Magma Partners Q3 2023 Investor Letter

November 29, 2023

February 23, 2021

We send quarterly investor letters to our LPs about what we're seeing in the Latin America startup market. We share edited versions with our portfolio. We’ve decided to share a further edited version publicly.

Magma LPs,

Hope you had a great Q3. In this letter, we’ll cover:

Big divergence between startups that can raise VC money and those that can’t

  • Early stage - Much easier at pre-seed and seed than later stage, new rounds are happening at much lower valuations than in 2020/2021
  • Later stage - Good/great companies can raise, but on lower multiples than before, some companies have to take down rounds if they’re inefficient or mistimed the cycle
  • Everyone else - It’s hard to raise money and many of these companies are in big trouble
  • Much longer time between rounds. More pronounced delays going from Seed to Series A

Artificial Intelligence (AI): Overhyped or a must have for all startups?

  • Huge trend that is already improving productivity, all startups should try to use AI.
  • We’ve made a few AI investments and are on the lookout for more.
  • AI is super early, lots of hucksters, similar to 2021 when every pitch deck had blockchain whether it made sense or not.

Macro Uncertainty = Startup Uncertainty

  • The economy is sending mixed signals, some of the economic indicators show a high likelihood of soft landing.
  • Others are flashing yellow and red, which adds risk to the system and higher chances of a recession.
  • Two wars are adding instability and unpredictability.
  • Later stage VCs are mostly underwriting this uncertainty and investing at lower valuations.

Latin American startups should relearn the pre-2019 playbooks, before the VC bubble

  • Capital was scarce in Latin America until ~2019.
  • Great companies could raise money, but they had to be efficient.
  • Efficient growth and metrics that fell out of favor during the boom like revenue per employee burn per employee, burn multiple, CAC payback periods, are going to be increasingly important.

Early stage deals happening at lower valuations; Post Series A, good companies can raise, but even good companies sometimes have to take down rounds.

Early Stage: Few competitive rounds, but capital available, AI booming

Most pre-seed and seed deals were still priced for the 2021/2022 boom market until Q2 2023. By Q3 2023, most pre-seed and seed stage companies were back to pre-bubble valuations.

Series A is a mixed bag. The best companies that are growing efficiently are raising rounds at a markup to 2020-2022 rounds. At the same time, we’ve seen good companies try to raise Series A, but end up with a Seed extension at a flat or lower valuation than previous rounds.

Companies without significant traction or that were burning too much are pursuing acqui-hire M&A transactions. Some are shutting down.

Later Stage: Efficient, fast growing companies with good unit economics command competitive rounds, everyone else is struggling to get multiple term sheets. Some down rounds, forced M&A, or shutdowns.

At later stage, there’s three groups of companies:

Startups that never found product market fit, or have really bad unit economics

  • These companies are likely to get acquired for large discounts to the capital they raised or shut down when they run out of money and come to market again.

Startups solving real problems, but raised at high valuations compared to traction

  • These companies can likely raise money, but face multiple compression and likely have to raise at lower valuations than their last round even with good growth.
  • Many companies that cut spending aggressively have not come to market yet, but will in 2024.

Startups that have great unit economics and are efficient

  • Most of these companies have not come to market because they have cash and can wait as they can grow into their valuations and hope the market improves.

Valuations and multiples are down, but the time between rounds is also increasing across the board, with the biggest change at early stage:

After 7+ quarters in the new startup environment and big name startups shutting down, some founders still seemingly believe we’re going to bounce back to the bubble times and are operating on a ZIRP-lite playbook. As we’ve consistently written, we don’t believe a 2021 type market will happen anytime soon.

Artificial Intelligence: Already starting to impact the market, but many hucksters

AI is booming, with many high priced AI rounds. Just like in 2020/2021 when huge numbers of startups pivoted to blockchain, nearly every startup we’ve seen recently has some sort of AI in their pitch deck. Most of the AI in the pitch decks are hand-wavy, buzzword-filled slides that don’t stand up to a few direct questions. Just like with any other boom/bust cycle, the vast majority of companies will fail.

At the same time, there will be massive AI winners. We have made multiple AI investments and are on the lookout for more, while trying to avoid paying extremely high valuations way ahead of traction. We’re also using AI to automate manual work and help us make better decisions. Using AI helps us further hone our investment thesis, identify which startups are hand-wavy pivots that likely won’t work and identify startups to invest in.

When you see something that actually uses AI, it’s magic. Runway is changing the video industry. Shopify says they’ve written 1M lines of code using Github Copilot, a huge productivity gain. Deepmind can forecast the weather better than a $200M+ per year supercomputer and AI can diagnose breast cancer better than NHS doctors. We’ve seen magic from some portfolio companies and are on the lookout for AI-magic every day.

AI is a game changer, but even though we are still in the first innings, things are changing quickly. These two graphs from the Financial Times are already showing AI’s impact in editing and freelance writing.

AI is also leveling the playing field, closing the gap between lower performing and higher performing workers:

Anecdotally, we’ve seen this same trend: lower performers or junior team members can get a huge boost by using AI proficiently.

Our hypothesis is that this trend flips as higher performing workers start to figure out how to use ChatGPT and other generative AI tools more efficiently, unleashing their productivity. Other current high performers could get left behind, like those in the early 2000s who refused to use the internet or word processing.

Macro Uncertainty is still affecting startups: Mixed signals… nobody really knows

The economy is sending mixed signals. The data is all over the place.

  • US GDP growth was higher than expected at 4.9%
  • US Inflation was 3.7%, down from 9.1% at the peak of the pandemic
  • The FED signaled it likely won't raise interest rates soon
  • The FED balance sheet is continuing to run off, with a $1.1T reduction since May 2022 and $100B more per month, a deflationary force, after 9xing from 2008 financial crisis to the 2022 peak
  • Bond holders like banks and insurance companies that bought in 2021 have taken huge losses in US treasuries they bought in 2021 at 1.2% that now are sitting at ~5%
  • The US has to issue more bonds to finance the debt and we could get another government shutdown or another debt ceiling crisis in Q4 or early 2024

There’s also two wars, consumers are looking like they might be struggling with debt, housing and commercial real estate are under strain.There’s still lots of risk in the system, but the economy continues to hold up.

Our stance is that startups should be capital efficient while still growing. It’s better to grow more slowly and control your own destiny than grow faster and rely on VCs to keep you alive in an uncertain market.

With fewer funds writing checks into Latin American later stage companies, raising a Series B or later round is make or break for startups that don’t control their own destiny. If startups don’t control their own destiny, or at least have the unit economics and a controlled burn, they risk down rounds, forced sales or even shutdowns. If the market changes, startups can accelerate growth, hire more quickly and adapt again.

Startups should use lessons from the pre-Covid, pre-Latin America startup boom operators manual

In the Q2 2023 update, we wrote about muscle memory. Many founders and investors are used to falling interest rates and ZIRP, reflexively thinking there will be a bounce back to what life was like before the reset. Interest rates have generally fallen from the early 1980s peak all the way to 2022. Many founders and investors are still using the same 2008-2022 ZIRP-era playbooks, but operating the ZIRP playbook a bit more efficiently. The ZIRP era is the outlier.

Frank Rotman, QED’s Chief Investment Officer, recently tweeted that founders need to relearn what worked 10 years ago before the ZIRP distortions. We agree. The best companies in our portfolio have done just this.

Latin American startups would do well to go back to their roots when capital was less abundant and focus on healthy, efficient growth.

Nubank, the most successful Latin American startup since Mercado Libre in the 2000s, only had one core product in one geography, Brazil, for many years before expanding to other countries and adding new products. They rode Brazil and their core product to a $40B IPO.

Our portfolio has a mix of startups that were able to raise big rounds during the bubble and went after hyper growth and had to do major restructures mixed with startups which continued to grow while being efficient because they didn’t have any other choice.

The best performing companies cut burn decisively in late 2021 and early 2022, focusing on 1-2 products. Some even gave up revenue with unsustainable unit economics to focus on higher margin, more scalable software revenue. These companies had been able to raise big rounds of VC and followed the ZIRP growth playbook, but then made big changes, cutting their team between 45-75%, focusing on 1-2 products and reducing burn between 50-95%. Some of these companies are already profitable, or are likely to be profitable in early 2024.

The companies that were always efficient because they had to be and weren’t able to raise big rounds during the boom cut burn and continued to focus on efficient growth. It remains to be seen whether VCs will reward underestimated founders with great unit economics, but we hope that they will.

More difficulties for inefficient companies, opportunities for investors

As we wrote in Q2 2023, “the next few quarters are shaping up to be extremely hard for companies that are burning too much money, no matter their stage.” It’s still the case. 

VC investment in Latin America may be hitting a bottom, but with the exception of some later stage deals, the market doesn’t seem to be heating up much yet.

It still feels like we’re bouncing around the bottom now, but unless there’s a US interest rate cut, we don’t see Latin America deal and investment velocity rebounding materially until at least mid 2024, maybe longer.

Later stage investments might be the exception as companies come to market to extend runway. There is money available, but it’s very expensive money for all but the absolute best startups operating efficiently. According to Carta’s Q3 US data, more companies are taking down rounds. We think it’s likely similar in Latin America.

Late Stage VC investment grew in Q3 2023, but is still well below Q4 2022.

Early stage:


Our advice to founders is the same: control your own destiny, keep growing, but don’t push to grow faster if it means bad unit economics. Try to extend runway into 2025, and be in a position where you can get runway to 2026, or even profitability, if you can’t raise on the terms you want.

Innovation doesn’t stop. Great startups keep growing, no matter the market conditions. AI is likely a game changer. We’ll be active in the market trying to find these opportunities, while helping our current portfolio navigate these choppy times.


Nathan, Pedro, Mak, Francisco and the Magma team