On the surface, it might seem logical that artificial intelligence, as a tool to automate threat detection and replace manual security processes could displace some functions of current threat protection apps, including SASE (Secure Access Service Edge).
On the other hand, it is pretty hard to find any major industry analyst report that supports that line of thinking. AI represents new attack surfaces, for example, arguably increasing the need for SASE.
On the other hand, financial analysts seem to universally believe the AI danger to enterprise software is significant. And there’s no absolutely-clear way to know which view is correct.
There are pros and cons to the argument, as you would guess.
As sometimes happens, market analysts and financial analysts tend to disagree, for reasons related to business models.
Large market researchers are paid by vendors and suppliers who buy research subscriptions, commission custom reports, and pay for placement in analyst programs.
Enterprise buyers also subscribe, but vendors are typically the bigger revenue source and the more active relationship.
The resulting biases:
Market size inflation. A large total addressable market forecast makes a vendor's pitch deck look compelling, justifies investment in the space, and makes the analyst firm look like it spotted a major trend early. There is almost no commercial downside to a bullish forecast, as nobody fires their Gartner subscription because a market grew slower than predicted. Forecasts are inherently unauditable in the short run, and by the time they're proven wrong, a new forecast has replaced them.
“New” markets and categories create buyer demand. When a vendor wants to differentiate their products, they often work closely with analyst firms to define and name a new category. The vendor gets a category it conveniently leads; the analyst firm gets cited as the authoritative source of the framework. The bias is toward proliferating categories rather than consolidating them, because each new category is a new revenue opportunity.
Optimistic adoption curves. Researchers consistently underestimate the friction of enterprise adoption. Their models tend to treat "total addressable market" as if it were "realistically serviceable market in the next three years," producing forecasts that flatter suppliers' sales projections.
Vendor-funded research. Commissioned studies. where a vendor pays for research that it then cites, are structurally compromised. The findings rarely bite the hand that feeds.
Financial analysts (Sell-Side and Buy-Side) have different revenue models. Sell-side analysts at investment banks are ultimately paid through trading commissions and investment banking relationships (equity research is largely a loss leader that supports deal flow).
The resulting biases:
Structural bullishness on covered stocks. Issuing a Sell on a company damages the relationship with that company's management, threatens future access to executives, and risks losing investment banking business. This means technology assessments of publicly traded companies are systematically skewed upward.
Recency and momentum bias. Analysts are rewarded for being right in the near term. A technology with strong recent earnings will get upgraded; one stumbling will get downgraded.
Narrative over fundamentals during hype cycles. Missing a major rally in a sector you cover is more career-damaging than being wrong alongside everyone else. This produces herd behavior..
Coverage selection bias. Analysts choose what to cover, and they tend to cover companies where there's trading volume and banking opportunity. Small, potentially disruptive competitors often go uncovered until they're large enough to matter.
Market researchers inflate the supply-side opportunity (how big is the market, how fast will it grow).
Financial analysts inflate the demand-side story (which incumbent captures value).
Market researchers tend to see AI as an unambiguous expansion of the enterprise technology market. Their instinct is additive and are structurally inclined to frame AI as a rising tide.
Financial analysts face a much harder problem, because AI introduces several simultaneous dynamics that are deeply ambiguous for incumbent valuations:
Commoditization risk: If AI compresses the differentiation between enterprise software products, then the moats that justified premium multiples erode
Capex displacement (some categories might shrink as others grow)
Margin uncertainty
Value uncertainty (will value for app-layer firms be threatened by alternatives?)
Market researchers tend to view AI as more incumbent friendly, where financial analysts see more threats to traditional seat license revenue models, for example.
So one might argue market researchers are looking at “how much is being spent” where financial analysts are looking at “who captures the value?”
Either way, there is huge uncertainty about the “right” level of valuation for enterprise software firms.
No comments:
Post a Comment