The Era of 10x Software Innovation is Ending, But Deep Tech Offers New Frontiers
Software has been the darling of VC for decades. Since the dot-com bubble, consumer internet, enterprise SaaS, fintech and other software verticals have offered the tantalizing mix of scalability, network effects, and disruption that can produce 10x or even 100x returns. However, as software eats more of the world, the opportunities for transformational innovation decline. Most industries have been digitized in some capacity, making it difficult for startups to deliver an order of magnitude improvement over incumbents. The low-hanging software innovation opportunities have already been picked.
We can see this in the maturing of sectors like CRM, HRIS, and payments which have clear category leaders and adoption inertia. Compare the software company founded today with similar core technology to a predecessor from 10 or 20 years ago — the prior likely captured outsized economics while the newcomer struggles with point solutions for diminishing problems.
The Inevitable Slowing of Software Innovation
Does this mean the days of high startup returns are behind us? Perhaps for consumer/SaaS software, but deep tech is emerging with environments ripe for 10x innovation. By deep tech, I mean startups focused on highly technical solutions requiring significant R&D cycles before productization across areas like space, robotics, energy, and cutting-edge hardware.
Before diving deeper into why deep tech looks attractive, it’s worth elaborating on why incrementalism prevails in software. There are four core drivers of this trend:
1. Prior Art & Abstracted Layers — Virtually every business or consumer software problem has been partially solved. New entrants build on existing abstractions provided by open source communities and cloud vendors. This raises the innovation baseline but makes 10x improvements less likely as most of the low-hanging fruit around core functionality is gone. The competitive moat for a software startup today often comes down to UI/UX improvements, pricing, or verticalization — all very incremental areas.
2. Peak Smartphone — The release and fast adoption of iPhone & Android triggered a Cambrian explosion in software innovation. However, smartphone technology and use cases are maturing, limiting the opportunity for novel software with high attach rates among consumers. Apps today fight for declining marginal attention and dollars.
3. Maturing Cloud Infrastructure— The scalability, flexibility, and cost structure of cloud computing removed virtually all infrastructure advantages. Once again, essential functionality has been abstracted away — cloud itself is the ultimate incrementalist horizontal. Emerging areas like serverless and edge computing will provide less impactful base layers relative to their predecessors.
4. Increasing Abundance of Capital — Finally, record levels of capital availability in the VC asset class increased competition. More money inevitably leads to lower relative returns across a sector as less efficient players receive funding. We see these signals today in the form of multi-faceted dilution, delayed or avoided exits, and increased reliance on follow-on financing to grow into valuations.
In summary, the SaaS business model will continue driving economic growth and productivity for enterprises, but the opportunity for build-to-flip or build-to-IPO with software has structurally declined over the past decade. This trend will persist into the 2020s. Deep tech offers the greenfield frontiers for outsized startup returns.
Deep Tech Capital Efficiency Rising as Costs Fall
Deep tech is conventionally saddled with the “too much time, money, and risk” label among startup investors. Hardware does take longer to build and commercialize relative to software. However, new trends are driving greater capital efficiency including:
Agile Hardware Development
Hardware startups historically faced lengthy, expensive R&D cycles along with high costs for custom components, manufacturing tools, and minimal viable product runs. Today, the combination of easier 3D printing, modular design approaches, and accessible contract manufacturing enable rapid prototyping iterations. Instead of needing to raise and burn millions before having an alpha product, deep tech startups can leverage agile development with regular customer feedback checkpoints. This means less waste, quicker learning, and lower costs to product-market fit.
Offshore Supply Chain & Manufacturing Flexibility
Increased access to overseas electronics components through export/import marketplaces combined with overseas contract manufacturers able to handle lower volume runs has reduced hardware costs. Startups are less dependent on domestic suppliers or factories requiring large minimum order quantities. With global shipping infrastructure, designers can tap into China’s electronics ecosystems alongside Indian assembly labor to manage BOMs and scale production far more capital efficiently.
Specialized Engineering Talent
The growth of electrical, mechanical, materials and software engineering programs has increased the talent pool for technical co-founders. This makes it easier for deep tech founders to pull together multi-disciplinary teams with specialized skills needed to architect complex systems. Abundant PhDs graduating from leading research labs also create valuable hiring funnels in specific domains like robotics, aerospace, or neurotechnology.
Secondary Markets
Finally, the rise of online marketplaces for private company shares provide early liquidity opportunities and value signaling. This makes recruiting and retaining talent easier for deep tech startups if employees have a path to liquidity outside of traditional exit outcomes. Secondaries also influence primary round valuations in a more data-driven fashion to reduce dilution as deep tech startups scale.
These tailwinds combine with steady decreases in compute, sensor, robotics, synthetic biology, and space launch costs to make building deep tech companies cheaper than ever. Unit economics can work even serving niche customers. Startups need less funding today to reach initial traction milestones compared to predecessor deep tech companies. Exit time horizons are compressing as well.
On the flip side, the perpetual access to cheap capital created zombie software startups kept alive by follow-on rounds despite stalled growth. Deep tech is paradoxically becoming more efficient than software! This brings us to the key reason VCs should be looking closely at deep tech today: massive greenfield market opportunities.
Where is Non-Incremental Innovation Still Possible?
Deep tech serves markets where no digitized solution previously existed or capacity remained capped by physical limitations. Now exponential progress in underlying hardware capabilities allows startups to address these greenfields across sectors like:
Industrials:
Robotics — new robotic capabilities (computer vision, manipulation) and declining costs are driving automation into manual / hazardous jobs in warehouses, agriculture, construction, factories etc. Startups like Bright Machines, Covariant, and Intrinsic are leading the way.
Space — launch costs dropping 10x in last decade enables satellite constellations for telecoms, earth imagery/analytics, space transport, orbital infrastructure and more. See SpaceX, Planet Labs, Orbital Sidekick.
EVTOLs — electric vertical takeoff aircrafts will expand air transport capacity. Joby Aviation and Archer are emerging leaders.
Climate Tech
Grid-scale energy storage — new battery chemistries and control systems needed to scale renewable electricity. Form Energy, ESS, EnerVenue innovating here.
Carbon capture & sequestration — removing legacy CO2 emissions from atmosphere is paramount to climate goals. ClimeWorks, Heirloom, Planetary Hydrogen critical to supply chain.
AgTech — emerging biotech and autonomous farm equipment to boost crop yields amid climatic changes and population growth. See Indigo Ag, Bear Flag Robotics, Carbon Robotics.
Frontier Tech
Quantum computing — unlocking new tier of computational power to advance material science, drug discovery, finance modeling and more. Leaders include Rigetti and IonQ.
Brain-computer interfaces — moving beyond keyboards and touchscreens towards seamless man-machine interaction. Neuralink, Paradromics, and NextMind show massive potential.
I’ve highlighted just a sample of remarkable deep tech startups above. These companies exploit advances in material sciences, AI/ML systems, robotics, networks and sensors to push practical limits across computational power, environmental engineering, transportation methods, agriculture, communications infrastructure and human-computer interaction.
The business models work — often direct sales or contracting with governments and large enterprises eager to implement cutting-edge solutions. With diligent execution and go-to-market, deep tech represents the new frontier where software-style venture returns remain achievable over 5 to 10 year durations. The critical ingredients for outsized outcomes are present:
Massive TAMs
- High defensibility moats based on technical complexity and patents
- Large incumbent partners looking to adopt innovation
- Robust secular tailwinds around sustainability and technology permeation
For these reasons, I believe the broader VC asset class will continue shifting a larger percentage of capital towards deep tech startups through the 2020s. The erosion of software’s transformational promise dictates this migration into new frontiers. Disciplined lifecycle management and portfolio construction will remain critical given higher variance, but the next 10X giants will emerge from domains like quantum, robotics and commercial space — not another messaging app. Deep tech represents the future of venture.
References
[1] Snel, Jordi. “An LP’s Guide to Venture Funds — Part I.” Jordi Snel, 27 June 2022, https://jordsnel.substack.com/p/an-lps-guide-to-venture-funds-part.
[2] Mandel, Josh. “Four Theses on Deep Tech.” Josh Mandel, 1 Aug. 2022, https://mandel.substack.com/p/four-theses-on-deep-tech?utm_source=profile&utm_medium=reader2.
[3] “Manifesto.” Founders Fund, https://foundersfund.com/2017/01/manifesto/#/introduction. Accessed 12 Dec. 2023.
[4] Xu, Koko. “Deep Dive into Deep Tech.” Notion, https://kokoxu.notion.site/kokoxu/Deep-Dive-into-Deep-Tech-0ad0b64ee5864812af7acdad4856ccdc. Accessed 12 Dec 2023.
[5] McGuire, Shaun. “Venture Capital Model Is Broken, Not Because Too Much Money Raised, but Because It’s Too Diffuse Now…” Twitter, 12 Feb. 2023, https://twitter.com/shaunmmaguire/status/1651616138323988482.
