🤔 TVPI AFTER 2023 SURGE IN FAILURE RATE

Mar 27, 2024•12 min read• | | |

― By Luke Carroll, CIO

Following the notable surge in company failure rates during 2023, the question arises: Can LPs now rely on TVPI? This issue stands as one of the most hotly debated topics among LPs presently.

Over the past 8 years, interest rates remained at historic lows, while inflationary pressures stayed subdued. The unforeseen black swan event of Covid-19 prompted a loosening of monetary policy to spur economic recovery, ushering in an era of abundant liquidity. Under these unique circumstances, technology companies played a pivotal role in facilitating remote operations across various sectors such as remote work, cybersecurity, e-commerce, digital health/telemedicine, and fintech, experiencing accelerated growth during the pandemic.
 
 This surge in public market valuation, propelled by excess liquidity, along with a rise in dry powder in VC, intensified competition, bidding wars, and overall valuations. The dynamic between founders and capital deployment became a supply and demand issue, exacerbated by the SPAC boom and speculative behavior, where investments seemed destined for continual growth. Consequently, this led to a scarcity of attractive offers, a detachment from fundamentals, and inadequate benchmarking. Moreover, VCs faced substantial incentives to deploy capital swiftly, driven by the influx of LPs seeking to invest more capital, resulting in a race to secure as much funds as possible.
 
 These factors culminated in an average failure rate lower than the median, particularly in the last 8 years, as illustrated in the table below.

“1,498 startups perished due to financial insolvency.
 Put bluntly, over a thousand startups survived
 solely because VCs extended financial support.”

– Maggie Po, Carta 2023

Despite some LPs believing that the major correction occurred in 2023 and that TVPI now accurately reflects the current market state, there are indicators suggesting that 2024 might witness the most significant TVPI reset yet. We share a few of these below:

  1. We have held discussions with a sample of our 50 GPs and performed a full review of their portfolio companies highlighting any company that had not raised follow on capital since 2021, or those that only raised convertible notes since then. From these discussions, it became evident that we should expect additional valuations adjustments in 2024. This is further amplified by the fact that only the top performing companies are finding capital (see #3).
  2. Data from Pilot.com revealed that 56% of companies are set to raise capital in 2024, a substantial figure considering most companies have 2–3 years runway. This is due to the fact that many companies between 2022–2023 riased convertible notes from existing investors and/or were forced to cut burn to extend runway. It’s unlikely these companies will receive further bridge round financing and will now be expected to test the fundraising market.

3. Capital deployment has been notably slower in 2023, with Q4 2023 marking the worst quarter in VC for over 6 years. This slower pace has raised the bar for investors at each stage, potentially posing challenges for companies performing below or at expectations in raising capital.

4. The trajectory of the first quarter of 2024 suggests it may be the toughest year yet in terms of shutdowns. 61 companies (which had raised $5m+) shut down in January and February alone, surpassing the total closures in Q4 2023, with March figures yet to be added.

Additionally, disheartening figures from the above chart:

  • January witnessed the highest number of startup shutdowns on Carta across Seed-stage, Series A, and Series B companies. For Series C and beyond, February was the worst month.
  • The first two months of 2024 saw more startups, which raised at least $10 million and $20 million, shutting down compared to Q4 2023, marking a prior record quarter.
  • The first two months of 2024 also marked the worst period for shutdowns in Fintech, SaaS, Healthtech, and Medical Device sectors across all Carta data.

Nevertheless, there are some positive aspects:

1. VC funds on Carta called more capital in January 2024 than at any point since mid-2022. This influx of capital, although yet to be invested, indicates a readiness among fund managers to deploy into startups. There may be several factors driving this trend but the most obvious is that the public markets are looking strong (see #2 below) and funds typically have a 5 year deployment period with many deploying very little in the last two years. The dry powder remains at all time highs and we may start to see valuation tick up when capital is forced to be put to use again.

2. VC trends often lag behind the public market, which has been experiencing a rally, especially in technology sectors, attributed to renewed enthusiasm around AI. AngelList data suggests there is a positive correlation between lagged public markets (9–18 months) and the private market. This lagged correlation suggests the venture market will continue to slump in early 2024 but will start to pick up towards the end of 2024/early 2025. 
 
 3. Positive signs in the IPO market — such as the Reddit IPO trading 50% higher on first trading day has instilled fresh hope that the IPO market is starting to heat up and potentially begin to open the way for more companies to explore and exit.

Our analysis suggests that companies are likely to face increased challenges in 2024, with failure rates potentially reaching all-time highs as the competitive bar remains elevated. However, amidst this turbulence, a new trend is emerging: there is a significant accumulation of dry powder, prompting funds that previously hesitated to deploy capital to now accelerate their investment over the next 2–3 years. Concurrently, there are signs of recovery in the prices of public market tech stocks.
 
 We anticipate that these developments may culminate in intensified competition towards the latter part of 2024, leading to even greater disparities in valuations among successful companies, supported by elevated public multiples. Moreover, we foresee a gradual lowering of the performance expectations for companies to accommodate the active deployment of the accumulated capital for the coming years. 
 
 If our projections hold true, successful venture capital firms may be those that maintained a consistent investment pace throughout 2022 and 2023, keeping standards high while avoiding attempts to time market fluctuations in valuations. Anecdotally, an observation from the secondary market indicates that the pricing floor was reached in the summer of 2023, and since then, the gap between buyer and seller has narrowed in favor of the sellers. 
 
 Read more on the subject ⬇️
 
 Reddit closes nearly 50% higher on 1st trading day in latest sign IPO market heating up — Fortune
 
 The venture market downshifts further in 2023, with dealmaking and funding totals falling to 6-year lows — CB Insights
 
 Pilot: 57% of Venture Startups Will Need to Raise More In 2024 — Saastr

In case you missed it…

General Technologies 🚀

🤖 Reddit’s IPO Triumph: Reddit made headlines with its impressive initial public offering, with shares surging 48% on the first day (MAR 21, 2024), closing at $59.80 from a debut of $34. This significant jump reflects strong investor confidence, raising around $750 million. Notably, it marked the first major social media company to go public since Pinterest in 2019. Reddit’s IPO came the same week as data centre hardware vendor Astera Labs, which also skyrocketed by 72% on its opening day on the Nasdaq. These events indicate a potentially revitalizing IPO market amidst prior concerns over rising interest rates and inflation. Read more about Reddit IPO here.

⚖️ EU’s Tech Legislation Movement: The EU has launched probes under the Digital Markets Act (DMA) against Apple, Alphabet, and Meta for potentially breaching tech regulations. Apple’s past violation resulted in a €1.8 billion fine for restricting developers from informing iOS users about alternative and cheaper music subscription services available outside of the app. The new investigations challenge practices around user choice in app stores and services. Alphabet faces scrutiny over its search results’ fairness, and Meta is examined for its ad-free subscription models. Fines could reach up to 10% of annual global turnover, increasing to 20% for repeat offences. The EU aims to conclude these investigations within 12 months, underscoring its commitment to enforcing digital market laws. A complete description of the probes is available here.

✈️ Boeing’s Leadership Overhaul: Boeing announced major leadership changes; CEO Dave Calhoun is set to depart by year-end amidst ongoing safety challenges, including the recent Alaska Airlines incident and two fatal crashes of the 737 Max in 2018 and 2019 that killed 346 people. The chairman and the head of the commercial airplane unit are also leaving. The reshuffle signals a strategic pivot to restore trust and stability in Boeing’s operations and relations with major airlines frustrated by delivery delays and quality issues. Read more about these changes here.

🧠 Innovations in AI Training: A new approach, Direct Preference Optimization (DPO), is streamlining the training of large language models (LLMs). The new DPO method enhances AI training by allowing LLMs to learn directly from data, bypassing the need for a separate reward model based on human feedback (Reinforcement Learning from Human Feedback (RLHF)). Its ease of use allows smaller companies to tackle the alignment problem, and leading companies such as Mistral and Meta are testing it. Read more about Dr. Sharma’s discovery here.

🎧 What We’ve Been Listening To This Week

🎙️ Lex Fridman X Sam Altman: The CEO of Open AI joined Lex Fridman to discuss the company’s upcoming model, GPT-5, their video-creating AI model, Sora, and to share thoughts on Artificial General Intelligence (AGI).

Sustainability 🌍

🌍 AI’s Environmental Impact: Sustainable Promise or Resource Drain?

As AI skyrockets, a pressing question emerges: Is it reasonable to presume that AI will ultimately prove more environmentally beneficial than the energy and water resources it consumes to operate?

🚨 The Environmental Toll of A.I.

By 2027, the AI sector’s annual energy consumption could range from 85 to 134 terawatt hours — an amount comparable to the Netherlands’ yearly energy demand.

  • Energy Intensive: Data centers consume ten times the energy of a typical household, with AI training requiring up to three times more power than standard cloud workloads. For instance, training large language models like GPT-3 consumes nearly 1,300 MWh of electricity annually — equivalent to the annual power consumption of 130 US homes.
  • Expanding Data Centers: The proliferation of cloud data centers, estimated between 9,000 to 11,000 globally, is driving a substantial surge in electricity consumption. The International Energy Agency (IEA) forecasts that data centers’ electricity usage in 2026 will double compared to 2022, reaching 1,000 terawatts — nearly equivalent to Japan’s current total consumption.
  • Water Usage: Not only electricity, but A.I. also impacts water consumption. Google and Microsoft data centers saw significant increases in water usage, with Google’s data centers consuming about 5 billion gallons of fresh water for cooling in 2022, marking a 20% surge compared to 2021. Meanwhile, Microsoft recorded a 34% rise during the same period.

🌻 A.I.’s Role in Climate Mitigation

A recent report from BCG, in collaboration with Google, titled “Advancing Climate Action through AI,” reveals that the current deployment and expansion of AI technologies have demonstrated their capacity to significantly decrease GHG emissions, bolster climate-related adaptation and resilience initiatives and could lead to revolutionary scientific breakthroughs, such as advancements in nuclear fusion, offering novel pathways for addressing climate challenges. Here are a couple of examples:

  • Industry Integration: Aviation is one of the hardest sectors to decarbonize. According to the IPCC, contrails — the thin, white lines you sometimes see behind flying aircraft — account for roughly 35% of aviation’s global warming impact. Google Research teamed up with American Airlines and Breakthrough Energy to combine AI and huge amounts of data to predict where contrails will form and how planes can avoid making them. The trial reduced contrails by 54% across 70 live American Airlines flights.
  • Energy Solution: A collaboration of Google DeepMind and the Swiss Plasma Center at EPFL, a Swiss university, has leveraged AI to create the first deep reinforcement learning system for fusion research. It simulates EPFL’s Variable Configuration Tokamak (TCV) and has successfully modeled ways to stabilize and sculpt plasma that have subsequently been validated in the actual TCV — opening new avenues to advance nuclear fusion research.
     
     Sam Altman, CEO of OpenAI, has allocated hundreds of millions of dollars towards fusion research. In recent interviews, he has hinted at the potential of this futuristic technology, often regarded as the holy grail of clean energy, to meet the global demand for vast amounts of power. Fusion energy not only has the capacity to fulfill the 3–4% estimated power needs of next-gen AI (S&P) but also holds promise in transforming numerous industries, many of which are significantly more polluting, into more sustainable ones.

The highlighted examples merely scratch the surface of the myriad applications where AI can play a crucial role. While it’s clear that there are some negative points and questions to explore in more depth, there is also compelling evidence that AI is poised to yield more positive outcomes than negative one in various sustainability endeavors.
 
 That said, fully unlocking AI’s potential in sustainability demands meticulous attention to ethical considerations, data quality, environmental impacts, and equitable access. By navigating these complexities with foresight and responsibility, we can leverage AI’s transformative power to propel meaningful progress towards a more sustainable and resilient future for all.
 
 To achieve this vision, effective regulatory frameworks, inclusive collaboration, and ongoing innovation are indispensable. These efforts will guide the seamless integration of AI into sustainable practices, ensuring that technology serves as a catalyst for positive change, safeguarding our planet and enhancing the well-being of present and future generations.
 
Read more on the subject ⬇️

How AI Can Speed Climate Action — BCG
As Use of A.I. Soars, So Does the Energy and Water It Requires — Yale Environment 360
How much electricity does AI consume? — The Verge

Blockchain & Crypto 💸

⚖️ Regulations

  • Honk Kong’s Central Bank launches a regulatory sandbox for stablecoin issuers to facilitate oversight.
  • Europe adopts new Anti-Money Laundering regulation, but removes the €1’000 limit on cryptocurrency payments from self-hosted wallets.
  • Hedge fund industry groups sue US SEC over Treasury market dealer rule.
  • Genesis agree to pay $21m to the SEC to settle down.

🏦 Financial Institutions

  • SWIFT is set to launch a central bank digital currency (CBDC) platform in 12–24 months.
  • Japan’s Pension Fund (GPIF) is considering deploying capital into Bitcoin.
  • BlackRock launched their first $100m tokenized fund through Securitize.
  • BlackRock is set to add spot bitcoin ETFs to its Global Allocation Fund.

🔥 Top Stories

🔎 Research

📄 Shayon Senhupta (Multicoin) discusses how crypto facilitates direct value transfer, proposing the emergence of ‘Publisher-Exchanges’ — apps that blend content publishing with financial exchanges.
 
 📄 Justin McAffee (1kx) explores the intersection of AI and crypto for consumer applications.

📹Videos
Collab+Currency’s Amanda and BlockTower’s Winnie welcome Sharon (Multicoin) and Mark (M13) to discuss consumer trends in crypto.
 Bankless discussed the 7 most bullish crypto narratives.
 Bankless invited Robert Leshner to discuss the state of RWAs.
Bankless invited Nic Carter to discuss the state of Bitcoin.

Life Sciences 🔬

🧠 Neuralink’s Groundbreaking Leap with First Human Brain-Chip Interface
 Neuralink, Elon Musk’s brain chip company, has released a video featuring its very first patient. Mr. Arbaugh, who is paralyzed below the shoulders, is shown playing chess after receiving the device in January. He controls a computer cursor through the device, a brain-computer interface that can translate his thoughts into action. The technology can record and decode neural signals and then transmit information back to the brain using electrical stimulation. Challenges such as refining the technology and navigating regulations remain, but the prospect of mind-controlled devices signals a transformative trajectory for technology.

💡 A.I. Is Learning What It Means to Be Alive
Carl Zimmer for the New York Times takes us on a fascinating journey into the potential impact of AI on biology. He tells the incredible story of how AI discovered Norn cells, kidney cells essential for producing a key hormone, in just six weeks — something that took humans 134 years to do. Researchers at Stanford University programmed computers to autonomously analyze raw data from millions of real cells and their genetic makeup. In the process, the computers categorized cells into more than 1,000 types, including the discovery of the Norn cell, all without prior knowledge of its existence. This breakthrough highlights AI’s ability to independently uncover new biological insights. Challenges such as data availability, ethical concerns, and model limitations are also acknowledged, adding depth to the discussion.

🏥 NVIDIA Healthcare Launches Generative AI Microservices 
 Healthcare companies take advantage of the latest advances of generative AI. The microservices allow companies to quickly integrate specific tools from across the NVIDIA platform into their R&D workflow with flexible pricing. These include solutions for image processing, natural language processing, and digital biology generation, prediction, and simulation. This launch aims to democratize access to advanced AI tools in healthcare, enabling tasks such as drug compound screening, improved patient data collection for early disease detection, and smarter use of digital assistants. Read more here.

Not subscribed yet?

Subscribe Here

More from Reference Capital

View all reports