Lattice 002: Accelerants

This is Issue Two of a series covering startup ecosystems and the most impactful innovation acceleration models.

I2 goes deep on accelerants: how orgs like EWOR work, the impact of national innovation agencies like UKRI, and how does it all fit together into the global matrix of progress. This is a primer for 3 separate Deep Dives into micro / meso / macro ecosystem accelerant layers.

Happy reading.

I once wondered what could be conceivably called the oldest progress-focused lab in history, defined as an organised, deliberate effort to produce useful invention.

The first organised R&D lab is often credited to Thomas Edison. The team at Menlo Park’s “Invention Lab“ were brought together to systematically produce technological innovation with commercial potential. You could reasonably call it the precursor to modern R&D clusters.

Still, that’s barely a century and a half ago.

The earliest ”labs” were probably workshops akin to the one run by Pythagoras in Ancient Greece times. You could hardly call it a modern laboratory, but these workshops were the testbed for key experiments and breakthroughs.

What of research institutions, then? There are some pretty old universities in the world. The oldest in continuous operation is the University of al-Qarawiyyin in Fez, Morocco. As often happened, it began as a mosque and gradually took on a larger role as a hub for learning.

Are we counting the Chinese Civil Service Exams into this, with their impact on governance? Roman roads? The systematic proliferation of railroad-driven commerce?

I digress from the basic question, of course. There are two reasons for this little exercise of imagination:

1. Progress Accelerants go way beyond startup accelerators and research labs. If you’re going very broad and focusing on the maximum impact, the most powerful accelerator might end up being a well-staffed school or even agriculture programs. Cascade effects extended over 20-30 years are a big deal, and equally hard to detect.
2. The Modern Innovation Complex is a very recent development. There have always been pockets of scientific activity and invention. Our world stands on their shoulders. That said, the differences in scale and speed are astonishing. We went from very little GDP per capita growth for 800 years (1000-1800) to 10x-ing it globally in 2 centuries. The pace of invention also went from 10-15 “substantial“ innovations per century to thousands every single year.

That by itself solves the Lattice question from Issue One — what is the most effective large-scale model to scaling innovation. Such a technological acceleration mesh already exists — in the form of the decentralised engine of progress that is the modern world.

Thousands of institute-attached labs, corporate R&D, national innovation. Schools, fellowships, apprenticeships. Trainings, grants, seed funding.

The modern innovation machine is colossal by historical standards. And it works through casting a net the size of the entire planet.

Think of it as 10,000 concurrent operations, each focused on its own pocket of the innovation space. Breakthroughs in materials science combined with AI development might end up producing new, lighter metals — enabling cheaper orbital manufacturing fabs for zero-g pharmaceutical production.

It’s a lot harder for civilisation-level bottlenecks to persist when there are millions of people solving all kinds of problems.

It is also on some level unknowable which keys open which locks. Penicillin is the famous example here — Fleming wasn’t looking for antibiotics. The concept didn’t even exist.

In short, it is good to cast a very wide net. The more people there are involved in the process of innovation, looking into an ever-larger number of directions, the better.

The Alchemy of Progress

The above principle works well in theory. Suppose, however, you are a researcher, founder, perhaps a Minister of Innovation.

Where do you look? What do you focus on?

Exact science, none of this is not.

Below is a snapshot of key areas of research + the types of outcomes & patterns it generally looks at. This is for reference — factors that affect innovation the most are leverage points which all such organisations are built on. The factor defines the model. Or sometimes vice versa, I suppose.

Aggregated based on looking at a lot of Google Scholar & Perplexity Deep Research. For now I’m less interested in precise interventions and more in what sorts of models — accelerators, VC, fellowships, etc — can do what.

If we look only at ecosystem scale, all innovation support models can be split up into 3 layers:

1. Micro (Single city)
2. Meso (National & Regional)¹
3. Macro (Global)

Folding that together with startup ecosystem factors, we get a matrix covering some 95% of programs and organisations.

This table above is what the rest of this series will be digging into.

Now, the majority of programs blend several or even all of these factors. Public programs tend to be built around specific KPIs and gaps, while accelerators or fellowships are more thesis- or selection-driven..

Two examples.

UKRI, the UK Government’s Innovation support arm, runs a comprehensive suite of programs that includes research grants, funding for lab equipment, skill development initiatives and more. Many countries run similar models, each with its own quirks, like Enterprise Singapore (Or perhaps A*STAR is a closer match here) or NSF in the US.

These programs don’t “create“ startups. Instead, they exist to give the ecosystem itself a boost — through talent, infrastructure, funding, and even just a vote of confidence. And they do this at scale: UKRI allocated ≈1.6bn GBP in R&D grants, ≈700m GBP in talent development investments, and much more across their spectrum of initiatives.

In short, this is national-level support designed to speed up the entire machinery of innovation.

An organisation like EWOR by contrast is entirely selection-driven. If you have potential, a burning desire to solve a problem, and are among the most promising in your field — you’re in. If you can be among the 35 out the 35,000 applicants, that is.

They, too, have been immensely successful. EWOR Founders often go on to raise 1-10M rounds post-program. Artifact AI, Thaleron (The largest pre-seed in Europe at the time) and Tillhub are just some of the high-profile alumni. Really, it’s a moonshot factory: the boldest dreamers backed by mentors who’ve made it before, one of the best networks in the business, and EWOR’s brand behind them.

EWOR is something of a unique case, and while each has its own methods, many accelerators like Antler, Techstars or indeed YC are fundamentally built on the principle of selecting for the highest potential. They differ in just how selective they are (EWOR or YC are single programs VS multiple-branch accelerators like Antler), but the methods are in the same ballpark.

And so, to simplify. Lift an entire ecosystem to raise its innovation output (UKRI), or find the most promising talent already ready to build (EWOR).

Which has the larger impact?

It’s a somewhat rhetorical question. No one program or organisation covers everything. You want to both raise the overall output and speed up the most promising talent and companies, among others. But what mix is the most effective? Where are the gaps? Just how exactly do all these initiative synergise?

I ask these questions from personal experience. Besides co-founding START London, I often ask myself what is missing. A space for builders to collaborate? Something to enable everyone to innovate in small ways? Programs that zero-in on a specific tracks like automated manufacturing or drug discovery innovation?

There just isn’t a straightforward answer. And I want to be clear — you don’t need an answer to begin having an impact. The world is a positive-sum game.

On a global scale, the more the merrier.

A Thousand and One Founders

To answer all of this, I built a database of over 100+ programs, organisations, and historically successful cases like Bell Labs.

The next three issues will go layer by layer, from targeted programs serving a single city to regional initiatives and agencies with global reach.

Frankly, this is a tough problem to crack. What I am really trying to accomplish here is to create a synergy model: a foundry-of-foundries. A type of org or foundation that itself produces different kinds of ecosystem-level accelerants.

The endgame is to build a series of initiatives targeting early-stage founders or people looking to become founders. All functioning in synergy.

Until next one.

Footnotes

1. Might seem worth to split the Meso one up, but the model won’t be very different in practice. The layers are porous. The only substantial difference would probably be pan-regional programs agreed upon by countries or supranational bodies like the EU. Regional accelerators are basically just a larger version of a country-based one with added structures & complexity. I’m perhaps playing a little loose with the boundaries, but otherwise you can slice and dice it ‘till the sun expires.