelerators ever bother with the first step of this process- and if you were to ask them why they don’t have this vision, the answer will be one of two things- either they “leave the future up to their founders” or “they are looking for the best possible outcomes for LPs”.  Both of these answers have profound flaws in logic.


World-class technology companies have a deterministic view of the future- we as founders have dedicated our lives to creating a specific outcome in the future that we believe will arrive, and nothing can stop us.  The accelerator model so often takes the opposite approach- we don’t know which one of these visions of the future will come true, so we’re going to minimize risk by diversification, and in some heinous examples, hedging. 


You are hastening founders into the future, but first you must select which founders to hasten.  This part of the job is a playbook borrowed from the VC world, as most accelerators want to be able to then hand off the companies into the arms of waiting VCs, so there is a strong incentive to use the same process.  The problem here, however, is that a great VC might do 5 deals a year- you are doing 25, maybe more.  At 25 deals a year, the degree of insights you can draw from founders, coupled with the fact that you must do 10 or more of these deals simultaneously, means that you cannot take the time to understand the future of every founder.  You must have a point of view on what the future will be, and then you must select for people and companies who you can hasten to that future.


Being focused solely on returning value to your LPs is not a good reason for acceleration, because we’ve already spent 80 years attempting to create a model that works well for that exact purpose in the venture world, and things are not looking great. Those that have figured it out don’t spend a whole lot of time talking about value back to their LPs, they spend their time talking about the future.  Furthermore, lacking funds for pro-rata is a rotten place to be working from if you believe that venture is really based on a power-law of returns.



Why do you, poor misinformed person, THINK you are starting an accelerator?


If your goal is to create more startups you are playing a numbers game in an indeterminate world.  This is the approach of post-secondary accelerators, accelerators based entirely on government funding for innovation, and the “farmteam” type of accelerator whose job is really focused on deal-flow within a supply-constrained market for companies.  If this is your goal, then stop charging people.


If your goal is to educate then you should be publishing!  Paul Graham’s blog, I would argue, will create more value in the world than Paul Graham’s accelerator has.  Those essays are the Hammurabic Code of startups (which explains why the website looks like archeologists dug it from the ground). If this is your goal, then stop charging people.


If you goals is to get companies funded then your model should put your money where your mouth is.  From here on in, you take no equity- all you have is an option on 25% of the next round of financing at a capped valuation.  Success-based pricing and a firm structure!


If your goal is to make money for yourself and buy a vineyard then you cannot start here.  You will fail because you are not fundamentally aligned with your founders, and because the model is not proven to return.


If your goal is to hasten the arrival of the future, then read on, I have some ideas for you.


One of the most profound experiences of my life was the first time I heard Peter Thiel speak concerning his theory of Determinant Optimism. Watch this talk.


The deterministic nature of your bet on the future needs to be the central tenant of your accelerator.  You must know what the future will be, and invest your time and money in a group of individuals who are building companies that will make that future arrive.   You must also be able to say no, to scream no, at things that are just going to make money.  This makes me sound like a jaded, hippie I know, but the success of the community, and the success of the program that that community is built on, is reliant on creating a group of people who have a common vision of what the future will be.  THAT is the genius of YC, because in many cases, the idea that is accepted does not find its anchor in the vision for the future, but in time, the company comes to understand why their product, their business, fits into a larger vision for what 2020, 2050 will be.


Go get ‘em. 




(to Ian, Zoe, Semyon, Ty, Eveline, thank you.  You made me who I am.)