Two things inspired this week’s AI update 1) The calendar, which is always a fun source of inspiration. With earnings season largely behind us, we are forced to mark to market our call on AI differentiation. And 2) High frequency data that fits with our HUNCH – based on comments from Jenson Huang and Ben Thompson at Stratechery HERE and HERE – that the death of all software companies is exaggerated. Ara Kharazian, an economist at Ramp, a corporate payments processor, detailed on Andreessen Horowitz’s a16z podcast how Ramp’s spending data shows incumbent Software has been resilient (so far) (HERE).
At the beginning of earnings season, we argued there may be some upside in S&P 500 Software ex Hyperscalers IF enough companies began proving they weren’t being displaced by AI, given the hits to forward earnings growth and cash return already priced in. Our fair value modeling* implied valuations were discounting a 6pp hit to the forward earnings CAGR or a ~20pp hit to cash return (dividends + buybacks / net income) (HERE). Our main point was if some Software companies are able to defend their moats/utilize AI themselves, the index was not a compelling short. Today we are marking to market that thesis…
Net net, earnings and the narrative shift around earnings reinforce our view that Software ex Hyperscalers is still no longer an obvious short. There are still significant hits to earnings and/or cash return priced in, while earnings and cash return held up in 1Q and the evidence is that businesses are not yet pivoting away from paying seat-based subscriptions. Software has not yet proven it won’t be disrupted, so it’s not an obvious long either. The practical implications are 1) the asymmetry is higher for the group as a whole and 2) stock picking between winners and losers will matter more going forward. Correlations within Software have declined, supporting this thesis.
Earlier this week, the Quant team highlighted a list of Software stocks with high EPS revisions and positive earnings sentiment (an objective score of how optimistic management teams sounded about their company earnings during their earnings calls). These names have outperformed the S&P 1500 by +10% YTD, accelerating higher during earnings season. Full report HERE and stock list below.
Details and charts below…
S&P 500 Software ex Hyperscalers are up +7% MoM relative to the S&P 500. Earnings helped the group stabilize; on aggregate, the average excess return from earnings was positive (+1.2%), though the median was negative (-1.2%). In other words, the hit rate of earnings wasn’t great, but the skew of returns was positive. That was the point we made about the asymmetry heading into earnings.
Narrative shifts help explain the price performance more than earnings releases. For that, we can draw on some evidence. Ara Kharazian, an economist at Ramp, a corporate payments processor, detailed on Andreessen Horowitz’s a16z podcast how Ramp’s spending data shows incumbent Software has been resilient (so far). Businesses are still paying seat-based subscriptions, not consumption-based models — AI hasn’t disrupted the buying behavior yet. Podcast HERE. This lines up with conversations we are having with investors.
The move in valuations now discounts less, but still significant, declines in earnings estimates or cash return. On their own, valuations imply the 5y forward EPS CAGR falling from 12% to 9% (or 2y forward EPS CAGR falling from 19% to 13%), or Cash Return falling from 88% to 77%. Together, something like an ~83% cash return ratio and an EPS CAGR of 10%. In 1Q, earnings estimates increased, and cash return remained stable. One quarter in, there is little evidence of the downside developing in aggregate. There will be relative losers, but some winners as well.
The below matrix shows Fair Value (% from current price) of different combinations of cash return and earnings, assuming that the equity risk premium was moving ahead of earnings and cash return estimates.
It is still plausible that Software margins or revenue will be impaired. We are not disputing that. Our point is how much is already priced in. IF Software can continue to prove resilience, the equity risk premium is biased towards falling back in-line with a more “normal” spread to the S&P 500. Recovering the entire spread implies a +32% relative return.
Fair Value with different ERPs is charted below, for reference.
Correlations within Software dropped over earnings season. Stock picking is effective within Software again.
Earnings estimates increased in 1Q and cash return improved. Downside cases are not yet being realized on aggregate.
*Our Fair Value model is based off the work of valuation guru Aswath Damodaran. It is a DCF style approach that is akin to valuing an index like a single stock. The cash that is discounted is the cash returned to investors through dividends and buybacks. This approach has helped us work through valuations in a way that is more tangible than say an NTM PE. Please let us know if you would like a copy of our model (or a clearer explanation).
Software names with positive EPS revisions and earnings sentiment have outperformed this year, accelerating higher during earnings season. This is a good list to look for longs within Software as Software correlations decline.
