A key observation about the "deep tech" benchmarks versus all the other categories of startup is that they are fundamentally subjective. To make that situation worse, there is frequently a dearth of qualified outside expertise that an investor can lean on to evaluate the startup. This makes deep tech a very different kind of fundraising process than the other types of startups.
Tangential to that, it is nearly impossible to raise money as a deep tech company in software specifically. Some of the reasoning is valid (where will the revenue come from?) but it also greatly stunts development of novel and advanced capabilities in software that require substantial investment.
I tend to agree with you. Even though there are specialist funds in deep tech, the available capital is less than traditional SaaS. The cycles are longer in deep tech, and require a bit more 'patient capital.
That said, deep tech startups know to access a range of subsidies and grants that are equity free in a way that more traditional startups don't.
There are definitely more diverse funding sources available for deep tech but mostly if you limit "deep tech" to more traditional physical domains like hardware, biotech, medtech, etc. For domains like computer science, the non-traditional deep tech funding sources tend to be few outside US intelligence agencies.
That said, having raised capital for software deep tech, I think the best path forward today is to position it as a SaaS startup. The execution strategy is quite different than a traditional SaaS startup and requires significantly more front-loaded personal investment on the part of the founders, but I think it is a lower hurdle to clear in terms of effectively raising investment than positioning as deep tech.
My own view is that as an industry we are currently quite poor at fostering the formation of software deep tech startups (which is an opportunity). This is both because they are difficult, unusual startups from an investor perspective and there are very few entrepreneurs with experience successfully executing this type of startup. Thousands of articles exist on executing and investing in SaaS or consumer tech companies, and approximately none for software deep tech.
For biotherapeutics (the largest subsector of deeptech in terms of VC funding, receiving $17B in 2018, and also the most idiosyncratic), in general the milestones are:
Series A:
* lead product with robust data in gold standard animal models + strong additional supporting data OR
* validated, novel biology platform with a team of experienced drug hunters
* path to first in class or best in class product in market with significant unmet clinical need
* good IP including composition of matter (or experienced team of drug hunters who knows this is important, if too early for having this IP)
* either a lead development candidate or a clear path to one (can be earlier stage if you have a really good biology platform, proven team, and tractable chemistry)
* path to human POC on venture dollars (doesn't have to be on the Series A, but ideally on Series B, though later can be acceptable)
* good team
Series B:
* same as above, plus: should be able to get through early human studies on this round
Seed is pretty variable but generally you need a plan to achieve Series A readiness with your seed money
Of course this is a broad generalization -- some startups license very advanced drugs from pharma and use the Series A to fund a clinical study in a new indication
Note that pretty much all of these milestones relate to products, not platforms. A platform is only worth as much as the products it generates. In general Series A investment in platform companies are focused on generating the first products and getting them into preclinical development
Tangential to that, it is nearly impossible to raise money as a deep tech company in software specifically. Some of the reasoning is valid (where will the revenue come from?) but it also greatly stunts development of novel and advanced capabilities in software that require substantial investment.