By Evan Powell, DeepTempo Founder and CEO

The cybersecurity landscape is already going through significant changes driven by more sophisticated attackers and aging platform players. Here are a few key predictions, with the good and the bad, about the evolving dynamics of the industry:

1. Significant Increase in Cybersecurity Spending leave us Further Behind

Global cybersecurity budgets are projected to exceed $250 billion by 2025, according to Gartner. On the other hand, with most analysts estimating a flow of trillions of dollars to attackers, they have the resources to continue to out innovate us. 

2. Cloud-Based Data Platforms Enable Safe AI for Security

Enterprises will increasingly use shared data platforms that serve multiple purposes, including analytics and cybersecurity. Some of these platforms, such as Snowflake, have become a preferred environment in which to run advanced analytics capabilities powered by AI for real-time threat detection and response while retaining the security and data controls of the enterprise. The emergence of these "NativeApp" platforms allows the intelligence to run on the data, reducing the cost and risk of moving data into proprietary security systems. This trend will continue to drive the development of specialized security applications, foster cross-functional data sharing, and create a demand for data engineers and analysts skilled in leveraging these powerful environments.

3. Vendor Lock-in will sap more budgets

While Broadcom's aggressive approach to paying off their VMware acquisition has drawn the most headlines as customers have experienced the costs of vendor lock-in through 200-300% and greater price increases, the same sort of behavior is playing out in cybersecurity, where Cisco is leveraging their acquisition of Splunk to drive up their control of enterprise accounts and in some cases to extract significant price increases. Vertically integrated platforms by their nature lock customers in - and as such every CFO and CISO in the industry should be looking at their exposure to these platforms.  The question is when, and not if, these aging platforms will be used to extract massive price increases.   

4. Under Threat Regulators Try to Stop Cyber Contagion

 At a time of deep skepticism of possible over-regulation - after all reportedly some large enterprises spend almost as much on regulatory compliance as they do on actually preventing and responding to cyber threats - these increasingly scrutinized regulators will be forced to play a more active role in limiting the risks of cyber contagion from attacks on major enterprises.  One approach we are already seeing is regulators insisting that large organizations have an accelerated "get well" plan so that in case of breach, they can quickly restore trust in attacked institutions.  A cyber contagion may be exactly what certain nation-state actors would like to see - so in 2025 we all need to work on containing this risk. 

5. AI Skepticism Will Increase in Volume

Security influencers will repeatedly warn against over-reliance on AI for cybersecurity while offering little in terms of actionable solutions beyond basic best practices. While the talking heads focus on better patching the flood of CVEs and improving our access policies and so on, attackers will leverage AI and augmented insider based attack vectors. To some extent we are still learning that what got us here - humans using simple rules and overwhelmed index based logging systems to see and respond to suspicious behaviors - will not protect us going forward, as attackers morph their attacks to evade rule and traditional ML based indicators. 

6. Burn Out Continues

Along with increased extractive behavior by leading platforms - another sign of the end of times for an industry is the people in the industry becoming burned out.  Unfortunately in 2025 we will continue to see unsustainable levels of burn-out.  There are a number of reasons for this burn out, including the tendency of decision makers within cybersecurity to push favored solutions down to their teams, as opposed to allowing their teams to select and integrate preferred solutions. 

7. Learned Indicators from purpose built AI proliferate

In 2024 we saw forward-looking organizations starting to leverage deep learning and purpose-built Log Language Models to analyze vast amounts of log data, detecting anomalies and potential threats with higher accuracy and speed. This trend will accelerate in 2025, enabling security teams to respond more quickly to threats, while optimizing their security spending, and boosting the productivity and well-being of their teams.

Conclusion

On the whole, I am optimistic.  The problems we face in cybersecurity - and the opportunities to unlock enterprises from an increasingly unsustainable burden - are too great to allow the conditions causing us all to fall further behind to persist.  Savvy organizations that are prepared for ever more sophisticated attacks by their increasingly extractive legacy vendors - and of course by innovative cyber criminals - will build into a capability that allows them to move faster with much greater resilience than their competitors.  Could 2025 be the year that is acknowledged as a competitive strength as opposed to primarily a cost center? 

##

ABOUT THE AUTHOR

Evan Powell, DeepTempo Founder and CEO

Evan Powell is a serial entrepreneur who has founded several enterprise software companies that have built and leveraged open data architectures. His companies have consistently achieved broad adoption, driving the IT industry towards greater openness. His latest company, DeepTempo, developed Log Language Models designed to run on open data lakes, enhancing the detection of cybersecurity threats.