Thread regarding SAS Institute layoffs

@r3

Unfortunately all of that is happening more rapidly at my employer. Translation is mostly done, the senior folks are retiring. Demand for newer use cases is more around LLM usage, so inevitably, budgets will follow. When AI can do the math, you likely need fewer, if any, more senior analysts to help you interpret. What language they use isn’t a large concern. Those days are long gone.

@py

Thanks. Sadly, I agree with you, and see no long-term path for SAS to survive. The AI stacks will belong to the hyperscalers. Nobody else is spending enough to compete.

In the short term, though, an enterprise could decide to change their whole stack -- and yet also renew SAS until their elderly data analysts retire.

So I think @kk has it right: the demographics will end SAS before the AIs do.

@pc

Your trenchant observations ring true. However I think the decline of enterprise software including SAS may be accelerated by the following factor.

Existing customer Enterprises who are rethinking their overall technology strategy with everything evolving to interoperate within an AI-based stack.

SAS will very quickly need to rethink its product innovation strategy and direction if it plans to survive in this kind of world.

"danger for SAS is its aging customer user base"

This is the existential danger.

AI reduces translation costs to near zero. Once AI translates SAS code into R or Python, the younger data analysts, who are trained in both, can easily fix bugs in the translation.

Thus AI and Open Source satisfy two conditions of an inflection point for SAS. But they are not sufficient conditions, because SAS customers still have a switching cost.

Many older analysts love SAS. Retraining them, and waiting until they become fluent in R or Python, would not be cheap. So some SAS customers, even with AI support, still find it cheaper to pay renewal fees than to switch to Open Source.

Once those older analysts retire, the calculation will change. The switching cost will become cheap. Customers will have every incentive to cut expenses by eliminating renewal fees.

Like Open Source, AI is an inflection point for the industry. Both these changes impact SAS new sales and renewals. But its aging customer base is the third condition that creates an inflection point for SAS.

Their functional usage is eroding due to changes in the businesses that deployed them

Because truth and fact are not valued or wanted anymore. Welcome to he-l, you’ll fit right in.

A clear and danger for SAS is its aging customer user base, most who were established during the "golden age" of traditional SAS language programming (1980-2005).

Retirement will accelerate among this population, or the best of them have already or will continue to learn new open source analytics and data management tools. The smart ones are braving the new world of AI.

The massive acceleration of AI coding capabilities and models' abilities to translate language code will likely accelerate the decline of SAS usage in customer sites.

In some cases entire legacy enterprise software systems, including SAS, SAP
along with other vendors will likely see accelerated useagr decline in the short years ahead.

The reasoning is pretty simple. A lot of these systems are near EOL anyway. Their functional usage is eroding due to changes in the businesses that deployed them in the first place. Given these systems actively represent 10s to 100s of $Billions in annual licensing cost, they are prime targets for new AI based products and consultancies to displace.

AI can easily replace the Art Department, which is safe.

AI is an inflection point for the industry. I wish it would be one for SAS.

But SAS will not "effectively eliminate unnecessary fat" and institute "startup discipline". That would be a cultural change, and as such, completely unlikely. More likely, SAS will simply continue milking its renewal revenue stream.

For enterprise software vendors, the impact of AI may be similar to that of Open Source. Open Source is an attractive choice for new customers. But for existing customers, there's a switching cost, so it's often cheaper to keep paying the renewal fees.

Some existing customers do switch; combined with fewer new customers, that has caused SAS revenues to decline against expenses. To ensure profitability, SAS has lately reduced headcount by ~2-3% each year.

AI will attract new customers and give existing customers another reason to switch. Instead of ~2-3%, AI could require SAS to reduce headcount by ~4-5% each year -- not sudden death, but a faster decline.

If SAS fails to effectively eliminate unnecessary fat in a timely matter and institute "startup discipline", then the game is going to be over soon sooner than later."

When belt tightening is not across the board, morale suffers and workers feel that the founders simply do not care.

Saving the company should always be more important than protecting pet area .

There are other subtle cues in this presentation going beyond the "layoffs are necessary to fund AI CapEx due to insufficient free cash flow" narrative.

Around @2:00, as this narrative unfolds, the notion that reducing management layers and "operating as the world's largest startup" Is nevertheless wise -- for massive hyper skaters like Amazon and by comparison relatively small legacy data management and analytics companies like SAS.

We don't know what the SAS balance sheet looks like. It is opaque to all but a Relative handful of people, none of whom comment here.

Clearly SAS does not have a business model that requires building out extensive AI data centers, yet is going to have to radically step up its AI game to survive.

@9:00 the presentation gets more to the heart of "what can SAS learn?" ... anyone who spent years or decades at SAS, ACTUALLY AWAKE, can see the parallel.

If SAS fails to effectively eliminate unnecessary fat in a timely matter and institute "startup discipline", then the game is going to be over soon sooner than later.

If this video points to anything, it's that the entire tech industry has reached an inflection point and there is no going back. Naysayers can crow all they want, but AI is accelerating the irrelevance of legacy enterprise software at a sobering and astonishing rate.

The video argues that even mighty Amazon had insufficient cash flow to invest in AI. So they laid off 30,000 people to free up the money. To become an "AI company" requires that level of investment.

But there are AI applications that don't require such massive investment. E.g. if I load my data into Viya and command it "Analyze This!" does it deliver useful results?

SAS should build such niche applications -- or buy them -- before the big companies do.

Nothing?

We can hope that SAS learns but does not imitate.