Voice of AI: The leap is coming. Those who wait will lose. 🚀
- Ralph Schwehr
- Mar 23
- 4 min read
A single number has been shaking up the strategic halls of the global economy this month: 83.0 percent. That's the score achieved by OpenAI's new model, GPT-5.4 "Thinking," on the GDPVal benchmark, a test that doesn't solve academic puzzles but evaluates real-world professional work: sales presentations, accounting spreadsheets, and production drawings across 44 professions and nine industry sectors. This means AI outperforms human experts in four out of five measurable tasks. Morgan Stanley calls it a "non-linear quantum leap" and predicts the next surge will come this quarter. The question is no longer technical. It's strategic: Which companies are prepared?
From promise to measurement
AI is no longer a gamble on the future. It's a key metric in the profit and loss statement. Morgan Stanley surveyed around 1,000 executives in five countries, and the result was both sobering and revealing: On average, workforces were reduced by a net 4% in the past twelve months, directly attributable to AI adoption. At the same time, economists at the University of Chicago are reporting, for the first time, macroeconomically visible productivity gains from AI in aggregated economic data. The evidence from real-world practice has overtaken the evidence from the models.
OpenAI CEO Sam Altman summed it up perfectly at the Morgan Stanley TMT Conference: The world is unprepared. "Extremely capable models" will arrive faster than originally thought. His vision of companies employing one to five people yet competing with established corporations sounds radical. It will become reality in the coming years.
The Agentic wave: 40% by year-end
What's currently happening in companies has a name: Agentic AI. These are systems that don't wait for prompts, but plan, decide, and act independently. Gartner predicts that by the end of 2026, 40% of all enterprise applications will be integrated with task-specific AI agents, compared to less than 5% at the beginning of the year. The global market for Agentic AI has surpassed the $9 billion mark.
But Deloitte warns clearly and with evidence: The majority of agentic AI implementations fail. The reason rarely lies in the technology itself. It lies in the lack of process design. Those who implement agents on top of existing structures exacerbate their weaknesses instead of addressing them. The winners redesign their processes before automating. The difference between transformation and costly disappointment isn't a budget issue. It's a question of strategic maturity – we're here to support you at oakai.de .
Industrial giants show how it's done
Two examples from this week illustrate the difference between talk and action particularly clearly. At NVIDIA GTC 2026, SAP presented its "Joule Agent" system: AI agents that coordinate tasks across the entire SAP portfolio , model-agnostic and deeply embedded in business processes. The goal is not a chatbot for HR. It's the AI-native enterprise where AI forms the operational backbone.
Mastercard is taking a different, equally instructive approach. Instead of language models, the payment service provider relies on a Large Tabular Model (LTM), trained on structured transaction data. Use cases include cybersecurity, loyalty programs, and portfolio optimization. The underlying message is important: AI transformation is not a one-size-fits-all solution. Those who understand and leverage their company's specific data foundation build genuine competitive advantages.
The invisible bottleneck: Energy
Behind the AI boom lies a physical problem that is still underestimated in the strategic calculations of many companies. Morgan Stanley's "Intelligence Factory" model projects a power deficit of 9 to 18 gigawatts for the US alone by 2028, which corresponds to 12 to 25% of the required capacity. The industry is already reacting: Bitcoin mining farms are being converted into data centers, gas-fired power plants are being reactivated, and fuel cells are being installed.
For European companies, this has a strategic consequence that goes far beyond cloud provider preferences. The question of where AI infrastructure is located is becoming a location decision with geopolitical dimensions. Those who secure capacity today will have a competitive advantage tomorrow.
M&A: The New Gamebook
The global M&A market in the technology sector grew by 76% to $903 billion in 2025. AI transactions alone accounted for $117 billion, an increase of 125% compared to the previous year. McKinsey describes the current phase as the "industrial maturity" of AI M&A, and thus as a structural shift in deal logic.
Today's acquisition targets aren't looking for economies of scale. They're securing access to proprietary data, model IP, and talent unavailable elsewhere. The challenge: In so-called "access-driven deals," it's often unclear at the time of contract signing exactly what's being transferred. What remains with the seller, what's shared, and what's licensed? The old due diligence frameworks simply don't work in this new reality. Companies that recognize this and adapt their acquisition strategy will lead the consolidation wave, not be swept away by it.
The most important information at a glance
83% on GDPVal: GPT-5.4 outperforms human experts in 4 out of 5 professional tasks (OpenAI / Morgan Stanley, March 2026)
4% net job reduction in 12 months, directly attributable to AI adoption (Morgan Stanley, 1,000 executives, 5 countries)
40% of all enterprise apps will integrate AI agents by the end of 2026; today it's less than 5% (Gartner).
$117 billion in AI M&A volume in 2025, up 125% compared to 2024 (GlobalData)
Up to 18 GW of power deficit by 2028 threatens global AI infrastructure growth (Morgan Stanley)
Sources
title | Publisher | Date | link |
Morgan Stanley: AI Breakthrough 2026 | Fortune | March 13, 2026 | |
AI Job Displacement: Morgan Stanley TMT Conference | Fortune | March 12, 2026 | |
40% of Enterprise Apps with AI Agents by 2026 | Gartner Newsroom | August 26, 2025 | |
Agentic AI in Enterprise 2026: $9B Market Analysis | Tech Insider | March 18, 2026 | |
AI Key Driving Force for TMT Deal Activity 2025 | GlobalData/Advanced Television | March 12, 2026 | |
Technology M&A: AI Enters Its Industrial Phase | McKinsey & Company | February 13, 2026 | |
How SAP and NVIDIA Advance AI for Enterprise | SAP Newsroom | March 18, 2026 | |
Inside Mastercard's New Gen AI Engine | Mastercard Newsroom | March 17, 2026 | |
Agentic AI Strategy | Deloitte Insights | Dec. 2025 |
Ready for the next step?
The AI transformation is no longer a question of "if". It's a question of "when" and "how well prepared". Companies that are still taking a cautious, observational approach today will find in two years that the turning point has long since passed.
OAKAI supports companies from the initial assessment of their current situation to the operational integration of AI. From AI impact analysis and strategy development to implementation in core processes.
Let's talk: info@oakai.de
oakai.de | The future is not a matter of chance. It is a decision.
Comments