Crossed Wires: Venture capital â where is all the money going?

This story has significance for readers across Africa and beyond.

I originally set out to research and write an article about the venture capital (VC) industry, and where it has been investing and whether any trends were emerging. VC is not some arcane field for financial specialists; it is the primary bellwether of most of the stuff all of us end up interacting with a few years down the road. Apple, Uber, Airbnb, Waze, X, Facebook, Google, Amazon and their cohort. A very long list of ubiquitous participants in our lives.

The big winner in attracting risk capital is artificial intelligence (AI) â no surprise there. But I want to dispense with it quickly because something else has emerged as I was doing this research, which we will get to.

Out of the approximately $500-billion doled out by the major VC companies, 75% went to AI, and nearly half of it went to Anthropic and OpenAI. Two companies gobbled up nearly as much as all the early-stage VC-funded startups on the planet. A much more interesting matter is who picked up the crumbs that were left over. They are a harbinger of a massive change of destinations for risk capital, and its inevitable downstream effects on everyone on the planet.

There is a backdrop to this narrative. Venture capitalist Marc Andreessen wrote an essay in the August 20, 2011 edition of The Wall Street Journal titled Why Software is Eating the World. To say that it went viral is an understatement. It stands as possibly one of the most prescient and widely read technology predictions of its time. Andreessen nailed it â software, and particularly Software-as-a-Service (SaaS), was going to rampage through every industry on Earth. Which it duly did.

But then AI arrived (which is also software, albeit optimised by running on top of specialised chips called GPUs). Very quickly it became evident that AI software could write other software (Claude Code, OpenAI Codex, Cursor and others). And it can now do the job better than humans on almost all axes of software development (the new-frontier AI models are now writing their own next versions).

The consequence? Software is no longer that interesting. Even AI software is becoming a cheap commodity. A utility. A reliable and very clever substrate for everything on top of which everything else is being built â cars, fridges, mining equipment, medical devices, manufacturing plants, TV sets. Software is not eating the world any more. It is now feeding the world.

And so back to the story, who is picking up the overflow after the AI feeding trough is empty?

The answer is stuff. Things you can touch, hold, see. The physical world. Here are the sectors that are seeing VC love (if not lust), and a romance that has just started and is likely to deepen.

Defence technology

The most remarkable engagement is between risk capital and the military-defence complex.

Not long ago, defence technology occupied a peculiar moral no-manâs-land inside Silicon Valley. Venture capitalists happily funded companies that disrupted taxi drivers, newspapers and hotel owners, but anything connected to military applications tended to provoke uncomfortable conversations (including a walkout of engineers at Google in 2018 as a protest against Google software being provided to defence-aligned Project Maven). Many venture funds explicitly excluded defence altogether. It was easier to raise money by promising to optimise grocery deliveries than by improving missile guidance.

Russiaâs invasion of Ukraine demonstrated how quickly warfare was becoming software-defined and AI-fuelled. Cheap drones destroyed expensive tanks. Computer vision became a battlefield weapon. Autonomous systems moved from speculative research projects to military necessities. At the same time, escalating tensions between the United States and China, and multiple belligerents in the Middle East have raised the stakes alarmingly.

As the world has recently become more dangerous, governments have redirected their budgets and the venture industry has adjusted (somewhat gleefully) with remarkable speed. VC giant Andreessen Horowitzâs defence investments are proudly housed in a practice called American Dynamism, making its position loud and clear.

So defence technology has become one of venture capitalâs fastest-growing sectors outside pure AI itself (180% up year on year in the US). Investors are pouring money into autonomous drones, battlefield intelligence, satellite analytics, cyber defence, secure communications and autonomous logistics. European venture firms, once especially reluctant to finance military technologies, have reversed course, fuelled by the US administrationâs irate departure from its traditional role as military sugar daddy.

US investors have scarcely bothered to disguise their enthusiasm.

The ideological reversal is extraordinary. Silicon Valley spent years insisting that technology should connect people rather than divide them. Now many of its brightest entrepreneurs are building machines whose primary purpose is to deter, disable or destroy adversaries more efficiently than before. Nobody seems particularly embarrassed by the transition.

Robotics

Robotics tells a similar story. For decades, robotics occupied a curious place in technology. Every few years somebody produced a machine capable of dancing, climbing stairs or pouring coffee, accompanied by predictions that the robot revolution had finally arrived. It rarely did. The hardware remained expensive, the software brittle and commercial applications frustratingly limited.

The commodification and accelerating capabilities of AI have changed that equation almost overnight.

Machines can now recognise unfamiliar environments, adapt to unexpected obstacles and learn new tasks with far less human intervention. Warehouses, factories, ports and mines suddenly look like fertile territory for automation. Investors who once backed enterprise software helping office workers schedule meetings are increasingly financing robots designed to replace those workers altogether.

The data? Risk capital investment into âphysical automationâ has increased by 205%.

The movement was visible across most of the harder edges of technology over the past year. US semiconductor start-up financing rose 85%, space-tech funding climbed roughly 48%, and investment in specialist AI-fuelled drug-discovery labs was nearly four times its 2023 level. Quantum startup investment has risen more than sixfold.

And the losers? SaaS, crypto startups, fintech startups, social media and other digital-only companies.

Escaping the screen

So I go back to my original thesis. Viewed collectively, these supposedly secondary growth sectors reveal something more profound than the AI boom. Venture capital is quickly moving on from the mirage of apps living inside smartphones. Investors are looking to escape the screen.

It is moving into factories, satellites, drones, power stations, laboratories, autonomous vehicles, construction sites, warehouses, restaurants and targeting systems. The winning companies are those combining clever software with expensive physical assets, difficult engineering challenges and deep technical expertise that competitors cannot reproduce over a weekend.

That may ultimately prove to be the lasting legacy of the AI revolution. AI has become so good at producing code that code alone no longer commands the premium it once did. As software becomes abundant, boring and ubiquitous, scarcity reasserts itself elsewhere.

And so the short answer to the question that opened this column â the money is starting to go to things made of atoms, just as it did before the information revolution. DM

(Sector VC growth numbers were aggregated from PitchBook, Crunchbase, FT, Bloomberg, MarketWatch, McKinsey and others, provided by ChatGPT, sources checked.)

Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg, a partner at Bridge Capital and a columnist-at-large at Daily Maverick. His new book, Itâs Mine: How the Crypto Industry is Redefining Ownership, is published by Maverick451 in South Africa and Legend Times Group in the UK/EU, available now.

Reporting originally appeared via Daily Maverick. Read the full source for additional context.