Decoding a decade of AI investing through the lens of Seedcamp
Read on for Will Bennet's breakdown of the Seedcamp Investment Playbook in AI as well as David's taking of the pulse on European Venture.
As Europe's seed fund, Seedcamp requires no introduction, boasting a formidable legacy since 2007. The recent launch of Fund VI in May plants the seed 🌱 for even greater dedication to pioneering AI-driven solutions across diverse industries.
This is precisely why we're excited to share this insights article by William Bennett, Associate at Seedcamp. This piece is a great example of the value of being an angel LP in funds like Seedcamp. It offers unparalleled insights into emerging and potentially groundbreaking technologies.
Synthesia and Viz.ai, two recent AI investments by Seedcamp, are undoubtedly on the lips of every tech investor and AI enthusiast. But Seedcamp's first investments in AI date back to a time when AI was mostly confined to academia 🎓. Notable early investments include the likes of James Finance (a personal favorite, and yes, it's a Portuguese 🇵🇹 startup) and Elliptic, which recently raised a Series C to continue powering financial crime detection 🕵️.
We’ve all noticed that the rapid innovations in GenAI have accelerated the AI space, representing a remarkable 18% of all investment in the venture industry. But how does Seedcamp think about this space and what binds together the 70+ commitments the firm has made to date in AI and ML? Let’s explore! 👀
This is a condensed version of the original piece published on the Seedcamp blog.
6 micro-themes
Applications that ‘do work’
Enterprise workflow stack and automation
MLOps
DevOps and deployment
Guardrails
Sector-specific discovery engines
Applications that ‘do work’
Large language models mark a pivotal moment in AI, becoming the first tools capable of tasks traditionally reserved for humans 🤖. Sarah Tavel of Benchmark delves into this point of inflection, highlighting how software, once merely enhancing efficiency, now actively performs or generates work.
Seedcamp has invested in Metaview which writes interview notes, Flowrite which can write emails and messages across Google Chrome, and Synthesia which automates video production and recently achieved unicorn status.
The concept of software that does work is already catalysing a wave of startups and new business models and philosophies for the enterprise.
How do we redefine 'productivity' in a world where software can be infinitely reproduced? 🔄
Where 'time' once stood as the denominator in the traditional productivity equation [productivity = rate of output/rate of input], could the correct denominator now be the GPU's power?
In a future where humans might primarily validate model accuracy and precision, the real opportunity lies in finding the 'fast moving water,' as aptly coined by NfX. This is where AI is capable of the biggest volume of work.
Amidst this early wave of LLM applications, a clear pattern emerges. Companies are laser-focused on automating time-consuming, high-complexity activities conducted in written natural language. Procurement RFPs, compliance questionnaires, legal documents, and code generation stand out as the obvious early use cases. The companies that can seamlessly translate problem descriptions in these areas into sound and reliable output are poised to capture significant value. 💼🚀
Enterprise workflow stack and automation
In the past decade, the majority of enterprise AI applications have honed in on enhancing employee efficiency, with a lion's share dedicated to this pursuit. The standout innovations have manifested in horizontal software, addressing 'hard and boring' challenges like process mining, application automation, and project management. The goal? To empower employees with more time ⏰ to accomplish meaningful work.
UiPath (NYSE: PATH) has emerged as a category leader in the enterprise work stack since its public debut in 2021. Specializing in automating high-volume, low-complexity tasks within the enterprise, UiPath has solidified its position by acquiring the communications automation platform, re:infer, also backed by Seedcamp. Seedcamp’s investment portfolio extends into AI businesses transforming different facets of the enterprise, including Rossum, a document gateway for business communication, Juro, a contract collaboration platform, and tl;dv,which records meetings and helps you tag important moments on the fly.
Newer robotic process automation products and features increasingly include ‘self-compounding’ elements that use previous activities to prioritise future activities based on ROI and other key metrics 📈.
In the longer-run, there may be a significant difference between apps that ‘do work’ and more general efficiency-enhancing enterprise software. AI businesses in the former category may use models trained on sector specific data, which are necessarily vertical software (a co-pilot for every industry). In the latter category, the market opportunity includes all business with processes, for which broader approaches may be more relevant. Vertical or sector-specific winners can of course still be massive businesses, and may capture 60%+ market share whereas horizontal winners rarely get more than 15%.
MLOps
'MLOps' has evolved into a nebulous yet indispensable branch of the AI taxonomy, serving as the most fitting descriptor for companies facilitating the operation, maintenance 🛠️, analysis, and fine-tuning 🌌 of AI models.
MLOps has several layers that are each prerequisites for model development; the GPU layer, the data layer, the observability layer, etc. Companies typically have an insertion point that tackles one of these and move on to add features that target several at once where there are obvious adjacencies. MLOps companies serve as the 'picks and shovels,' essential tools for the AI gold rush. ⛏️💎
In Seedcamp’s portfolio, Embedd.it parses data from ML datasheets to generate drivers for semiconductors and Fluidstack aggregates spare capacity in datacenters to rent out to customers.
The rapid adoption of MLOps tools by developers underscores the possibility that demand is surpassing supply in this category. 🚀 For instance, within the data layer, there remains ample white space for AI-native tools dedicated to observability, integration, and transformation. The ongoing race to develop tools for debugging 🔍 and enhancing model performance is heating up 🔥, reflecting the fervor of innovation in this space.
Guardrails
Sustaining the MLOps ecosystem is a cluster of software dedicated to averting the general misuse of AI. If generative AI is an open-road race, envision the guardrails as the police, the red lines, the car bumpers, the rolling barrier, and the orange cones. 🚔🚧
The guardrails entail a broad set of companies spanning from cybersecurity 🛡️ to governance 💼 – Resistant.ai protects AI-native systems from complex fraud and sophisticated machine learning attacks through to governance; Enz.ai helps organisations adopt policies to build and deploy AI and track their practices against them.
Guardrails are plausible at many insertion points within a model pipeline. Software that ensures only the right company data is sucked into in an open-source model, software that sits on top of a model to ensure brand alignment and software that generally ‘troubleshoots’ the software supply chain in complex cloud environments, such as RunWhen.
These tools are increasingly integrated into infrastructure rather than at the application level, signaling a chronological shift leftward. This shift allows companies to enforce compliance and security across the entire model production sequence. While this approach may result in a higher rate of false-positives due to checks occurring before project completion, there's potential for reinforcement learning layers to enhance workflow, categorising issues as they arise. 🧠💻
Devops and deployment
The trickiest category to neatly compartmentalize is the tooling layer that aids developers in model integration. 🧩 These tools don't directly contribute to model 'building' but rather guide developers within enterprises on optimal model selection and utilization.
Typically modular, these tools concentrate on how models interface with both the enterprise and the end-user. Deployment channels become crucial, particularly in industries with fewer software developers available for custom model integrations. Companies within the Russell 3000, for instance, may be actively considering effective strategies for this intricate task.
This is one of the spaces that is moving incredibly quickly in the current LLM wave as the answers to many questions are yet to emerge.
Will all businesses need to integrate an LLM?
Do all sectors require a co-pilot based on industry-specific data?
Do all employees need to be able to interface with a model or only developers?
At Seedcamp, we have invested in several early-stage businesses in this space. Dust is a platform for creating sophisticated processes based on large language models and semantic search. AskUi is a prompt-to-automation platform that understands applications and helps users build UI workflows that run on every platform. Kern.ai helps companies connect external models to a internal data in a secure fashion.
Sector-specific discovery engine
Lastly, we navigate the predictive applications of 'old-school' AI—less in vogue but still invaluable. This category excels at identifying unknown unknowns, like uncovering suspicious trading patterns. Moreover, it's adept at spotting known unknowns earlier in workflows, particularly evident in settings like healthcare 🩺.
Despite the heady excitement around generative AI, predictive AI still has a lot of room to improve its own value proposition. For example, intelligent AML, KYC, and transaction monitoring technologies are still trying to crack financial crime and scams. The amount of risky and illicit crypto flowing into financial platforms is still in the hundreds of billions despite trending downwards YoY and represents 2-3% of all inflows.
At Seedcamp, we have focused on healthcare and fintech in this space to-date. In healthcare, we are investors in Ezra, an early cancer detection technology that combines advanced medical imaging technology with AI, and viz.ai, which alerts care teams to coordinate care, connect professionals to specialists and facilitate communication. In fintech, we are investors in 9fin, a platform for financial institutions to leverage analytics and intelligence and Elliptic.
AI's perpetual excitement lies in its ability to consistently challenge and reshape investor beliefs. Seedcamp’s first AI/ML commitment in 2012, James Finance, built fintech infrastructure to determine customer eligibility. In Europe, Lendable and Funding Circle became breakout companies in this space, but AI advancements didn’t propel a large group of winners in the intervening period and investor theses about the space decayed.
With every evolution of AI technology, the landscape of the category undergoes transformation. Notably, in 2021 and 2022, predictive fintechs like Marshmallow achieved substantial scale by leveraging technology to determine customer eligibility. This success prompted yet another pivot in investor sentiment, showcasing the dynamic nature of the AI sector.
Seedcamp is optimistic about the way AI upends sectors and is always interested in new solutions at every interval of the software value chain. We are especially excited about entrepreneurs with a differentiated perspective on the future. That might be that every company will operate myriad small open-source models via a routing layer, that the entire data pipeline needs to be rebuilt or that federated learning between farms will unlock a new wave of agricultural prosperity. It might even be that language models are a flash in the pan until they are in the hands of humanoid robot butlers, doctors, and waiters.
If you're working on a groundbreaking AI-driven solution, Seedcamp would love to hear from you. Get in touch!
*Funding Circle, Lendable, and Marshmallow are not Seedcamp companies.
Enjoyed these insights? These are the types of insights we uncover from being an LP, as investors don’t usually line up at our doors to share their insights. By marrying our investing, talking and writing, hopefully these learnings will help you on your journey in venture 💡 Want to explore more?
EU VC Pulse 🩺
While private market fundraising saw a year-over-year decline worldwide, Europe worked diligently to regain its capital share. The surge in semiconductor investments, with Europe actively participating in the high-stakes tech race alongside the US and China, highlighted a strategic realignment towards innovation-driven growth. Meanwhile, the impact of COP28 outcomes and debates surrounding the EU's AI Act reflected the delicate balance between regulatory oversight and fostering innovation.
Deep tech startups celebrated exits, serial entrepreneurs excelled, and infrastructure emerged as a darling in impact investing, signaling diverse opportunities in the European market. Yet, challenges persisted for SaaS companies, and the tech IPO market remained uncertain, reinforcing the need for investors to navigate this dynamic landscape with agility and foresight.
Highlights
🌍 Global Fundraising Hurdles: The global private market fundraising faces a year-over-year decline, signaling tightened belts and cautious strategies. Europe's fighting to regain its capital share, but North America's still the big player in the room.
🌐 Global Semiconductor Surge: Governments globally are pouring over $40B into semiconductor development. Watch Europe jostle with the US and China in this high-stakes tech race!
🌪️ COP28: A Climate of Controversy (🙌 kudos to Romain for the insights): COP28 concludes with mixed reactions, a sentiment echoed in Romain's insightful feedback. Acknowledging the historic step towards fossil fuel reduction, yet criticizing the lack of a concrete plan and firm targets for greenhouse gas cuts, it raises the question: breakthrough or a missed mark?
🔍 EU AI Act Under Scrutiny: The EU's new AI Act steps into the limelight, sparking heated debates. Tech circles buzz with concerns over potentially stifling innovation, despite the Act's aim to rein in AI's sprawling impact. Are we over-regulating the future?
💡 Tech's Shining Stars: The semiconductor sector's not just buzzing; it's booming with a $5.2B injection in Q3 alone. Public firms like NVIDIA, crossing the trillion-dollar mark, are leading the charge.
📈 US Economy's Surprise Show: Defying recession fears, the US economy flexed its muscles in 2023 with solid growth and consumer spending. Could this influence European markets, investment trends, and consumer confidence, amidst ongoing global uncertainties, inflation and geopolitical tensions?
🌐 VC and AI's Dynamic Duo: Nvidia leads the charge in AI investments with $872 million in deals, eyeing major players like Inflection AI and Cohere. This surge in strategic AI investments underscores the sector's potential. Aligning with this trend, our involvement in Curiosity Fund I positions us to capitalize on groundbreaking AI opportunities.
🔒 Tech and Immigration Tango: The UK's tighter immigration policies might just throw a spanner in the thriving tech industry. Meanwhile, Europe's GenAI investments are set to skyrocket.
💼 AllianceBernstein's Big Bet: Jumping into NAV lending, AllianceBernstein is set to ride the wave of growing demand with a laser focus on investment-grade loans.
👥 VC Firms Rally for Innovation: 37 VC firms band together, opposing the FTC's appeal against the Microsoft-Activision deal. They're sending a clear message: don't stifle our innovation and investment mojo.
🚀 Deep Tech's Exit Extravaganza: Deep tech unicorns are not just exiting; they're soaring with a 550% increase in exits since 2018. AI startups, especially in generative AI, have amassed an impressive $68.7 billion. Our investment in First Momentum Ventures taps directly into this trend, fostering the next wave of deep tech innovators.
🏃♂️ The Serial Entrepreneur Edge: In Europe, serial entrepreneurs are weathering the VC dealmaking storm better than their rookie counterparts, snagging funds faster and larger.
💡 Impact Investing's New Turn: Infrastructure's the new darling in Impact investing, pushing assets under management to a hefty $740.9 billion. Europe and Africa are leading the charge in this socially conscious investment shift.
🌍 Climate Policy's Uphill Battle: Despite global efforts lagging in meeting net-zero emissions by 2050, there's a silver lining with Impact investing. Infrastructure is now a key player, pushing assets under management to $740.9 billion. Our commitments to Satgana Fund I and Nucleus Fund II align with this direction.
🌐 Europe's Venture Puzzle: Despite a funding gap, there's hope for Europe's 25,000 new companies in the next five years. But big exits are rare, and survival strategies are key.
📉 Reddit's IPO Puzzle: Reddit gears up for a 2024 IPO but might face a valuation chop. The tech IPO market's in a slumber.
🤖 Big Tech's AI Adventure: The tech titans are diving deep into AI, quantum computing, and AR/VR to keep their growth engines running. It's a strategy shift amid a market slowdown and regulatory glares.
📉 SaaS's Stumble: European SaaS companies are feeling the pinch, with median ARR growth dropping and profitability becoming a tightrope walk. It's a challenging landscape, but resilience is the name of the game.
🤖 AI's Mixed Bag: Generative AI's reshaping the tech world, but OpenAI's GPT-4 hits a snag, opening doors for rivals like Claude and Gemini. It's a tech tussle with high stakes.
🔐 Cybersecurity's Funding Uptick: Cybersecurity's back in the game with a 29% funding bump in Q3’23. It's a sector that's quietly outperforming its tech cousins in the funding arena.
Deep Dive
Global Trends and European Market Dynamics
The observed year-over-year decline in global private market fundraising, with Europe striving to regain its capital share against North America's dominance, underscores the broader economic challenges and strategic shifts in the investment landscape. As sectors like private debt show resilience through steady growth in total assets, the increasing prevalence of evergreen funds and retail offerings suggests a pivot towards more sustainable and long-term investment models. These trends indicate a need for European investors and startups to adapt their strategies, potentially leading to a more diversified and resilient portfolio.
Moreover, the global surge in semiconductor investment, with significant contributions from Europe, aligns with these broader trends. Europe's active participation in this high-stakes tech race, particularly against the backdrop of the US and China's dominance, signals a strategic realignment towards innovation-driven growth. This shift is further exemplified by the semiconductor sector's robust growth, as evidenced by NVIDIA’s trillion-dollar valuation, indicating the sector's economic potential and its crucial role in driving technological advancements.
Strategic Shifts in Investment and Policy Influences
The UK's tighter immigration policies juxtapose the European trend of increasing investments in General AI (GenAI), highlighting a dichotomy in the tech sector's growth and talent acquisition strategies. This situation points to a potential shift in the European tech landscape, necessitating a focus on local talent development and alternative work models to counteract potential talent shortages.
Simultaneously, the collective opposition of VC firms to the FTC's appeal in the Microsoft-Activision deal marks a significant stance on fostering innovation through mergers and acquisitions. This case could set a precedent for how future tech mergers and acquisitions are perceived, potentially impacting investment strategies and growth trajectories of tech startups.
This development comes at a time when the EU is actively scrutinizing AI through its proposed AI Act. The Act, aimed at regulating the sprawling impact of AI, has sparked debates in tech circles regarding its potential to stifle innovation. The juxtaposition of these policy changes against the backdrop of increasing AI investments highlights the need for a balanced approach that fosters innovation while ensuring ethical and responsible AI development.