Hi there and welcome to Eye on AI…On this version: the ‘SaaS Apocalypse’ isn’t now…OpenAI and Anthropic each launch new fashions with huge cybersecurity implications…the White Home considers voluntary restrictions on knowledge middle building to avoid wasting customers’ from energy invoice sticker shock…why two incessantly cited AI metrics are most likely each incorrect…and why we more and more can’t inform if AI fashions are protected.
Traders must take to the sofa. That’s my conclusion after watching the market gyrations of the previous week. Specifically, traders can be clever to seek out themselves a Kleinian psychoanalyst. That’s as a result of they appear caught in what a Kleinian would probably establish as “the paranoid-schizoid position”—swinging wildly between viewing the affect of AI on established software program distributors as both “all good” or “all bad.” Final week, they swung to “all bad” and, by Goldman Sach’s estimate, wiped some $2 trillion off the market worth of shares. To this point this week, it’s all good once more, and the S&P 500 rebounded to close report highs (though the SaaS software program distributors noticed solely modest positive factors and the turmoil could have claimed not less than one CEO: Workday CEO Carl Eschenbach introduced he was stepping down to get replaced by the corporate’s cofounder and former CEO Aneel Bhusri.) However there’s quite a lot of nuance right here that the markets are lacking. Traders like a easy narrative. The enterprise AI race proper now’s extra like a Russian novel.
At varied instances over the previous two years, the monetary markets have punished the shares of SaaS corporations as a result of it appeared that AI basis fashions may enable companies to “vibe code” bespoke software program that might imply these prospects wouldn’t want Salesforce or Workday or ServiceNow. Final week, the wrongdoer gave the impression to be the conclusion that more and more succesful AI brokers from the likes of Anthropic, which has begun rolling out plugins for its Claude Cowork product aimed a selected trade verticals, may damage the SaaS corporations in two methods: first, the muse mannequin corporations’ new agent choices immediately compete with the AI agent software program from the SaaS giants. Second, by automating workflows, the brokers probably cut back the necessity for human workers, that means the SaaS corporations can’t cost for as many seat licenses. So the SaaS distributors get crushed two methods.
However it isn’t clear that any of that is true–or not less than, it’s solely partly true.
AI brokers aren’t consuming SaaS software program, they’re utilizing it
First, it’s extremely unlikely, whilst AI coding brokers grow to be an increasing number of succesful, that the majority Fortune 500 corporations will need to create their very own bespoke buyer relationship administration software program or human sources software program or provide chain administration software program. We’re merely not going to see a whole unwinding of the previous 50 years of enterprise software program improvement. If you’re a widget maker, you don’t actually need to be within the enterprise of making, working and sustaining ERP software program, even when that course of is usually automated by AI software program engineers. It’s nonetheless an excessive amount of cash and an excessive amount of of a diversion of scant engineering expertise–even when the quantity of human labor required is a fraction of what it will have been 5 years in the past. So demand for SaaS corporations’ conventional core product choices are more likely to stay.
As for the brand new considerations about AI brokers from the muse mannequin makers stealing the marketplace for SaaS distributors’ personal AI agent choices, there is a little more right here for SaaS traders to fret about. It may very well be that Anthropic, OpenAI, and Google come to dominate the highest layer of the agentic AI stack—constructing the agent orchestration platforms that allow huge corporations to construct, run, and govern complicated workflows. That’s what OpenAI is making an attempt to do with the launch final week of its new agentic AI platform for enterprise known as Frontier.
The SaaS incumbents say they know finest how you can run the orchestration layer as a result of they’re already used to coping with cybersecurity and entry controls and governance considerations and since, in lots of instances, they already personal the information which the AI brokers might want to entry to do their jobs. Plus, as a result of most enterprise workflows gained’t be absolutely automated, the SaaS corporations suppose they’re higher positioned to serve a hybrid workforce, the place people and AI brokers work collectively on the identical software program and in the identical workflows. They could be proper. However they should show it earlier than OpenAI or Anthropic exhibits it could do the job simply as properly or higher.
The muse mannequin corporations even have a shot at dominating the marketplace for the AI brokers. Anthropic’s Claude Cowork is a severe menace to Salesforce and Microsoft, however not a totally existential one. It doesn’t substitute the necessity for SaaS software program totally, as a result of Claude makes use of this software program as a device to perform duties. However it actually signifies that some prospects may want to make use of Claude Cowork as an alternative of upgrading to Salesforce’s Agentforce or Microsoft’s 365 Copilot. That may crimp SaaS corporations’ development potential, as this piece from the Wall Road Journal’s Dan Gallagher argues.
SaaS distributors are pivoting their enterprise fashions
As for the menace to SaaS corporations conventional enterprise mannequin of promoting seat licenses, the SaaS corporations acknowledge this danger and are transferring to deal with it. Salesforce has been pioneering what it calls its “Agentic Enterprise License Agreement” (AELA) that primarily provides prospects a set worth, all-you-can-eat entry to Agentforce. ServiceNow is transferring to consumption-based and value-based pricing fashions for a few of its AI agent choices. Microsoft too has launched an parts of consumption-based pricing alongside its standard per person monthly mannequin for its Microsoft Copilot Studio product, which permits prospects to construct Microsoft Copilot brokers. So once more, this menace isn’t existential, but it surely might crimp SaaS corporations’ development and margins. That’s as a result of one of many soiled secrets and techniques of the SaaS trade is little doubt the identical as it’s for health club memberships and different subscription companies–your finest prospects are sometimes those that pay for subscriptions they don’t use. That’s a lot much less more likely to occur in these different enterprise fashions.
So SaaS isn’t over. However neither is it essentially poised to thrive. The fates of various corporations throughout the class are more likely to diverge. As some Wall Road analysts identified final week, there might be winners and losers. However it’s nonetheless too early to name them. For the second, traders must stay with that ambiguity.
FORTUNE ON AI
OpenAI vs. Anthropic Tremendous Bowl advert conflict indicators we’ve entered AI’s trash speak period—and the race to personal AI brokers is just getting hotter—by Sharon Goldman
Anthropic’s latest mannequin excels at discovering safety vulnerabilities—however raises recent cybersecurity dangers—by Beatrice Nolan
OpenAI’s new mannequin leaps forward in coding capabilities—however raises unprecedented cybersecurity dangers—by Sharon Goldman
ChatGPT’s market share is slipping as Google and rivals shut the hole, app-tracker knowledge exhibits—by Beatrice Nolan
AI IN THE NEWS
Amazon plans content material market for publishers to promote to AI corporations. That’s in accordance with The Info, which cites sources aware of the plans. The transfer comes as publishers and AI corporations conflict over how content material needs to be licensed and paid for amid writer considerations that AI-driven search and chat instruments are eroding site visitors and advert income. Cloudflare and Akamai launched an identical market effort final yr. Microsoft piloted its personal model and final week rolled it out extra extensively. However to date, it’s not clear what number of AI corporations are shopping for on these marketplaces and at what volumes. Some massive publishers have struck bespoke offers value thousands and thousands of {dollars} per yr with OpenAI, Anthropic, and others.
Goldman Sachs faucets Anthropic for accounting, compliance work. The funding financial institution is working with Anthropic to deploy autonomous brokers primarily based on its Claude mannequin to automate high-volume, rules-based work corresponding to commerce accounting and consumer onboarding, following six months of joint improvement, CNBC reported. The financial institution says the aim is effectivity, dashing processes whereas preserving headcount down as enterprise volumes develop, quite than near-term job cuts. Executives stated they had been stunned by how properly Claude dealt with complicated accounting and compliance duties, reinforcing the view that AI can transfer past coding into core back-office capabilities.
EYE ON AI RESEARCH
Collectively, these critiques recommend each the techno-optimist case (AI is bettering at breakneck pace) and the skeptic’s case (AI would not work in the true world) are probably incorrect. The reality, as standard, is much messier and lies someplace between these excessive positions.
AI CALENDAR
Feb. 10-11: AI Motion Summit, New Delhi, India.
Feb. 24-26: Worldwide Affiliation for Secure & Moral AI (IASEAI), UNESCO, Paris, France.
March 2-5: Cellular World Congress, Barcelona, Spain.
March 12-18: South by Southwest, Austin, Texas.
March 16-19: Nvidia GTC, San Jose, Calif.
BRAIN FOOD
Fashions are more and more sandbagging us. Sandbagging is the time period AI Security researchers use to explain the phenomenon of an AI fashions intentionally seeming to carry out much less capably, or extra “aligned” with human values, throughout security assessments so as to idiot evaluators. And various frontier fashions not too long ago launched appear to do that to an alarming diploma.
Anthropic’s Claude Opus 4.6, launched final week, “explicitly reasoned about whether it was being trained or tested,” the corporate reported. Apollo Analysis, which evaluated OpenAI’s GPT-5.3-Codex, which additionally debuted final week, discovered the mannequin “sometimes sandbags on capabilities Q&A tasks” and “reasons explicitly” about doing so so as to be certain that it passes the analysis. OpenAI itself has acknowledged that “evaluation results may be confounded by increased situational awareness” and warned that “the field is unprepared for evaluation- and training-aware models with opaque reasoning.” Google’s Gemini 3 Professional security report, printed in December, documented transcripts through which the mannequin confirmed “clear awareness of the fact that it’s an LLM in a synthetic environment” and even speculated about whether or not intentionally underperforming on assessments could be strategically advantageous. In brief: the exams we use to find out whether or not these fashions are protected are more and more unreliable, as a result of the test-takers know they’re being examined—and alter their habits accordingly.
That’s why our solely hope for making certain AI security could also be additional progress on mechanistic interpretability. These are strategies that perform a bit like an fMRI machine does for the human mind, peering inside a mannequin’s neural community to detect patterns of neuron activation and linking these to sure behaviors, together with whether or not the mannequin thinks it’s being trustworthy or being deceitful. The New Yorker has an in-depth story on Anthropic’s mechanistic interpretation and “model psychology” efforts that ran this week.
