Another counterexample -- Drudge Report.
Main Point: A carefully conceived and planned sole, solo founder information technology startup, e.g., a Web site, can be cheaper to start than even a common Main Street business, e.g., a pizza carryout, even a grass mowing service. But if the site becomes popular, then quickly it will generate more cash per year than in a seed round. From there the site might continue to be a significant business, e.g., Plenty of Fish, Drudge Report, and likely some more, still without equity funding.
Conclusion: The Web is one heck of an opportunity for startups without equity funding.
There is a common assumption in your argument, that the startup must have a period of wildly rapid growth to take the opportunity or lose out. Sometimes this is true, but as in my counterexamples and, really, just common sense, it does not have to be true and isn't always true.
Part of this argument about rapid growth and need for equity funding is that once a startup shows promise, it will be copied by others, and only a startup that executes well with lots of equity funding can win. This situation need not always be true: Instead, some startups can have crucial, core, powerful, valuable, proprietary, protected technology, secret sauce, locked up in a secure server farm, and difficult to duplicate or equal. That secret sauce can provide a powerful barrier to entry, "Buffett moat".
For my startup, I am a sole, solo founder. I've done all the work, have the software apparently running as intended, and am rushing to go live on the Internet.
From the timings of my software and some Mary Meeker data on ad revenue, the one server I plugged together, for less than $2000, if users like the site well enough to keep the server busy, should generate top line revenue and, essentially the same, pre-tax earnings of about $250,000 a month. I don't need the equity funding now, and with $250,000 a month in revenue, I will not need and would not accept equity funding.
My Web site is intended to be of high interest to essentially everyone in the world with a standard Web browser.
If my site gets to $250,000 a month in revenue, then that will be a significant sample of what users like from which we could make a projection to rapid growth to about $10 million a month, all just from my spare bedrooms and present Internet bandwidth. With such revenue per server dollar, the costs of the servers, etc. are next to trivial.
Thus, from pleasing a large fraction of everyone on the Internet, we could reasonably extrapolate the potential of my startup to a company worth $T+.
We have to keep in mind: Successful information technology startups are unusual things, each one essentially a unique business. So, we have to expect, at least not reject, some unique aspects.
Moreover, we have to see that the most successful startups were quite different from anything that went before. So being new and different, even radically different, are not sufficient for success but are nearly necessary -- again, we must not reject things just because they are new, different, unique.
Some products and services have such coveted utility that customers, users are really eager. My startup is so intended. Uh, my startup is squeaky clean, fully "safe for work", actually, say, culturally uplifting, an aid to better lives, families, communities, countries, and civilization, nothing to be ashamed of.
While my startup is a Web site, really what's important about it is just some original applied math I derived. The math is difficult to duplicate or equal because it is clever and original and exploits some advanced pure math prerequisites. Users will just get the utility and not be aware of the math.
But it is the math that is the crucial core secret sauce, difficult to duplicate or equal. In comparison, except for the crucial core, the software is next to trivial. So the site is not really a computer software development effort but an applied math effort.
There are not many people who know the math prerequisites and can do good original applied math, and the people that can nearly never want to do such applied math, want nothing to do with software, and are not in the information technology startup community.
There is a way to have a good idea about the promise of my work -- based on the utility for all the people on the Internet with just a standard Web browser: While Silicon Valley does not use such means of startup evaluation, several very important parts of our civilization do, do so routinely and quite successfully, with batting average far, far above that of Sand Hill Road.
So some parts of our society can look at a serious problem, see clearly that the problem is important and the first good or much better defensible solution will be very valuable, and evaluate proposed solutions with high reliability. Such parts include some internal industrial R&D, Ph.D. committees, editors of the better STEM field peer reviewed journals of original research, NSF, NIH, ONR, DARPA, USAF Cambridge, NASA, major engineering firms, etc.
My work would easily pass such filtering. So, I'm confident.
In strong contrast, Sand Hill Road wants to see a minimum viable product with traction significant and growing rapidly; all the project evaluation techniques of Ph.D. committees, NSF, ... are ignored and set aside with severe determination.
Sand Hill Road is not interested in anything from original STEM field research, maybe is forbidden by their limited partners to consider such research, and is not practiced in evaluating such research.
But I hold a good Ph.D. in pure and applied math from one of the world's best research universities. All my career, many times, I've successfully applied math, sometimes original, to practical problems.
So, I have a solid basis on which to be optimistic. Sand Hill Road can't understand that.
Sand Hill Road wants nothing to do with me before my business is so good I will want nothing to do with them.
Yes, my startup is unique, and that's nearly necessary for major success.
It would be short sighted to reject what I'm saying merely because it is unique or even different.
On crushing it, that's not part of my intentions or planning. The $1T+ is.
I've seen a lot in computing, e.g., in an AI, artificial intelligence, project at IBM's Yorktown Heights Watson lab, for US national security, saving FedEx twice, teaching ugrad computer science at Georgetown University and graduate computer science at Ohio State University, and I've seen much more in pure and applied math.
Net, at this point, my considered option is (1) academic computer science has run out of promising research directions they have the qualifications to pursue and (2) Sand Hill Road style computing is essentially dead and all but buried. Routine software will continue, much as routine sheet steel stamping has continued, but for now the glory days are over. In particular, for AI, etc., I see 99 44/100% hype and nearly all the rest water or mostly quite old applied math. We just had an AI spring of hope then a summer of hype and are now in a fall of failure and soon again in an AI winter. E.g., to me self-driving cars in current traffic on current roads is obviously an absurdity and a lot of hype, a bubble about to burst.
Computing remains a major opportunity, but nearly none of the software development community, academic computer science, or Sand Hill Road know how to make use of the opportunity.
For exploiting the amazing cycles, bytes, bandwidth, infrastructure software, etc. some new directions are needed.
IMHO the most promising such new direction is engineering based on very carefully done theorems and proofs of pure/applied math.
The number of people in the world ready, willing, able, and eager to do this could have a convention in a $20 a night motel room.
So the situation is bad, but the flip side is a grand opportunity. E.g., the processor I bought is an AMD FX-8350, 64 bit addressing, 8 cores, 4.0 GHz standard clock speed, now available quantity one, retail for $100-. ECC, error correcting coding, main memory is available for $10-/GB. The situation on hard disks, solid state disks, and Internet bandwidth is maybe even more amazing. So, the need and the opportunity is to make good use of such products.
First step: To heck with Kernighan and Ritchie, Stroustrup, and even with Knuth. Instead get a copy of W. Rudin, Principles of Mathematical Analysis, Third Edition, dig in, and work nearly all the exercises.
E.g., again, once again, over again, yet again, one more time, US national security is going deep long ball, if you prefer, swinging for the fences, into challenging pure and applied math with the new Space Force. For the role of math there, trust me!
No, don't trust me: Instead watch the start of the movie about John Nash where the flat statement is that "mathematics won WWII". Exercise: Discuss with arguments pro and con. Don't forget that Commander Rochefort's work and advice to Admiral Nimitz is what sank four Japanese carriers at Midway and turned the tide of battle in WWII in the Pacific. Uh, the Japanese brought 6 carriers to Pearl Harbor. Soon they lost one at Coral Sea -- also due to Rochefort. Then they lost 4 more at Midway. So early in the war, they were already running short of aircraft carriers. And don't forget much of the reason Rommel lost in North Africa was because most of his supplies went to the bottom of the Mediterranean due, right to code breaking. And note the importance of the cavity magnetron from Maxwell's equations in winning the Battle of Britain. And there was much more, e.g., from A. Turing. Then we have to count The Bomb, e.g., the critical mass calculations.
If you regard these views as new or different, then just remember you first learned them here.