AI’s rising capital demands
The AI thematic is still front and center of investors’ minds it seems, but sentiment has turned sharply from optimism to skepticism.
Speculation that AI would introduce a new productivity boom drove the narrative, and the market appeared to take in its stride the huge sums Microsoft, Amazon, Meta and Alphabet were committing to the AI race.
In 2026 they are expected to spend more than $600bn, eating up their free cash flows and reversing the ‘spend little, earn lots’ formula which characterized these companies. “This level of capex will consume almost 100% of hyperscalers’ cashflow from operations compared with the 10 year average of 40%.” UBS said.
When Microsoft announced stagnating growth in Azure, its cloud-computing business, and high capital spending, the market ignored the solid earnings report and sold the stock down 17.5% in the first two weeks of February.
Ameriprise describes it as a “no-win situation”, commenting “Investors were comfortable saying ‘so long as it happens in the future, I’m comfortable with Microsoft or Amazon or Alphabet spending the money.’ Now they want to know more immediately when the payback will come-and we don’t have a clear picture.”
This uncertainty has led to almost $1.5trn being wiped off the market capitalisation of these tech behemoths leading to the underperformance of the Nasdaq 100 this year.
Beyond the Hyperscalers
However, focusing on hyperscalers spending may miss the deeper structural shift underway.
We saw a preview of this last year when DeepSeek claimed its AI was faster, cheaper and less power hungry than its US peers, igniting fears that saw a selloff in AI leaders including Nvidia which fell 17%.
As AI becomes more sophisticated, it is becoming a greater disruptive force.
We have probably all had conversations about the potential for AI to disrupt various high knowledge occupations like lawyers and doctors, but far less conversations about the potential negative impact on leading edge technology companies, and in particular software as a service companies. These businesses with their high margins and low capex requirements have been keenly sought after by investors, including private equity and private credit. However, it appears they may in fact be among the most vulnerable of all.
Implications for software and private markets
As AI steadily rolls out new applications, investors are becoming increasingly worried about the businesses which will be swept aside or at least seriously disrupted. AI’s rapid improvement in its ability to write code with sophisticated models like Anthropic’s Claude are wreaking havoc amongst application software companies. Fund manager Nick Evans from Polar Capital warns, “We think application software faces an existential threat from AI”. He predicts that most companies will go the way of newspapers in the 2000s when print media was decimated by the internet.
Examples of companies Evans has sold are SAP, ServiceNow, Adobe and Hubspot. Closer to home Xero has fallen more than 60% since June last year.
He is obviously not alone. iShares Expanded Tech-Software Sector ETF (IGV) is down 22%. At the same time there are growing concerns around US BDC’s (Business Development Companies), which lend to small and medium sized US businesses, including significant exposure to software companies which may now be worth far less.
According to Apollo – one of the world’s largest private credit managers – “the (software) industry accounted for roughly 40% of all sponsor-backed private credit” adding that, “about 30% of private equity firepower went into the (software) industry this last decade”.
Earlier this month Anthropic released plugins for its Claude Cowork AI agent claiming the model could automate tasks for customer service, product management, market, legal and data analysis, among others. Insurance brokers tumbled on another programme tied to OpenAI, and yet another from a little know startup Altruist led to a selloff in wealth managers like Charles Schwab. and Raymond James. This vast improvement in AI coding tools means they can already replicate and modify most existing software.
Investing through uncertainty
But there are those who believe this is merely a new narrative without substance and will prove to be a buying opportunity in SaaS companies. LPL Financial, the largest broker-dealer in the US, writes that markets appear to be pricing in a worst-case scenario that may not fully account for the sector’s strength. Dan Ives of WedBush believes the sell-off is “the most disconnected trade I’ve ever seen in my career on Wall Street”. He is of the view that AI will boost many of these software companies, not destroy them. JP Morgan has also described the view that AI companies would disrupt the software industry as “broken logic”, and that investor concern is overblown.
At Saxe Coburg, we don’t chase thematics but allocate to skillful investors who we rely on to navigate markets through periods of disruption and uncertainty. Of all the views expressed above, we would tend to back Nick Evans, whose fund Polar Capital has proven to be a very good investor in technology over many years and is less likely to be conflicted than say a JP Morgan whose broader commercial interests could influence their stance.
Get in touch with Mark or Sam to discuss this article, or how we can help manage your portfolio through disruption.
To read more of our Investment Insights click here.