What to do after the SaaS sell-off
From the outset, I'll reiterate what I said about 2 months ago, Claude Code (and ChatGPT Codex) is the single most significant thing that has ever happened in my career.
The silly thing on my front is I should have been short tech names because in hindsight, if this affected me as someone who works in finance let alone tech workers, can you imagine the disruption in the tech sector? Well you don't even need to imagine, ASX XIJ (−43%), ASX XTX (−37%) are representative of what happens if you're exposed to "disruption" from AI. That is an utterly brutal drawdown in anyone's book, to get back to evens, XIJ has to bounce +75% to get back to previous peaks.
Not in the realm of impossible but things have fundamentally changed so the question we must answer is, can we get back there?
Let's actually try to answer this.
Feedback
I have spent the last few weeks talking to a few SaaS companies (CFOs, COOs, CTOs) about how they're approaching AI and their views. The majority of these companies are private, I have a holding in some of them (which I think are now worth a lot less). Insights below:
- All engineering teams are using Claude Code/ChatGPT Codex, I haven't talked to a single person who said they haven't incorporated this into their workflows. Existential crisis realised.
- V: Personally, this is one of the most shocking things I've ever heard as the pace of adoption is ridiculous. But this also lends to the idea that investing in software dev tools is a mugs game, there is no barrier to entry and users are incredibly fickle.
- V: On PA, obviously one of my bigger direct holdings as private SaaS is a devtool... ffs
- I'll quit digressing, when I've talked to CFOs the interesting thing is they can't point out to efficiency just yet and still challenged to see how things work, some comments (non-verbatim obviously):
- It has absolutely made everyone efficient
- Everyone is doing more and are a bit more productive
- Sales teams generate revenues, so picking up the phone calling people or meeting people hasn't changed and this has a particular cadence – AI has done nothing to unlock value here
- Engineering teams are cost centres
- One question I posed is if they can do more, can it be turned into revenue
- Answer is maybe; you can ship more stuff or improve outcomes but that doesn't necessarily turn into revenue
- Second question I've been asking is can you reduce headcount
- Blatant answer is "no"
- Surprisingly everyone I spoke to has already gone through the headcount exercise with absolutely 0 change
- "You can fire someone but then I need someone to train the agent"
- That is the main thing I kept hearing, you can't really reduce headcount because engineering teams are treated as a cost centre – they're already run as tight ships
- V: This is perhaps reflective of the zero cost-of-capital era expiring and the last tech blow-up has already forced everyone to think about their economics and profitability more closely.
- Perhaps reduction in headcount growth, but no direct benefit in headcount reduction. Yet...
- The final bit of feedback is everyone recognises that the barrier for new entrants is significantly lower, that if a few sales people and some strong engineers got together, that any industry can be disrupted now
This lends to the final idea of although you won't see revenues change all that much (in the near term) and perhaps there might be better profitability at SaaS companies, that the terminal values all need to be marked lower because the barrier for competition is just so much lower today.
So the direct answer is earnings haven't fallen, this is another valuation recession for Tech just like 2022. Earnings haven't gone anywhere but terminal values for every business is lower (for now?).
Noting that these are all <$500m Mcap, <$50m revenue companies so the caveat is headcount could be reduced at large companies that I haven't picked up on quite yet.
The Playbook
Frankly I think AI is utterly existential if you're a large tech company, but I still think large tech companies are better positioned to navigate who knows what than smaller ones because at the end of the day, it is all about sales.
How do we play this?
- Need to risk weight it appropriately
- Need to be aware that any new competitor for any listed company that gets any kind of traction will likely push around valuations significantly. One AFR article can easily halve Xero tomorrow if someone raised capital for an "AI Competitor".
- I think no change to fundamental earnings in the next 12 months – so an opportunity in this respect
- The bull case is potential upside earnings revision due to potential headcount reduction
- An outcome here is valuations might fall but earnings improve – there is a hedge there against each piece but unfortunately valuations are more volatile. Both risk and opportunity.
- Allocating here is taking a bet that valuations bounce back from your entry point + earnings improve due to headcounts. Both scenarios plausible.
- Picking bottoms (hehe) is challenging, there is still significant shareholder turnover in these companies.
On sale, everything is 50% off!
Imagine if a year ago someone told you that you could buy WTC, REA, TNE, CAR or PME at 30–50% off. Everything is on sale! Opportunities like this seldomly appear but need to climb the wall of worry and position appropriately, how much risk are you willing to take on? Are you ok if this goes down another ~50%? How much permanent capital destruction is ok?
- If you're looking at US stocks, stock-based compensation (SBC) is still a significant headwind.
- I think it was Joe from Bloomberg's Odd Lots (maybe, I can't remember) who said something this week along the lines of - if terminal values are falling, and these SaaS companies are all negative free cash flow after SBC, then they have no earnings and no terminal value so its not a surprise they're all falling.
I honestly don't know what the final outcome is and neither does anyone, which is reflected in lower terminal values and lower valuations across the board. But I don't think there is earnings erosion in the near term so the ultimate question is whether the market will reward this in again the future?
An idea you rent rather than own for now. Tech migrating to cyclicals.