by Bhawani Shankar

Is the cloud seducing software giants into the enemy lair?

Opinion
Dec 12, 20247 mins

Despite cynical theories of slowdown, the Big 4 software giants’ growth is rocket-fuelled by cloud migrations, as their recent quarterly reports show. But could the looming "service software" AI tsunami threaten the very business they are trying to sustain?

Bhawani Shankar
Credit: Bhawani Shankar

Sometimes enemies can be lifesavers. The $150m loan from Microsoft that rescued Apple from bankruptcy in 1997, for instance, is now the most widely read chapter of software folklore. Microsoft was under threat in the antitrust case and needed to show it wasn’t a greedy monopoly and that Apple was a credible rival; Apple, on the other hand, needed a lifeline — and cash — to survive. The rest may well be history but the question that begs asking is if the massive push into cloud migration will lead software giants like SAP and Oracle into relative obscurity.  

Gates and Jobs did it with their eyes wide open. Today’s software behemoths, however, are riding high on the crest of cloud business while the impact that generative artificial intelligence (genAI) could have on their core business feels more real every day. Here’s the story of the three key players that more or less rule the roost.  

Tentative starts to the cloud-ward journey  

A strongly premise-based software industry didn’t readily take to the cloud. For Oracle, the strategy was bottom-up: build the cloud infrastructure to ensure there was a foundation. At that time, Oracle’s perceived rival was IBM. The real boost came first with the Sun Microsystems acquisition that provided a hardware layer and then later when NetSuite came into the family in 2016 — that trained the guns at SAP.  

But that is history. From that start to today, Oracle has come a very long way indeed. Its second fiscal quarter of 2025, reported on December 9, shows a 24% increase in IaaS and SaaS revenues, which, at $5.9 billion, is a little over 40% of its quarterly earnings of $14 Billion.  

Exceeding expectations…for now  

The main story is the mid-term: Douglas Kehring, Oracle’s Exec V-P for Operations announced in the September 2024 financial analyst meeting that there is “more than $99 billion” of guaranteed revenue to be operationalized — this is what the SaaS industry calls remaining performance obligations (RPO) which are a sum of deferred and backlog revenue. Presumably, this is largely from future IaaS and SaaS work, and this has prompted Oracle to raise expectations for 2025-26 revenues significantly.  

SAP’s results, too, are leaving Wall Street analysts guessing: Its October 2024 earnings report showed that moving its large enterprise installed base to the cloud is not only big business but also highly profitable. This accounted for half of its $9.2 billion in quarterly revenue with profits up 27%. As with Oracle, SAP’s backlog of guaranteed revenue stands at about $16 billion — so there’s more to come.  

Enter genAI  

Health warning: In the genAI world, a week is equivalent to what we know conventionally as one year! An exploding amount of innovation means that what we discuss here is probably obsolete after its publication. Consider this: In 2023, GitHub’s new code converter app hit the road with the target of converting code from one programming language to another. It seemed fairly harmless as open-source coders played one language against another.   

Within weeks, and perhaps unconnectedly, many more experiments hit the road. Swimm, a product launched in 2018, found new life in providing documentation for legacy code combined with the power of LLMs to create new, AI-fuelled code in a different language. Parsing documentation is, by now, child’s play for genAI and paves the way for a gradual erosion of legacy code directories — exactly what many of these experiments are aiming at.  

As we speak today, converting Cobol — which still powers cash machines around the world — is becoming commonplace. IBM and many others (Ispirer, Astadia, Codingfleet which converts it to Python…) have implemented this; the conversion of SAS analytics code to Python is becoming widespread (Pythonconvert, SAS2PY, which claims 95% automated conversion, for instance, and many more).  

From source-available to open source  

Growing from that point, can we lean on genAI to help convert SAP and Oracle custom code to open source? Well…not yet, but it is starting. And in genAI timescales, this may be much sooner than we are used to. Migrating to the cloud has certainly heightened the possibility, which would have been harder when custom code was under lock and key on-premise. The first Oracle database code and schema conversions have begun using Gemini.  

In reality, the possibility of converting entire spans of custom applications that have decades of vested development behind them isn’t around the corner. But it is certainly true that the doors are starting to open — often by the major software players themselves. The complete source code for SAP’s ABAP system, for instance, is now available (“source available”).  

As of August 2024, we have the first open-source LLM trained on SAP APIs and ABAP. This has several advantages: The first open-source LLM to learn ABAP and SAP APIs, HANA SQL, is multilingual and allows commercial use — such as building your SaaS with it. The push of open source into software kingdoms is now inexorable. Corteza, ERPNext and several others are entirely open-source alternatives with Odoo gathering significant traction.  

Even a partial conversion of some custom code can be interesting. According to Diego Lo Giudice, vice president and principal analyst at Forrester, “This is the beginning of a new world, where the cost of migrating legacy systems to modern technology stacks is no longer prohibitive.” Organizations will go from not touching legacy to migrating or regenerating it into new and more modern applications.  

Agentic futures and their impact on the business  

So, what of SaaS firms that are already there in the cloud? If Salesforce is the beacon, agentic futures are the road ahead. Combined with the potential of conversion to open-source code, agentic will also disrupt the commercial model.  

As agents evolve, the service or function they provide will become the software. Pay-per-use or function will become an outcome-based pricing model. According to Alex Kroman, CPTO at LivePerson and formerly SVP at New Relic, “The promise of software will finally be fulfilled — not as a tool we have to master, but as a partner that helps us reach our full potential.”  

Could it be then that Salesforce and other prominent SaaS firms are better placed to make the transition from SaaS to agentic to a consumption-driven world of service software?   

If you consider Salesforce’s recent quarterly results, its 3rd quarter results show a mere 8% growth relative to Oracle and SAP’s 24% and 27% cloud revenue growth. On the one hand, you could say catching up with the cloud is proving profitable. But equally, the inevitable question is if an accelerated and, perhaps, belated migration might be leading them to a world where a more radical business change awaits them. 

For more than 30 years, Bhawani Shankar has worked across industries initially as a deputy CIO, then as an analyst and consultant — including eight years at Gartner — followed by business leadership roles with tier one technology services firms including HCL Tech, Cognizant and Wipro. Since 2022, he has been CEO at Cubix, an advisory firm that helps enterprises select and buy AAA products and services.